2017-10-01T00:15:52Z dougk__ joined #sbcl
2017-10-01T00:15:53Z dougk quit (Remote host closed the connection)
2017-10-01T00:22:06Z whoman` joined #sbcl
2017-10-01T00:26:41Z whoman quit (Ping timeout: 260 seconds)
2017-10-01T00:40:31Z abbe: Hi!
2017-10-01T00:40:47Z abbe: getting error building SBCL 1.4.0 on FreeBSD (amd64) -- https://pastebin.com/pe5By1Y3
2017-10-01T00:44:25Z attila_lendvai quit (Quit: Leaving.)
2017-10-01T00:44:49Z whoman` is now known as whoman
2017-10-01T01:11:32Z stassats: abbe: build using make.sh --without-immobile-space --without-immobile-code --without-compact-instance-header
2017-10-01T01:11:47Z stassats: as a workaround
2017-10-01T01:15:37Z abbe: thanks!
2017-10-01T01:16:03Z abbe: stassats: also like me to report this as bug on bugtracker ?
2017-10-01T01:16:28Z stassats: never hurts
2017-10-01T01:22:46Z abbe: workaround works, thanks!
2017-10-01T01:24:53Z Bike quit (Ping timeout: 258 seconds)
2017-10-01T01:26:35Z whoman quit (Ping timeout: 240 seconds)
2017-10-01T01:31:42Z stassats quit (Ping timeout: 240 seconds)
2017-10-01T01:41:41Z Bike joined #sbcl
2017-10-01T01:54:55Z dougk__ quit (Ping timeout: 240 seconds)
2017-10-01T01:55:13Z stylewarning: are there any ways to "trick" the runtime into thinking more memory has been allocated than really
2017-10-01T01:58:40Z dougk joined #sbcl
2017-10-01T01:59:59Z dougk quit (Remote host closed the connection)
2017-10-01T02:00:00Z dougk__ joined #sbcl
2017-10-01T02:02:46Z stylewarning: question #2: did I receive warning from someone here earlier that finalizers were a bit flaky/unreliable?
2017-10-01T02:28:26Z stylewarning: I cannot for the life of me get finalizers to run
2017-10-01T02:28:55Z stylewarning: finalizers on simple-array objects
2017-10-01T02:30:43Z stylewarning: there's no issue with a bad closure, and no issue with re-entrancy
2017-10-01T02:40:54Z Bicyclidine joined #sbcl
2017-10-01T02:40:54Z Bicyclidine quit (Read error: Connection reset by peer)
2017-10-01T02:41:42Z Bicyclidine joined #sbcl
2017-10-01T02:43:17Z Bike quit (Ping timeout: 255 seconds)
2017-10-01T02:43:19Z stylewarning: Maybe it has to do with page granularity
2017-10-01T02:46:44Z stylewarning: But granularity isn’t relevant for collection AFAICT, just pinning
2017-10-01T02:52:34Z dougk__ quit (Read error: Connection reset by peer)
2017-10-01T02:52:52Z dougk joined #sbcl
2017-10-01T02:56:33Z stylewarning: I'm dumb, it's because the collector wouldn't be in charge of collecting the simple-array anyway, because it's not in the Lisp heap
2017-10-01T03:29:35Z sjl_ quit (Ping timeout: 248 seconds)
2017-10-01T03:30:57Z sjl_ joined #sbcl
2017-10-01T03:35:25Z sjl_ quit (Ping timeout: 248 seconds)
2017-10-01T03:37:03Z sjl_ joined #sbcl
2017-10-01T03:58:26Z Bicyclidine quit (Ping timeout: 255 seconds)
2017-10-01T04:04:27Z stylewarning: can I allocate something in immobile but collected space?
2017-10-01T04:06:18Z stylewarning: all I care is that I can get a very large chunk (100s of MB of contiguous space) of memory that is managed by GC but won't move
2017-10-01T04:07:59Z stylewarning: (MAKE-STATIC-VECTOR looks promising)
2017-10-01T04:13:32Z stylewarning: hm, static space is too small :(
2017-10-01T04:34:07Z dougk quit (Ping timeout: 248 seconds)
2017-10-01T04:34:50Z dougk joined #sbcl
2017-10-01T04:44:19Z dougk quit (Read error: Connection reset by peer)
2017-10-01T04:44:34Z dougk joined #sbcl
2017-10-01T05:02:59Z dougk quit (Ping timeout: 240 seconds)
2017-10-01T05:17:26Z karswell_ joined #sbcl
2017-10-01T05:17:39Z karswell quit (Remote host closed the connection)
2017-10-01T05:39:11Z sjl_ quit (Ping timeout: 248 seconds)
2017-10-01T05:41:41Z whoman joined #sbcl
2017-10-01T07:24:52Z slyrus joined #sbcl
2017-10-01T07:48:23Z angavrilov joined #sbcl
2017-10-01T08:25:14Z attila_lendvai joined #sbcl
2017-10-01T08:30:24Z shka_ joined #sbcl
2017-10-01T08:42:50Z attila_lendvai quit (Quit: Leaving.)
2017-10-01T08:49:01Z knobo quit (Ping timeout: 248 seconds)
2017-10-01T08:57:32Z josemanuel joined #sbcl
2017-10-01T10:43:11Z Bike joined #sbcl
2017-10-01T10:49:34Z Bike quit (Ping timeout: 248 seconds)
2017-10-01T11:01:55Z stassats joined #sbcl
2017-10-01T11:02:13Z stassats: stylewarning: large objects wouldn't move anyway
2017-10-01T11:20:29Z les quit (Quit: "")
2017-10-01T11:20:45Z les joined #sbcl
2017-10-01T11:54:30Z stassats quit (Remote host closed the connection)
2017-10-01T11:54:46Z stassats joined #sbcl
2017-10-01T11:59:26Z stassats quit (Ping timeout: 248 seconds)
2017-10-01T12:08:48Z Bike joined #sbcl
2017-10-01T12:09:24Z knobo joined #sbcl
2017-10-01T12:36:30Z knobo quit (Ping timeout: 240 seconds)
2017-10-01T13:02:25Z dougk joined #sbcl
2017-10-01T13:09:21Z karswell_ quit (Ping timeout: 240 seconds)
2017-10-01T13:12:26Z stassats joined #sbcl
2017-10-01T13:40:33Z dougk: stylewarning: finalizers work; was it you who asked whether they run in the final GC prior to save-lisp-and-die? That's the one time they don't run.
2017-10-01T13:41:35Z dougk: and yes you can put vectors in immobile gc-able space. You'll be into new territory though; totally doable but hasn't been done
2017-10-01T14:54:37Z shka_ quit (Quit: Konversation terminated!)
2017-10-01T15:07:27Z shka joined #sbcl
2017-10-01T15:09:20Z whoman quit (Read error: Connection reset by peer)
2017-10-01T15:16:44Z dougk quit (Ping timeout: 248 seconds)
2017-10-01T15:26:23Z josemanuel quit (Quit: leaving)
2017-10-01T15:51:28Z stassats quit (Ping timeout: 240 seconds)
2017-10-01T15:57:59Z Xach: I don't really understand where this error is coming from - http://report.quicklisp.org/2017-10-01/failure-report/cl-freetype2.html#cl-freetype2-doc
2017-10-01T15:58:03Z Xach: any hints?
2017-10-01T15:59:21Z Xach: https://github.com/gwkkwg/cl-markdown/blob/master/dev/spans.lisp isn't helping me grok it
2017-10-01T16:17:25Z |3b|: is that the bit about not being able to use NIL to ignore things anymore?
2017-10-01T16:18:06Z |3b|: at https://github.com/gwkkwg/cl-markdown/blob/master/dev/extension-mechanisms.lisp#L78 maybe?
2017-10-01T16:18:16Z |3b| has no idea what BIND does with that
2017-10-01T16:19:27Z |3b|: seem to match though, T is 4th thing in body, NIL is 4th thing in BIND
2017-10-01T16:21:57Z |3b|: though looks like BIND is specified to allow that use of NIL, so would be a bug in that if so
2017-10-01T16:28:57Z |3b| is confused by the code in BIND that is supposed to handle that :p
2017-10-01T16:29:14Z |3b|: but it does look like it fails to handle it correctly
2017-10-01T16:29:30Z |3b|: Xach: ^
2017-10-01T16:53:06Z jack_rabbit quit (Quit: Leaving)
2017-10-01T16:55:31Z karswell_ joined #sbcl
2017-10-01T17:23:17Z whoman joined #sbcl
2017-10-01T18:04:29Z stylewarning: dougk_: that was me who asked that. My finalizer confusion turned out to be confusion about what the garbage collector handles
2017-10-01T18:15:35Z slyrus quit (Quit: Client Quit)
2017-10-01T18:16:28Z sjl_ joined #sbcl
2017-10-01T18:27:04Z stassats joined #sbcl
2017-10-01T18:32:34Z knobo joined #sbcl
2017-10-01T18:32:38Z shka quit (Read error: Connection reset by peer)
2017-10-01T18:33:02Z shka joined #sbcl
2017-10-01T18:39:10Z shka quit (Read error: No route to host)
2017-10-01T18:39:31Z shka joined #sbcl
2017-10-01T18:42:02Z shka quit (Read error: Connection reset by peer)
2017-10-01T18:42:28Z shka joined #sbcl
2017-10-01T18:49:39Z shka quit (Read error: No route to host)
2017-10-01T18:50:02Z shka joined #sbcl
2017-10-01T18:54:03Z leah2 quit (Ping timeout: 258 seconds)
2017-10-01T18:58:48Z leah2 joined #sbcl
2017-10-01T19:12:39Z Jesin quit (Quit: Leaving)
2017-10-01T19:19:38Z knobo quit (Read error: Connection reset by peer)
2017-10-01T19:19:40Z Jesin joined #sbcl
2017-10-01T19:30:54Z knobo joined #sbcl
2017-10-01T19:35:29Z knobo quit (Read error: Connection reset by peer)
2017-10-01T20:15:38Z dougk joined #sbcl
2017-10-01T20:19:58Z dougk quit (Ping timeout: 240 seconds)
2017-10-01T20:25:06Z solene quit (Remote host closed the connection)
2017-10-01T20:51:28Z whoman is now known as hooman
2017-10-01T20:58:15Z stassats: i think i've already encounter the issue of local functions using the same stack storage base
2017-10-01T20:58:41Z stassats: as a result both parent and child use more stack than necessary
2017-10-01T21:17:05Z dougk joined #sbcl
2017-10-01T21:19:16Z angavrilov quit (Remote host closed the connection)
2017-10-01T21:20:38Z stassats: even register sbs are shared
2017-10-01T21:21:28Z dougk quit (Ping timeout: 240 seconds)
2017-10-01T21:25:29Z stassats: ok, too complicated
2017-10-01T21:38:51Z stassats: and local calls force register spilling even if the callee doesn't actually use the registers
2017-10-01T21:52:06Z stassats: i'm always stumped by regalloc
2017-10-01T22:20:16Z dougk joined #sbcl
2017-10-01T22:24:43Z attila_lendvai joined #sbcl
2017-10-01T22:41:21Z shka quit (Ping timeout: 240 seconds)
2017-10-01T22:44:23Z attila_lendvai quit (Quit: Leaving.)
2017-10-01T22:46:44Z dougk quit (Ping timeout: 258 seconds)
2017-10-01T22:53:05Z dougk joined #sbcl
2017-10-01T23:02:18Z stassats: sbs, at least unbounded, should probably be attached to ir2-physenv
2017-10-01T23:02:33Z stassats: but that also means that life-time analysis is also affected, not just pack
2017-10-01T23:10:35Z stylewarning: stassats, what controls the order or depth of inlining?
2017-10-01T23:10:53Z stassats: order?
2017-10-01T23:11:55Z stylewarning: if F is defined in a labels and G is global and declared inline, then I would hope (G #’F) would in-line F, where G just funcalls F
2017-10-01T23:12:19Z stylewarning: (Both F and G are declared inline)
2017-10-01T23:13:27Z stassats: it can't inline first-class functions
2017-10-01T23:13:57Z stylewarning: I thought I’ve seen SBCL do it. I’ll have to pull up the example I have.
2017-10-01T23:14:50Z stylewarning: If something is (FUNCALL #’F) where F is statically determined, I thought I’ve seen SBCL do OK.
2017-10-01T23:15:15Z stassats: (funcall #'f) is not exactly first class...
2017-10-01T23:18:35Z stylewarning: That’s what I’m saying. If I have (declaim (inline h g)) (defun g (x) (funcall x)) (defun h () (labels ((f () 5)) (declare (inline f)) (g #’f)))
2017-10-01T23:19:01Z stassats: you keep using ’
2017-10-01T23:19:50Z stassats: it doesn't inline here cause it's in a different component
2017-10-01T23:20:07Z stassats: (defun h (x) (labels ((f () x)) (declare (inline f)) (g #'f))) will be inlined
2017-10-01T23:20:23Z stassats: cause closures are sharing the component with the parent
2017-10-01T23:21:59Z stylewarning: (I’m typing in a bad IRC client on my phone, hence the wrong quotes.)
2017-10-01T23:23:29Z stylewarning: stassats: are you saying that if I force sharing that it will work?
2017-10-01T23:23:38Z stassats: yes
2017-10-01T23:23:44Z stassats: until something else stops it
2017-10-01T23:24:25Z stylewarning: Hm, I’ll have to try. This kind of code is in our hot path, and I’ve been avoiding doing manual inlining since it breaks all of these nice abstractions
2017-10-01T23:25:23Z stassats: and insert dummy closed over variables is keeping it nice?
2017-10-01T23:27:44Z stylewarning: No that sounds awful and confusing
2017-10-01T23:28:03Z stassats: that's a perfect description for the sbcl compiler
2017-10-01T23:29:48Z stylewarning: But it sounds better than trying to refactor N functions (like G) usable within M contexts (like H) into this manually unlined gobbledygook
2017-10-01T23:30:01Z stylewarning: inlined *
2017-10-01T23:33:28Z dougk quit (Ping timeout: 240 seconds)
2017-10-01T23:35:51Z dougk joined #sbcl
2017-10-02T01:05:25Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-02T01:05:58Z DeadTrickster_ joined #sbcl
2017-10-02T01:22:37Z Bicyclidine joined #sbcl
2017-10-02T01:22:45Z stassats quit (Ping timeout: 258 seconds)
2017-10-02T01:23:10Z Bike quit (Ping timeout: 255 seconds)
2017-10-02T01:57:49Z dougk quit (Ping timeout: 248 seconds)
2017-10-02T03:20:24Z Bicyclidine quit (Ping timeout: 258 seconds)
2017-10-02T04:34:13Z dougk joined #sbcl
2017-10-02T04:47:11Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2017-10-02T04:53:49Z shka joined #sbcl
2017-10-02T05:27:09Z PuercoPop joined #sbcl
2017-10-02T05:27:33Z PuercoPop is now known as Guest43159
2017-10-02T06:24:33Z dougk quit (Ping timeout: 248 seconds)
2017-10-02T06:33:13Z scymtym quit (Ping timeout: 258 seconds)
2017-10-02T06:36:31Z angavrilov joined #sbcl
2017-10-02T06:44:56Z dougk_ quit (Ping timeout: 246 seconds)
2017-10-02T06:58:08Z dougk_ joined #sbcl
2017-10-02T07:21:31Z shka quit (Ping timeout: 258 seconds)
2017-10-02T07:47:21Z scymtym joined #sbcl
2017-10-02T08:21:37Z dougk joined #sbcl
2017-10-02T08:25:55Z dougk quit (Ping timeout: 258 seconds)
2017-10-02T09:22:37Z dougk joined #sbcl
2017-10-02T09:26:41Z dougk quit (Ping timeout: 240 seconds)
2017-10-02T10:23:48Z dougk joined #sbcl
2017-10-02T10:28:17Z dougk quit (Ping timeout: 248 seconds)
2017-10-02T10:33:48Z m00natic joined #sbcl
2017-10-02T11:24:02Z Bike joined #sbcl
2017-10-02T12:00:11Z rpg joined #sbcl
2017-10-02T12:02:33Z dougk joined #sbcl
2017-10-02T12:04:59Z panji joined #sbcl
2017-10-02T12:13:14Z foom2 joined #sbcl
2017-10-02T12:16:20Z foom quit (Ping timeout: 255 seconds)
2017-10-02T12:22:06Z sebastien_ joined #sbcl
2017-10-02T12:25:47Z Bike quit (Ping timeout: 255 seconds)
2017-10-02T12:47:29Z panji quit (Ping timeout: 248 seconds)
2017-10-02T12:59:51Z Bike joined #sbcl
2017-10-02T13:00:10Z panji joined #sbcl
2017-10-02T13:29:36Z dougk quit (Ping timeout: 240 seconds)
2017-10-02T13:29:37Z sjl_ quit (Ping timeout: 248 seconds)
2017-10-02T13:46:22Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2017-10-02T13:49:35Z cromachina quit (Read error: Connection reset by peer)
2017-10-02T14:27:58Z pkhuong: stylewarning: you might want to consider a real code generator and wrap that in a macro.
2017-10-02T15:05:36Z dougk joined #sbcl
2017-10-02T15:40:51Z Jesin quit (Quit: Leaving)
2017-10-02T15:43:24Z Jesin joined #sbcl
2017-10-02T15:46:38Z rpg joined #sbcl
2017-10-02T16:22:57Z panji quit (Ping timeout: 260 seconds)
2017-10-02T16:30:10Z Jesin quit (Quit: Leaving)
2017-10-02T16:33:40Z Jesin joined #sbcl
2017-10-02T16:35:49Z panji joined #sbcl
2017-10-02T16:53:27Z leah2 quit (Ping timeout: 258 seconds)
2017-10-02T17:00:44Z milanj joined #sbcl
2017-10-02T17:09:10Z leah2 joined #sbcl
2017-10-02T17:11:17Z shka_ joined #sbcl
2017-10-02T17:15:13Z leah2 quit (Ping timeout: 248 seconds)
2017-10-02T17:17:13Z slyrus joined #sbcl
2017-10-02T17:28:28Z stassats joined #sbcl
2017-10-02T17:29:24Z stassats: huh, so there's a complex-vector-widetag floating around with broken %array-data
2017-10-02T17:30:10Z stassats: and there's like 10 (vector nil 20), what's up with that?
2017-10-02T17:30:13Z stassats: they aren't broken
2017-10-02T17:30:42Z leah2 joined #sbcl
2017-10-02T17:33:22Z stassats: ok, they are all broken
2017-10-02T17:34:27Z stassats: what's in array-data looks 32-bit and starts with #xF
2017-10-02T17:37:07Z stassats: https://bugs.launchpad.net/sbcl/+bug/1720763/comments/9 is the way to get to it
2017-10-02T17:38:35Z stassats: 32-bit on macos, it's not 32-bit on linux, so that theory is out
2017-10-02T17:41:00Z stassats: and all the bogus vectors have the same bogus thing for array-data
2017-10-02T17:41:01Z leah2 quit (Ping timeout: 240 seconds)
2017-10-02T17:42:43Z stassats: ok, it's a stack address, explains the difference between linux and mac
2017-10-02T17:47:16Z stassats: NIL is actually not the type, but the initial element in array-element-type
2017-10-02T17:47:56Z stassats: changing that to :invalid to reduce future confusion
2017-10-02T17:54:19Z stassats: so, the (sb-vm:print-allocated-objects :smoke) test is just broken, you can't print random stuff and expect it to work
2017-10-02T17:57:12Z leah2 joined #sbcl
2017-10-02T17:58:40Z stassats: scymtym: i'm just going to remove room.pure.lisp?
2017-10-02T17:59:44Z stassats: i'm eager to snip PRINT-ALLOCATED-OBJECTS as well, it's just never going to work
2017-10-02T18:02:06Z scymtym: stassats: go ahead with the test. i have no opinion regarding PRINT-ALLOCATED-OBJECTS
2017-10-02T18:05:31Z stassats: well, nobody uses print-allocated-objects, so i guess no harm in doing anything
2017-10-02T18:05:35Z stassats: either removing or not removing
2017-10-02T18:07:09Z stassats: type-of now says (VECTOR :INVALID 20)
2017-10-02T18:07:23Z stassats: (could signal an error, but that was an easier thing to do)
2017-10-02T18:10:25Z m00natic quit (Remote host closed the connection)
2017-10-02T18:22:05Z stassats: but i guess type-of not signalling an error is a good thing
2017-10-02T18:22:49Z stassats: ok, to the next thing, another pfdietz test
2017-10-02T18:26:36Z panji quit (Ping timeout: 240 seconds)
2017-10-02T18:30:54Z stassats: just a NIL typed lvar getting past csubtypep, that was easy
2017-10-02T18:34:48Z leah2 quit (Ping timeout: 240 seconds)
2017-10-02T19:06:11Z stassats: oops, (vector :invalid) thing broke a lot of tests
2017-10-02T19:16:39Z dougk quit (Ping timeout: 246 seconds)
2017-10-02T19:18:07Z dougk joined #sbcl
2017-10-02T19:23:08Z dougk__ joined #sbcl
2017-10-02T19:24:29Z dougk quit (Ping timeout: 258 seconds)
2017-10-02T19:26:53Z scymtym quit (Ping timeout: 246 seconds)
2017-10-02T19:31:34Z ym quit (Ping timeout: 264 seconds)
2017-10-02T19:43:40Z attila_lendvai joined #sbcl
2017-10-02T19:48:05Z shka_ quit (Ping timeout: 240 seconds)
2017-10-02T19:49:28Z stassats: now i wonder, what's creating complex vectors with dx data arrays?
2017-10-02T20:00:28Z stassats: list-referencing-objects is silent
2017-10-02T20:05:58Z Jesin quit (Quit: Leaving)
2017-10-02T20:11:11Z Jesin joined #sbcl
2017-10-02T20:18:22Z scymtym joined #sbcl
2017-10-02T20:33:06Z dougk__ quit (Ping timeout: 258 seconds)
2017-10-02T20:35:55Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2017-10-02T20:40:29Z Jesin quit (Quit: Leaving)
2017-10-02T21:10:29Z rpg joined #sbcl
2017-10-02T21:15:06Z dougk joined #sbcl
2017-10-02T21:28:37Z Bike quit (Ping timeout: 260 seconds)
2017-10-02T21:40:30Z Jesin joined #sbcl
2017-10-02T21:42:10Z Jesin quit (Remote host closed the connection)
2017-10-02T21:42:26Z dougk quit (Ping timeout: 255 seconds)
2017-10-02T21:49:51Z Jesin joined #sbcl
2017-10-02T21:58:35Z ym joined #sbcl
2017-10-02T22:08:30Z Bike joined #sbcl
2017-10-02T22:15:37Z attila_lendvai quit (Quit: Leaving.)
2017-10-02T22:16:54Z sjl_ joined #sbcl
2017-10-02T22:24:27Z dougk joined #sbcl
2017-10-02T22:32:25Z stassats: how can you even get a complex vector with dx data?
2017-10-02T22:34:33Z stassats: ok, i guess displacing would work
2017-10-02T22:36:13Z stassats: found it, it's from SB-C::CDF-ENCODE-LOCS
2017-10-02T22:40:06Z sjl_ quit (Ping timeout: 240 seconds)
2017-10-02T22:49:42Z dougk quit (Ping timeout: 260 seconds)
2017-10-02T23:01:35Z sjl_ joined #sbcl
2017-10-02T23:08:49Z sjl_ quit (Ping timeout: 248 seconds)
2017-10-02T23:09:58Z Bike quit (Remote host closed the connection)
2017-10-02T23:10:19Z Bike joined #sbcl
2017-10-02T23:21:17Z Xach left #sbcl
2017-10-02T23:27:18Z sjl_ joined #sbcl
2017-10-02T23:28:04Z cromachina joined #sbcl
2017-10-02T23:33:06Z stassats: that motivated me to add support for stack allocating vectors with a fill-pointer
2017-10-02T23:49:28Z dougk joined #sbcl
2017-10-02T23:59:05Z stassats: which was surprisingly trivial
2017-10-03T00:09:36Z Posterdati quit (Ping timeout: 240 seconds)
2017-10-03T00:10:10Z Posterdati joined #sbcl
2017-10-03T00:27:42Z leah2 joined #sbcl
2017-10-03T00:45:20Z leah2 quit (Ping timeout: 248 seconds)
2017-10-03T00:52:06Z dougk: stassats: i can't believe i didn't remember that I did that
2017-10-03T00:55:53Z sjl_ is now known as sjl
2017-10-03T01:11:54Z stassats quit (Ping timeout: 246 seconds)
2017-10-03T01:24:27Z leah2 joined #sbcl
2017-10-03T01:24:34Z milanj quit (Quit: This computer has gone to sleep)
2017-10-03T01:44:01Z dougk quit (Ping timeout: 248 seconds)
2017-10-03T02:20:01Z leah2 quit (Ping timeout: 258 seconds)
2017-10-03T02:36:28Z leah2 joined #sbcl
2017-10-03T02:44:13Z Bicyclidine joined #sbcl
2017-10-03T02:46:51Z Bike quit (Ping timeout: 258 seconds)
2017-10-03T02:47:54Z dougk joined #sbcl
2017-10-03T02:50:36Z Bicyclidine quit (Ping timeout: 246 seconds)
2017-10-03T02:54:05Z leah2 quit (Ping timeout: 240 seconds)
2017-10-03T03:07:33Z dougk quit (Ping timeout: 258 seconds)
2017-10-03T03:31:45Z sjl quit (Ping timeout: 248 seconds)
2017-10-03T03:42:56Z leah2 joined #sbcl
2017-10-03T03:47:29Z sjl joined #sbcl
2017-10-03T03:58:25Z leah2 quit (Ping timeout: 248 seconds)
2017-10-03T04:15:55Z leah2 joined #sbcl
2017-10-03T04:22:32Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2017-10-03T04:26:30Z leah2 quit (Ping timeout: 246 seconds)
2017-10-03T04:43:49Z leah2 joined #sbcl
2017-10-03T05:13:15Z shka_ joined #sbcl
2017-10-03T05:33:41Z eschatologist quit (Ping timeout: 240 seconds)
2017-10-03T05:47:35Z leah2 quit (Ping timeout: 240 seconds)
2017-10-03T06:01:18Z milanj joined #sbcl
2017-10-03T06:15:38Z sebastien_ left #sbcl
2017-10-03T06:16:14Z leah2 joined #sbcl
2017-10-03T06:23:07Z salva joined #sbcl
2017-10-03T06:24:58Z leah2 quit (Ping timeout: 258 seconds)
2017-10-03T06:40:48Z leah2 joined #sbcl
2017-10-03T07:08:49Z shka_ quit (Ping timeout: 248 seconds)
2017-10-03T07:12:31Z leah2 quit (Ping timeout: 248 seconds)
2017-10-03T07:27:49Z leah2 joined #sbcl
2017-10-03T07:50:25Z leah2 quit (Ping timeout: 248 seconds)
2017-10-03T08:06:01Z leah2 joined #sbcl
2017-10-03T08:16:27Z leah2 quit (Ping timeout: 246 seconds)
2017-10-03T08:55:28Z milanj quit (Quit: This computer has gone to sleep)
2017-10-03T08:58:41Z scymtym quit (Ping timeout: 248 seconds)
2017-10-03T09:17:51Z ym quit (Ping timeout: 248 seconds)
2017-10-03T09:19:09Z ym joined #sbcl
2017-10-03T09:28:13Z scymtym joined #sbcl
2017-10-03T09:45:34Z panji joined #sbcl
2017-10-03T10:00:49Z attila_lendvai joined #sbcl
2017-10-03T10:00:49Z attila_lendvai quit (Changing host)
2017-10-03T10:00:49Z attila_lendvai joined #sbcl
2017-10-03T10:37:45Z m00natic joined #sbcl
2017-10-03T10:38:28Z stassats joined #sbcl
2017-10-03T10:49:03Z leah2 joined #sbcl
2017-10-03T10:54:50Z Bike joined #sbcl
2017-10-03T10:58:21Z leah2 quit (Ping timeout: 240 seconds)
2017-10-03T11:15:59Z scymtym: i would like to clean up one of the issues found by the ifdef checker. in interrupt.c line 282, the #if really is superfluous since the #if in line 217 tested the exact same condition, right?
2017-10-03T11:16:52Z stassats: right
2017-10-03T11:17:27Z stassats: and next  #endif #if !defined(LISP_FEATURE_WIN32) || defined(LISP_FEATURE_SB_THREAD) can be squashed too
2017-10-03T11:18:21Z stassats: but it's all really bad looking
2017-10-03T11:18:35Z scymtym: i'm a bit disappointed, that it found such a boring case
2017-10-03T11:19:34Z stassats: working on (the (function (string) (values string &optional)) #'evenp) => warning: The function EVENP conflicts with its assert type (FUNCTION (STRING) (VALUES STRING &OPTIONAL))
2017-10-03T11:22:06Z attila_lendvai quit (Quit: Leaving.)
2017-10-03T11:22:25Z attila_lendvai joined #sbcl
2017-10-03T11:24:17Z scymtym: nice
2017-10-03T11:26:08Z scymtym: the interesting cases are basically style issues. for example in pseudo-atomic.h line 32, the condition tested by the #elif is implied by the context. so it could be just #else, but i'm not sure which one is better
2017-10-03T11:26:46Z stassats: can emacs better deal with #ifdef?
2017-10-03T11:26:57Z stassats: or is it just sbcl that has so many ifdefs that nobody else cares?
2017-10-03T11:27:20Z stassats: the support (or lack thereof) for #ifdef in c-mode is appalling
2017-10-03T11:28:07Z scymtym: what features would you like to have? folding?
2017-10-03T11:28:23Z stassats: matching, at least
2017-10-03T11:28:41Z angavrilov quit (Ping timeout: 240 seconds)
2017-10-03T11:28:41Z scymtym: that can be done via flycheck
2017-10-03T11:29:20Z stassats: there's cpp-highlight-buffer
2017-10-03T11:29:53Z scymtym: interesting
2017-10-03T11:30:06Z stassats: and C-c C-p runs the command c-backward-conditional
2017-10-03T11:30:16Z stassats: but i'm not sure it how it works
2017-10-03T11:32:10Z scymtym: seems to work when point is after an #endif
2017-10-03T11:32:39Z stassats: doesn't work well with #else
2017-10-03T11:33:41Z stassats: ok, there are some commands
2017-10-03T11:33:47Z stassats: i concede
2017-10-03T11:33:57Z stassats: but highlighting would have been nice
2017-10-03T11:35:10Z stassats: c-forward-conditional c-backward-conditional c-up-conditional
2017-10-03T11:35:16Z stassats: but neither works with #else
2017-10-03T11:35:59Z scymtym: are you extending the MAP transform, DERIVE-CALL-TYPE for MAP, or the CALLABLE mechanism?
2017-10-03T11:36:46Z stassats: scymtym: first, functional casts checking, maybe it'll be enough, maybe not
2017-10-03T11:37:14Z stassats: just as a standalone feature first
2017-10-03T11:37:22Z stassats: it's been in my todo for a long time
2017-10-03T11:38:52Z scymtym: yeah, that's a good thing to have in general
2017-10-03T11:41:11Z stassats: and handle (lambda) types better
2017-10-03T11:41:17Z stassats: currently (find-if #'evenp "abc") conflicts
2017-10-03T11:41:36Z stassats: but not (find-if (lambda (x) (declare (integer x)) x) "abc")
2017-10-03T12:00:19Z rpg joined #sbcl
2017-10-03T12:08:03Z stassats: having some trouble with #'ERROR, ugh
2017-10-03T12:08:56Z Bike quit (Ping timeout: 246 seconds)
2017-10-03T12:11:51Z stassats: ok, it's a legitimate mismatch, yay
2017-10-03T12:13:17Z stassats: if it already caught an error in sbcl, how many quicklisp projects will fall?
2017-10-03T12:16:00Z dougk joined #sbcl
2017-10-03T12:16:08Z BitPuffin|osx joined #sbcl
2017-10-03T12:21:57Z attila_lendvai quit (Quit: Leaving.)
2017-10-03T12:30:07Z Bike joined #sbcl
2017-10-03T12:30:11Z Bike quit (Remote host closed the connection)
2017-10-03T12:32:48Z Bike joined #sbcl
2017-10-03T12:34:23Z stassats: (the (function (string) (values string &optional)) (lambda (x) 1)) now works, but not (find-if (lambda (x) (declare (integer x)) x) "abc")
2017-10-03T12:35:12Z stassats: cause it's using the old callable checker, which is still producing nicer error messages, so i guess it should just assert the type and wait for the new thingy to pick it up
2017-10-03T12:39:18Z Bike quit (Ping timeout: 246 seconds)
2017-10-03T12:50:08Z milanj joined #sbcl
2017-10-03T13:17:53Z sjl quit (Ping timeout: 248 seconds)
2017-10-03T13:23:39Z leah2 joined #sbcl
2017-10-03T13:29:05Z leah2 quit (Ping timeout: 248 seconds)
2017-10-03T13:39:14Z milanj quit (Read error: Connection reset by peer)
2017-10-03T13:47:35Z cromachina quit (Read error: Connection reset by peer)
2017-10-03T13:54:35Z panji quit (Ping timeout: 240 seconds)
2017-10-03T14:06:08Z dougk quit (Ping timeout: 255 seconds)
2017-10-03T14:08:15Z panji joined #sbcl
2017-10-03T14:15:17Z dougk joined #sbcl
2017-10-03T14:28:47Z stassats: next order: make (map-into (make-string 10) (lambda (x) (declare (fixnum x)) (code-char x)) (the simple-base-string abc)) warn
2017-10-03T14:33:18Z Jesin quit (Quit: Leaving)
2017-10-03T14:52:54Z stassats: The function (LAMBDA (X) :IN FOO) called by MAP-INTO with the elements of the third argument of type (SIMPLE-ARRAY CHARACTER (3)) which conflict with FIXNUM.
2017-10-03T14:53:12Z stassats: too bad it doesn't work with (lambda (x) (code-char x))
2017-10-03T14:53:44Z stassats: maybe we should really derive the type in that case
2017-10-03T15:06:02Z dougk quit (Ping timeout: 246 seconds)
2017-10-03T15:07:17Z angavrilov joined #sbcl
2017-10-03T15:13:15Z Jesin joined #sbcl
2017-10-03T15:13:52Z Bike joined #sbcl
2017-10-03T15:13:54Z Bike quit (Remote host closed the connection)
2017-10-03T15:18:42Z stassats: but MAP transforms things quite a bit
2017-10-03T15:28:30Z panji quit (Quit: Leaving)
2017-10-03T15:39:54Z Jesin quit (Quit: Leaving)
2017-10-03T15:42:52Z stassats: (let ((x (lambda (x) (declare (fixnum x)) x))) (funcall x 'a)) doesn't fail
2017-10-03T15:42:57Z stassats: so many prerequisites
2017-10-03T15:45:24Z milanj joined #sbcl
2017-10-03T15:53:17Z dougk joined #sbcl
2017-10-03T16:02:27Z scymtym: seq.pure.lisp line 178. any idea what that was meant to do? my best guess is that it just wasn't noticeable in all the noise
2017-10-03T16:03:48Z stassats: yeah, weird
2017-10-03T16:04:45Z scymtym: that whole file is such a mess
2017-10-03T16:09:42Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2017-10-03T16:16:30Z stassats: managed (let ((x (lambda (x) (declare (fixnum x)) x))) (funcall x 'a)) to conflict
2017-10-03T16:16:30Z stassats:  
2017-10-03T16:17:03Z BitPuffin|osx quit (Ping timeout: 248 seconds)
2017-10-03T16:17:43Z stassats: should help with map transforms
2017-10-03T16:21:08Z scymtym quit (Ping timeout: 255 seconds)
2017-10-03T16:30:44Z rpg joined #sbcl
2017-10-03T16:36:13Z scymtym joined #sbcl
2017-10-03T16:47:21Z dougk quit (Ping timeout: 240 seconds)
2017-10-03T16:55:36Z shka_ joined #sbcl
2017-10-03T17:11:56Z sebboh left #sbcl
2017-10-03T17:23:03Z m00natic quit (Remote host closed the connection)
2017-10-03T17:28:40Z Bike joined #sbcl
2017-10-03T17:31:26Z leah2 joined #sbcl
2017-10-03T17:55:05Z leah2 quit (Ping timeout: 240 seconds)
2017-10-03T18:12:31Z milanj quit (Quit: This computer has gone to sleep)
2017-10-03T18:14:37Z leah2 joined #sbcl
2017-10-03T18:26:59Z stassats: and i have the bug-report test case issuing a type conflict
2017-10-03T18:27:30Z stassats: albeit with a bogus source location
2017-10-03T18:29:32Z Jesin joined #sbcl
2017-10-03T18:39:17Z leah2 quit (Ping timeout: 255 seconds)
2017-10-03T18:46:49Z stassats: figured that out, but fully expanded map transforms produce really confusing contexts
2017-10-03T18:47:06Z stassats: but it's a start, and i'll ponder on that in the future, time to close the bug
2017-10-03T18:52:53Z milanj joined #sbcl
2017-10-03T18:53:01Z milanj quit (Client Quit)
2017-10-03T18:56:25Z leah2 joined #sbcl
2017-10-03T19:03:57Z shka_ quit (Ping timeout: 246 seconds)
2017-10-03T19:07:06Z dougk joined #sbcl
2017-10-03T19:14:44Z eschatologist joined #sbcl
2017-10-03T19:31:45Z milanj joined #sbcl
2017-10-03T19:31:59Z phoe joined #sbcl
2017-10-03T19:32:03Z phoe: stassats: that was quick, thanks!
2017-10-03T19:32:10Z milanj quit (Client Quit)
2017-10-03T19:36:47Z milanj joined #sbcl
2017-10-03T19:37:11Z milanj quit (Client Quit)
2017-10-03T19:46:12Z stassats quit (Ping timeout: 260 seconds)
2017-10-03T19:48:51Z Jesin quit (Quit: Leaving)
2017-10-03T20:01:12Z angavrilov quit (Remote host closed the connection)
2017-10-03T20:09:05Z rpg quit (Ping timeout: 248 seconds)
2017-10-03T20:11:29Z attila_lendvai joined #sbcl
2017-10-03T20:15:35Z dougk quit (Ping timeout: 255 seconds)
2017-10-03T20:38:38Z dougk joined #sbcl
2017-10-03T20:41:01Z rpg joined #sbcl
2017-10-03T20:54:54Z Jesin joined #sbcl
2017-10-03T21:10:05Z Bike quit (Ping timeout: 240 seconds)
2017-10-03T21:15:00Z dougk_: stassats: I'm getting a build failure after one your changes, not sure which yet. Try building with #+sb-fasteval #-sb-eval. It gives an error that seems untrue
2017-10-03T21:15:23Z dougk_: "Derived type of #'PROTO-FUNCTIONIZE is {mumble} conflicting with {othermumble}"
2017-10-03T21:15:57Z dougk_: considering that I never explicitly asserted the type anywhere
2017-10-03T21:26:17Z sjl joined #sbcl
2017-10-03T21:26:59Z Xof: do we build with *derive-types* on?
2017-10-03T21:29:49Z dougk_: we do
2017-10-03T21:29:51Z sjl quit (Client Quit)
2017-10-03T21:30:11Z dougk_: in order to not have it on - which I'd like - I need to add about 5000 FTYPE proclamations
2017-10-03T21:32:18Z rpg quit (Read error: Connection reset by peer)
2017-10-03T21:32:44Z rpg joined #sbcl
2017-10-03T21:34:47Z sjl joined #sbcl
2017-10-03T21:43:21Z Bike joined #sbcl
2017-10-03T21:45:11Z Bicyclidine joined #sbcl
2017-10-03T21:45:14Z Bicyclidine quit (Remote host closed the connection)
2017-10-03T21:45:36Z Bicyclidine joined #sbcl
2017-10-03T21:47:05Z Gorgias joined #sbcl
2017-10-03T21:47:07Z Gorgias quit (Remote host closed the connection)
2017-10-03T21:47:25Z Gorgias joined #sbcl
2017-10-03T21:47:45Z Bike quit (Ping timeout: 248 seconds)
2017-10-03T21:50:05Z Bicyclidine quit (Ping timeout: 255 seconds)
2017-10-03T22:18:34Z Bike joined #sbcl
2017-10-03T22:21:46Z Gorgias quit (Ping timeout: 258 seconds)
2017-10-03T22:41:37Z stassats joined #sbcl
2017-10-03T22:43:53Z stassats: dougk: oops
2017-10-03T22:43:56Z stassats: some editing snafu
2017-10-03T22:44:33Z stassats: phoe: most of the stuff was already in place
2017-10-03T22:45:32Z dougk quit (Ping timeout: 258 seconds)
2017-10-03T23:00:47Z milanj joined #sbcl
2017-10-03T23:06:17Z cromachina joined #sbcl
2017-10-03T23:29:16Z Bicyclidine joined #sbcl
2017-10-03T23:29:59Z Bike quit (Ping timeout: 255 seconds)
2017-10-03T23:35:48Z stassats: (defun foo (x) (map-into (make-string 10) #'evenp x)) => "Derived type of THING is"
2017-10-03T23:35:52Z stassats: ugh, THING
2017-10-03T23:35:56Z stassats: from %coerce-callable-to-fun
2017-10-03T23:36:14Z stassats: transforms make error report hard, especially define-source-transform
2017-10-03T23:37:25Z dougk_: just be glad we don't name our local variables SHIAT, i guess?
2017-10-03T23:45:00Z stassats: just stuff the leaf name instead of mucking about with %coerce-callable-to-fun
2017-10-03T23:58:36Z dougk joined #sbcl
2017-10-04T00:01:30Z stassats: ok, i think i'm finally done for the day
2017-10-04T00:01:50Z stassats: but there's no end to it, better reporting, more through checking, etc
2017-10-04T00:08:54Z stassats: dougk_: mmap MAP_FIXED on darwin is translated to vm_map VM_FLAGS_OVERWRITE and VM_FLAGS_FIXED
2017-10-04T00:09:01Z stassats: so vm_map could be used with just VM_FLAGS_FIXED
2017-10-04T00:12:02Z dougk quit (Ping timeout: 246 seconds)
2017-10-04T00:17:21Z dougk joined #sbcl
2017-10-04T00:17:38Z dougk_: i don't think i'm passing MAP_FIXED for darwin. It always uses 'addr' as a hint.
2017-10-04T00:18:00Z dougk_: It's FreeBSD and others that *don't* assume that you want the specified 'addr' to have relevance if MAP_FIXED isn't given
2017-10-04T00:18:56Z dougk_: afaict, I'm supposed to use MAP_TRYFIXED to "prefer" an address. I just tried installing DragonFly under qemu and it didn't work, so, .... I don't care. Let someone tell me what to do who actually has {Net,Open,Dragonfly}BSD.
2017-10-04T00:18:58Z stassats: but os_validate is the same for all bsds
2017-10-04T00:20:05Z dougk_: right, ok. will revisit this for the Nth time
2017-10-04T00:20:18Z stassats: not that i'm having any problems
2017-10-04T00:20:35Z stassats: but i remember map_fixed being a fickle thing to diagnose
2017-10-04T00:21:14Z dougk_: indeed. that's why I added the massive comment. To get back to it.
2017-10-04T00:22:04Z stassats: was just looking through xnu to see if there's MAP_32BIT
2017-10-04T00:24:06Z stassats: and darwin seems to actually work without MAP_FIXED
2017-10-04T00:26:40Z stassats: well, it used to work before, so
2017-10-04T00:27:29Z stassats: oh, it was always MAP_FIXED
2017-10-04T00:28:01Z dougk_: yes
2017-10-04T00:28:19Z stassats: well, it works without, at least now and least for me
2017-10-04T00:30:34Z dougk quit (Ping timeout: 258 seconds)
2017-10-04T00:31:06Z stassats: and it actually places where requested, not just low addresses
2017-10-04T00:32:11Z stassats: i need to update my xnu sources
2017-10-04T00:32:52Z stassats: (to the new fancy git repo (not really fancy))
2017-10-04T00:43:08Z stassats: i think i'm for removing MAP_FIXED for darwin
2017-10-04T00:48:50Z dougk_: i agree
2017-10-04T00:48:57Z dougk_: i'll do that
2017-10-04T00:49:41Z dougk_: did you see my question about overalignment in alloc_tramp? "and $-32,%rsp" should be "and $-16,%rsp" ?
2017-10-04T00:50:54Z dougk_: because after the next 10 pushes, rsp is on a 16-byte boundary anyway, not a 32-byte boundary.
2017-10-04T01:00:05Z dougk joined #sbcl
2017-10-04T01:01:52Z stassats: dougk: yes, it is overaligned
2017-10-04T01:02:05Z dougk_: no reason for that, right ?
2017-10-04T01:02:31Z dougk_: more like someone just forgot how 2's-complement works?
2017-10-04T01:04:10Z stassats: can't think of any reason
2017-10-04T01:05:40Z stassats: movaps only needs 16 bytes of alignment
2017-10-04T01:17:05Z attila_lendvai quit (Quit: Leaving.)
2017-10-04T01:20:32Z stassats: what does /* FIXME Restore the NPX state. */ even mean?
2017-10-04T01:20:59Z stassats: x87
2017-10-04T01:21:48Z stassats: i don't think that's relevant anymore
2017-10-04T01:30:59Z rpg quit (Quit: Textual IRC Client: www.textualapp.com)
2017-10-04T01:32:42Z stassats: although x86 CW is callee saved
2017-10-04T01:32:46Z stassats: but who cares, really
2017-10-04T01:38:35Z stassats quit (Ping timeout: 240 seconds)
2017-10-04T02:01:09Z dougk quit (Ping timeout: 246 seconds)
2017-10-04T02:15:07Z Bike joined #sbcl
2017-10-04T02:16:21Z sjl quit (Quit: WeeChat 1.3)
2017-10-04T02:18:17Z Bicyclidine quit (Ping timeout: 258 seconds)
2017-10-04T02:26:49Z karswell_ is now known as karswell
2017-10-04T03:11:35Z Bike quit (Ping timeout: 240 seconds)
2017-10-04T03:25:17Z dougk joined #sbcl
2017-10-04T03:25:42Z milanj quit (Quit: This computer has gone to sleep)
2017-10-04T03:30:12Z dougk quit (Read error: Connection reset by peer)
2017-10-04T03:30:31Z dougk joined #sbcl
2017-10-04T03:54:41Z stylewarning: This code fails on the second call to (FAIL 0): http://paste.lisp.org/display/357602
2017-10-04T03:54:56Z stylewarning: Is it because the code is bad or because SBCL doesn't know how to handle it?
2017-10-04T03:58:36Z pkhuong: the code is bad.
2017-10-04T03:58:44Z stylewarning: :[
2017-10-04T04:46:51Z dougk quit (Read error: Connection reset by peer)
2017-10-04T04:47:07Z dougk joined #sbcl
2017-10-04T05:09:33Z shka_ joined #sbcl
2017-10-04T05:28:01Z dougk quit (Ping timeout: 240 seconds)
2017-10-04T06:04:16Z scymtym quit (Ping timeout: 248 seconds)
2017-10-04T07:16:13Z milanj joined #sbcl
2017-10-04T07:16:48Z shka_ quit (Ping timeout: 240 seconds)
2017-10-04T07:18:02Z carenz_ joined #sbcl
2017-10-04T07:27:52Z scymtym joined #sbcl
2017-10-04T08:05:52Z angavrilov joined #sbcl
2017-10-04T08:32:11Z pchrist quit (Quit: leaving)
2017-10-04T08:32:56Z pchrist joined #sbcl
2017-10-04T08:49:52Z m00natic joined #sbcl
2017-10-04T10:06:46Z attila_lendvai joined #sbcl
2017-10-04T10:06:46Z attila_lendvai quit (Changing host)
2017-10-04T10:06:46Z attila_lendvai joined #sbcl
2017-10-04T10:14:19Z Bike joined #sbcl
2017-10-04T10:17:48Z carenz_ quit (Ping timeout: 246 seconds)
2017-10-04T10:23:28Z DeadTrickster_ quit (Remote host closed the connection)
2017-10-04T10:26:35Z DeadTrickster joined #sbcl
2017-10-04T11:14:07Z Bicyclidine joined #sbcl
2017-10-04T11:15:45Z Bike quit (Ping timeout: 248 seconds)
2017-10-04T11:24:09Z Xof quit (Ping timeout: 258 seconds)
2017-10-04T11:47:14Z dougk joined #sbcl
2017-10-04T12:03:04Z scymtym_ joined #sbcl
2017-10-04T12:07:20Z scymtym quit (Ping timeout: 255 seconds)
2017-10-04T12:33:26Z Bicyclidine quit (Ping timeout: 255 seconds)
2017-10-04T12:42:41Z dougk quit (Ping timeout: 248 seconds)
2017-10-04T12:52:31Z dougk joined #sbcl
2017-10-04T13:00:19Z milanj quit (Quit: This computer has gone to sleep)
2017-10-04T13:02:11Z Bike joined #sbcl
2017-10-04T13:34:56Z dougk quit (Ping timeout: 248 seconds)
2017-10-04T13:34:56Z milanj joined #sbcl
2017-10-04T13:42:29Z stassats joined #sbcl
2017-10-04T13:43:11Z stassats: i think i need some better mechanism of expression casts that depend on some other lvars
2017-10-04T13:43:25Z stassats: like cast this to the element type of this array lvar
2017-10-04T13:44:20Z stassats: but first, type derivation across transforms really should be fixed
2017-10-04T14:05:28Z stassats: which is the problem with (find 10 (the vector z)), FIND derives 10, but then it gets transformed into %find-position and loses that, but say it doesn't lose it
2017-10-04T14:05:54Z stassats: then if 10 is not known at the time the FIND deriver runs, it won't get picked up again
2017-10-04T14:06:58Z stassats: so %find-position should be deriving, but that way we get an explosion of derivers
2017-10-04T14:07:48Z stassats: maybe if transforms leave a corps of the original combination behind to aid type derivation? and error reporting
2017-10-04T14:09:08Z scymtym__ joined #sbcl
2017-10-04T14:09:49Z stassats: but, transforming find to %find-position doesn't aid any optimization stages, maybe it should be delayed until the last moment
2017-10-04T14:10:04Z cromachina quit (Read error: Connection reset by peer)
2017-10-04T14:10:42Z stassats: that's true for many transforms, like map, the call to map already has all the info to derive and cast types
2017-10-04T14:11:38Z stassats: i think that's a better idea than leaving corpses, just make them live longer
2017-10-04T14:13:00Z scymtym_ quit (Ping timeout: 246 seconds)
2017-10-04T14:18:06Z stassats` joined #sbcl
2017-10-04T14:18:34Z stassats quit (Ping timeout: 240 seconds)
2017-10-04T14:34:21Z stassats` quit (Ping timeout: 246 seconds)
2017-10-04T14:47:34Z dougk joined #sbcl
2017-10-04T14:51:55Z dougk quit (Ping timeout: 258 seconds)
2017-10-04T14:58:01Z oleo quit (Quit: Leaving)
2017-10-04T15:43:07Z edgar-rft quit (Quit: edgar-rft)
2017-10-04T15:43:10Z Jesin quit (Quit: Leaving)
2017-10-04T15:44:10Z dougk joined #sbcl
2017-10-04T15:49:48Z Jesin joined #sbcl
2017-10-04T15:50:32Z dougk quit (Ping timeout: 255 seconds)
2017-10-04T16:01:19Z stassats` joined #sbcl
2017-10-04T16:25:16Z stassats joined #sbcl
2017-10-04T16:29:17Z m00natic quit (Remote host closed the connection)
2017-10-04T16:54:45Z shka_ joined #sbcl
2017-10-04T16:58:34Z stassats` quit (Ping timeout: 240 seconds)
2017-10-04T17:01:12Z stassats` joined #sbcl
2017-10-04T17:02:14Z dougk joined #sbcl
2017-10-04T17:09:24Z dougk quit (Ping timeout: 246 seconds)
2017-10-04T17:21:56Z dougk joined #sbcl
2017-10-04T17:29:49Z Jesin quit (Quit: Leaving)
2017-10-04T17:39:40Z dougk__ joined #sbcl
2017-10-04T17:42:35Z dougk quit (Ping timeout: 255 seconds)
2017-10-04T17:44:02Z dougk__ quit (Ping timeout: 258 seconds)
2017-10-04T18:01:05Z stassats` quit (Ping timeout: 248 seconds)
2017-10-04T18:07:01Z eschatologist quit (Ping timeout: 240 seconds)
2017-10-04T18:12:00Z eschatologist joined #sbcl
2017-10-04T18:20:58Z Jesin joined #sbcl
2017-10-04T18:33:45Z stassats` joined #sbcl
2017-10-04T18:34:53Z stassats` left #sbcl
2017-10-04T18:52:29Z carenz_ joined #sbcl
2017-10-04T19:19:04Z milanj quit (Quit: This computer has gone to sleep)
2017-10-04T19:19:32Z dougk_: scymtym: what do you want (DESCRIBE EMPTY-GF) to do for a generic-function that is not standard-generic-function? I should have asked before I broke it.
2017-10-04T19:20:03Z dougk_: but the implementation of %fun-name should not assume that it should access (slot-value f 'name). That's precisely why there is a GF named GENERIC-FUNCTION-NAME in the first place
2017-10-04T19:20:35Z carenz_ quit (Ping timeout: 240 seconds)
2017-10-04T19:21:27Z scymtym__: dougk_: sorry, i have to run now. will think and reply later
2017-10-04T19:21:47Z dougk_: ok, similar consideration applies to interface.impure, similar test
2017-10-04T19:31:51Z scymtym__ quit (Ping timeout: 246 seconds)
2017-10-04T19:32:20Z scymtym__ joined #sbcl
2017-10-04T19:37:06Z scymtym__ quit (Ping timeout: 246 seconds)
2017-10-04T19:46:39Z shka_ quit (Ping timeout: 248 seconds)
2017-10-04T20:24:38Z edgar-rft joined #sbcl
2017-10-04T20:40:07Z Ionic joined #sbcl
2017-10-04T20:41:06Z Ionic: is anyone either aware of issues building sbcl on x86_64 with GCC 6.x?
2017-10-04T20:41:14Z stassats: no
2017-10-04T20:42:17Z scymtym joined #sbcl
2017-10-04T20:42:20Z Ionic: or rather, has anyone been able to compile sbcl 1.4.0 with GCC 6.x successfully? :)
2017-10-04T20:43:26Z Ionic: so far I haven't been on a Gentoo system, which is odd. it always fails with a corruption warning and a memory fault at nil
2017-10-04T20:45:09Z Ionic: one user, albeit on x86, tracked it down to *FLAGS now being passed down to the build process, but I don't think that's spot on. even with empty values in the build environment for {C{,XX},LD}FLAGS, I get a build failure
2017-10-04T20:45:37Z Ionic: the interesting part is that retrying with GCC 5.4.0 works fine, either with my default *FLAGS values or with an empty set
2017-10-04T20:46:17Z Ionic: my current suspicion is that GCC's code generation changed in 6.x and is triggering this issue
2017-10-04T21:04:18Z stassats: sbcl builds fine
2017-10-04T21:05:03Z stassats: is it not being built without nopie again?
2017-10-04T21:05:18Z stassats: to many negatives
2017-10-04T21:05:34Z stassats: Ionic: show me one gcc invocation
2017-10-04T21:05:44Z stassats: after doing make clean; make in runtime/
2017-10-04T21:10:24Z Ionic: nothing out of the ordinary: x86_64-pc-linux-gnu-gcc -Wall -Wsign-compare -Wpointer-arith  -I. -DSBCL_PREFIX=\"'/usr'\"  -c -o funcall.o funcall.c
2017-10-04T21:11:08Z stassats: is that the complete line?
2017-10-04T21:12:06Z Ionic: yeah, one complete line
2017-10-04T21:12:28Z Ionic: and of course... the linking stage... let me look that up
2017-10-04T21:12:33Z stassats: well, everything's out of the ordinary about it
2017-10-04T21:12:42Z Ionic: x86_64-pc-linux-gnu-gcc -no-pie -o sbcl alloc.o backtrace.o breakpoint.o coreparse.o dynbind.o funcall.o gc-common.o globals.o hopscotch.o interr.o interrupt.o largefile.o monitor.o murmur_hash.o os-common.o parse.o print.o purify.o pthread-fute
2017-10-04T21:12:47Z Ionic: x.o regnames.o run-program.o runtime.o safepoint.o save.o sc-offset.o search.o thread.o time.o validate.o var-io.o vars.o wrap.o x86-64-arch.o linux-os.o x86-64-linux-os.o fullcgc.o gencgc.o traceroot.o marknsweepgc.o x86-64-assem.o ldso-stubs.o
2017-10-04T21:12:51Z Ionic:  -ldl -lpthread -lz -lm
2017-10-04T21:12:51Z stassats: cc  -g -Wall -Wsign-compare -Wpointer-arith -O3 -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -fno-omit-frame-pointer -momit-leaf-frame-pointer -fno-pie -I. -DSBCL_PREFIX=\"'/usr/local'\"  -c -o funcall.o funcall.c
2017-10-04T21:12:53Z Ionic: meh, terminal...
2017-10-04T21:12:56Z stassats: that's what the line is
2017-10-04T21:13:15Z stassats: yours is not only missing -fno-pie but a lot of other stuff
2017-10-04T21:13:45Z stassats: are the sbcl sources unadulterated?
2017-10-04T21:13:48Z jdz quit (Ping timeout: 240 seconds)
2017-10-04T21:13:57Z Ionic: hm, like the omit-leaf-frame-pointer bit
2017-10-04T21:14:28Z Ionic: no, gentoo applies some patches
2017-10-04T21:15:03Z Ionic: I'll take a look at why -fno-pie gets removed
2017-10-04T21:16:03Z Bike quit (Ping timeout: 240 seconds)
2017-10-04T21:16:35Z stassats: so, your complaints should be directed at gentoo
2017-10-04T21:16:41Z jdz joined #sbcl
2017-10-04T21:16:41Z angavrilov quit (Remote host closed the connection)
2017-10-04T21:18:13Z Ionic: probably, yes. in a clean sources set I see CFLAGS+=-fno-pie in src/runtime/GNUmakefile
2017-10-04T21:18:30Z stassats: https://gitweb.gentoo.org/repo/gentoo.git/tree/dev-lisp/sbcl/sbcl-1.4.0.ebuild#n111
2017-10-04T21:18:31Z Ionic: that's missing in the source tree of a failed gentoo build
2017-10-04T21:19:08Z stassats: it removes -O3 -g
2017-10-04T21:19:11Z stassats: what's up with that?
2017-10-04T21:19:12Z Ionic: ah, because they are just replacing every CFLAGS+= occurnace, makes sense
2017-10-04T21:19:33Z stassats: i guess they expect -O3 to be configured with CFLAGS, ok
2017-10-04T21:19:42Z stassats: but they break CFLAGS anyway
2017-10-04T21:19:47Z stassats: without +=
2017-10-04T21:19:59Z Ionic: yeah
2017-10-04T21:21:07Z Ionic: that's rather stupid
2017-10-04T21:21:55Z stassats: gentoo was the first to force -pie stuff and it had different syntax
2017-10-04T21:22:38Z stassats: but i can't be bothered to install gentoo to figure it out
2017-10-04T21:23:04Z Ionic: and as far as I understand, even totally unnecessary now, since *FLAGS defined in the environment should be passed down to the build
2017-10-04T21:23:28Z Ionic: which seems to have been the original intend of this replacement hackery
2017-10-04T21:23:39Z stassats: well, they want to override -O3 and -g and other stuff, but they also break it completely
2017-10-04T21:24:03Z Ionic: agree'd
2017-10-04T21:24:30Z stassats: well, it replaces CFLAGS at sed time
2017-10-04T21:24:41Z Ionic: yeah
2017-10-04T21:24:47Z fe[nl]ix: no problem compiling with gcc 7.1.1 here
2017-10-04T21:25:03Z stassats: fe[nl]ix: it's just the pie business
2017-10-04T21:25:37Z fe[nl]ix: they removed the custard
2017-10-04T21:25:37Z Ionic: yeah, gentoo is stripping out -fno-pie by stupidness and that fails
2017-10-04T21:26:02Z stassats: probably because they don't have -fno-pie
2017-10-04T21:26:50Z Ionic: that sounds improbable
2017-10-04T21:27:15Z stassats: didn't https://bugs.launchpad.net/sbcl/+bug/1668986 fix it?
2017-10-04T21:27:30Z stassats: as in d905bdd3b4ce68682ea611390624b78c8d8dbdcf
2017-10-04T21:30:02Z Ionic: likely, but their sed line is overzealously removing the -fno-pie flag
2017-10-04T21:30:29Z stassats: and more other stuff
2017-10-04T21:30:37Z Ionic: and more stuff
2017-10-04T21:31:21Z stassats: can you build vanilla sbcl?
2017-10-04T21:31:35Z stassats: i think there still may be a problem with -fno-pie
2017-10-04T21:31:38Z scymtym: dougk_: i think DESCRIBE for a symbol naming a gf that is not a STANDARD-GENERIC-FUNCTION is fine except that it should say "a generic function" instead of "a compiled function"
2017-10-04T21:31:41Z Ionic: of course
2017-10-04T21:31:41Z stassats: but gentoo is a pain to install
2017-10-04T21:33:14Z stassats: Ionic: paste to http://paste.lisp.org/ your gcc -dumpspecs
2017-10-04T21:33:27Z stassats: or paste wherever, just not here
2017-10-04T21:33:54Z Ionic: naturally
2017-10-04T21:35:01Z Ionic: https://paste.pound-python.org/raw/sYm4rAx6XRQcpSPuqV9G/
2017-10-04T21:35:46Z Ionic: looks like no-pie is the default and doesn't require a special flag
2017-10-04T21:36:29Z Ionic: and, of course, GCC 6.4.0 neither recognizes -no-pie nor -nopie
2017-10-04T21:36:37Z Ionic: but that should be fine I guess
2017-10-04T21:36:45Z stassats: ok, then d905bdd3b4ce68682ea611390624b78c8d8dbdcf does fix it
2017-10-04T21:37:30Z stassats: in theory
2017-10-04T21:37:48Z stassats: well, just play with the ebuild and throw it back to the gentoo people
2017-10-04T21:39:09Z stassats: basically the two nopie -e lines should be deleted
2017-10-04T21:39:41Z stassats: but wait, if your gcc defaults to nopie, why does it break?
2017-10-04T21:40:50Z stassats: because it cuts -Wl,--export-dynamic
2017-10-04T21:40:54Z stassats: due to the lack of +=
2017-10-04T21:41:52Z Ionic: well, vanilla sources seem to work fine, the compiled binary doesn't throw an error on run-sbcl.sh --no-sysinit --no-debugger --no-userinit
2017-10-04T21:42:12Z Ionic: so it's definitely forcing *FLAGS in the makefile
2017-10-04T21:42:15Z stassats: i just tested here without -Wl,--export-dynamic, it throws a fit
2017-10-04T21:43:51Z stassats: faulting at 0, specifically
2017-10-04T21:44:36Z Ionic: yeah
2017-10-04T21:44:51Z Ionic: and if course they are also overriding LINKFLAGS, so the initial value is lost
2017-10-04T21:45:36Z Ionic: no, what they do is completely broken at that point
2017-10-04T21:46:15Z Ionic: thanks a lot for your help in tracking this down
2017-10-04T21:46:17Z stassats: and removing -O3, dubious
2017-10-04T21:46:32Z sjl joined #sbcl
2017-10-04T21:46:40Z stassats: sbcl really shouldn't be built without -O2 or 3, or at least s
2017-10-04T21:47:02Z stassats: or you'll get very poor results due to not inlining
2017-10-04T21:47:05Z Ionic: well, that will mostly be the case, since people use either -O2 or -Os as their default CFLAGS value
2017-10-04T21:47:14Z stassats: (i guess -Os prevents inlining)
2017-10-04T21:47:27Z Ionic: together with other flags of course
2017-10-04T21:47:33Z Bike joined #sbcl
2017-10-04T21:47:43Z Ionic: the "safe recommended value" is -march=native -pipe -O2 IIRC
2017-10-04T21:48:29Z stassats: how unsafe is O3, really? or is just tall tales at this point?
2017-10-04T21:49:03Z Ionic: but that's actually a non-problem - if the user environment stuff were just appended, the last option would take precedence which should be fine
2017-10-04T21:49:14Z Ionic: that may not be true for -g but oh well
2017-10-04T21:50:05Z Ionic: I seriously don't know, my best guess is that -O3 is not widely tested and I wouldn't be surprised to see strange things happen with unsafe optimizations
2017-10-04T21:50:22Z stassats: well, sbcl is
2017-10-04T21:50:40Z stassats: but O2 should be ok, gc performance-wise
2017-10-04T21:50:57Z stassats: (there's nothing else in sbcl that really needs any optimization)
2017-10-04T21:51:30Z stassats: see 3c69f04495c23868f26160de76ea62e7098b9df6 for an example
2017-10-04T21:52:03Z Xof joined #sbcl
2017-10-04T21:52:25Z Ionic: by a factor of 25, that's huge
2017-10-04T21:52:38Z Ionic: although compared to no optimization at all, that is
2017-10-04T21:52:49Z Ionic: I'd like to know the speedup as compared to -O2
2017-10-04T21:52:57Z stassats: it was just due to inlining
2017-10-04T21:53:34Z stassats: 12 years of poor gc performance, by some temporary workaround
2017-10-04T21:53:55Z Ionic: okay, so -O2 would have likely ended in the same result, since at that level the compiler is deinitely able to inline
2017-10-04T21:54:01Z Ionic: (though maybe not as agressive)
2017-10-04T21:54:07Z stassats: and i noticed it completely accidentally
2017-10-04T21:55:00Z stassats: "huh, this gc test took n seconds, which is x times slower than on my x86-64 machine, but the machine is only y times slower"
2017-10-04T21:55:41Z stassats: and i noticed it because the test output was sitting at the same place for me to notice the timings
2017-10-04T21:56:13Z stassats: had it not been such a hug factor, i wouldn't have noticed
2017-10-04T21:56:22Z stassats: huge, rather
2017-10-04T21:59:46Z Ionic: it's rather weird nobody complained
2017-10-04T22:03:40Z stassats: Ionic: nobody uses ppc...
2017-10-04T22:06:19Z Ionic: apart from a dimishing Mac user base, yeah. ppc64 is not widely in use either, but there are special interest groups. but this issue also affected ARM, right? and ARM is widely deployed
2017-10-04T22:06:54Z stassats: cross with the number of people running sbcl there
2017-10-04T22:07:58Z Ionic: I honestly don't really use sbcl either, but indirectly via maxima from time to time and I'd assume that's a tool that's also being used by people on devices like the rpi
2017-10-04T22:08:17Z Ionic: though maybe that's wrong, because I've seen images bundled with wolfram mathematica for free...
2017-10-04T22:08:26Z stassats: rpi is like 25 slower already than x86-64 machines
2017-10-04T22:08:35Z stassats: who would notice another factor or two?
2017-10-04T22:08:52Z stassats: but that wasn't 12 years for the arm port to that point, just 1 or 2
2017-10-04T22:10:40Z Ionic: well, true
2017-10-04T22:21:31Z stassats: ir1-transforms with &keys make a mess to the IR, it gets cleaned up later, but that's a lot of unnecessary stuff
2017-10-04T22:22:20Z stassats: (except for the derived type, that doesn't come back)
2017-10-04T22:23:06Z stassats: about two hundred new ctrans just for a single transform, maybe that says something about &key handling
2017-10-04T23:07:24Z karswell quit (Read error: Connection reset by peer)
2017-10-04T23:18:42Z cromachina joined #sbcl
2017-10-04T23:27:42Z milanj joined #sbcl
2017-10-04T23:48:43Z attila_lendvai quit (Quit: Leaving.)
2017-10-05T01:36:54Z stassats: really painstaking to track IR1 changes
2017-10-05T01:38:37Z AlexOn365 joined #sbcl
2017-10-05T01:53:33Z stassats: yet again, failing to figure out type loss from let conersion
2017-10-05T02:02:13Z ccl-logbot joined #sbcl
2017-10-05T02:02:13Z 
2017-10-05T02:02:13Z names: ccl-logbot milanj cromachina Xof Bike sjl jdz scymtym Ionic edgar-rft eschatologist stassats DeadTrickster pchrist phoe leah2 ym salva Posterdati slyrus foom2 dougk_ Guest43159 hooman les White_Flame galdor kini adlai myrkraverk tonton easye Colleen fe[nl]ix specbot minion Blkt jsnell Shinmera Intensity abbe dim malm pkhuong alandipert jibanes christoph_debian flip214 nicdev joshe dustinm` sigjuice irsol p_l jyc swflint copec _death fiddlerwoaroof
2017-10-05T02:02:13Z names: Lord_Nightmare luis josemanuel[m] hzp lonjil DGASAU stylewarning gko carvite aeth jrm libre-man gingerale |3b| angular_mike___ nimiux gendl brucem drmeister jackdaniel ave1 mood mjl
2017-10-05T02:16:56Z Bike quit (Ping timeout: 258 seconds)
2017-10-05T02:44:45Z ccl-logbot joined #sbcl
2017-10-05T02:44:45Z 
2017-10-05T02:44:45Z names: ccl-logbot milanj cromachina Xof sjl jdz scymtym Ionic edgar-rft eschatologist stassats DeadTrickster pchrist phoe leah2 ym salva Posterdati slyrus foom2 dougk_ Guest43159 hooman les White_Flame galdor kini adlai myrkraverk tonton easye Colleen fe[nl]ix specbot minion Blkt jsnell Shinmera Intensity abbe dim malm pkhuong alandipert jibanes christoph_debian flip214 nicdev joshe dustinm` sigjuice irsol p_l jyc swflint copec _death fiddlerwoaroof Lord_Nightmare
2017-10-05T02:44:45Z names: luis josemanuel[m] hzp lonjil DGASAU stylewarning gko carvite aeth jrm libre-man gingerale |3b| angular_mike___ nimiux gendl brucem drmeister jackdaniel ave1 mood mjl
2017-10-05T02:47:51Z stassats: but i think i figured out how to make the stay for transform-call, by annotating the main-entry
2017-10-05T03:05:10Z stassats quit (Ping timeout: 264 seconds)
2017-10-05T03:22:26Z milanj quit (Quit: This computer has gone to sleep)
2017-10-05T03:55:33Z slyrus quit (Ping timeout: 240 seconds)
2017-10-05T05:30:02Z panji joined #sbcl
2017-10-05T05:40:42Z angavrilov joined #sbcl
2017-10-05T05:50:35Z shka_ joined #sbcl
2017-10-05T06:01:19Z milanj joined #sbcl
2017-10-05T06:37:27Z scymtym quit (Ping timeout: 260 seconds)
2017-10-05T06:45:03Z shka_ quit (Ping timeout: 240 seconds)
2017-10-05T07:14:52Z carenz_ joined #sbcl
2017-10-05T07:48:30Z scymtym joined #sbcl
2017-10-05T07:50:53Z panji quit (Read error: Connection reset by peer)
2017-10-05T07:54:35Z jibanes quit (Ping timeout: 240 seconds)
2017-10-05T08:00:36Z scymtym_ joined #sbcl
2017-10-05T08:04:26Z scymtym quit (Ping timeout: 246 seconds)
2017-10-05T08:19:46Z milanj quit (Quit: This computer has gone to sleep)
2017-10-05T08:28:07Z attila_lendvai joined #sbcl
2017-10-05T09:34:46Z jibanes joined #sbcl
2017-10-05T10:44:58Z Bike joined #sbcl
2017-10-05T10:47:09Z carenz_ quit (Ping timeout: 258 seconds)
2017-10-05T11:29:19Z Bike quit (Ping timeout: 258 seconds)
2017-10-05T11:37:39Z Bike joined #sbcl
2017-10-05T11:49:13Z stassats joined #sbcl
2017-10-05T12:39:05Z Bike quit (Ping timeout: 258 seconds)
2017-10-05T12:58:46Z slyrus joined #sbcl
2017-10-05T13:04:51Z milanj joined #sbcl
2017-10-05T13:13:17Z Bike joined #sbcl
2017-10-05T13:19:48Z |3b| quit (Quit: b)
2017-10-05T13:21:23Z |3b| joined #sbcl
2017-10-05T13:51:33Z m00natic joined #sbcl
2017-10-05T13:56:05Z sjl quit (Ping timeout: 240 seconds)
2017-10-05T14:26:15Z cromachina quit (Read error: Connection reset by peer)
2017-10-05T14:34:36Z scymtym_ quit (Ping timeout: 246 seconds)
2017-10-05T14:42:39Z stassats quit (Ping timeout: 246 seconds)
2017-10-05T15:29:01Z dougk_: scymtym: but what about the problem that %fun-name was a slot-exists-p tests followed by slot-value? Seems like it should call generic-function-name, but that fails if it's a nonstandard gf with no -NAME method
2017-10-05T16:10:03Z Jesin joined #sbcl
2017-10-05T16:20:20Z tonton quit (Ping timeout: 255 seconds)
2017-10-05T16:21:53Z tonton joined #sbcl
2017-10-05T16:50:54Z m00natic quit (Remote host closed the connection)
2017-10-05T17:15:43Z scymtym joined #sbcl
2017-10-05T17:43:20Z shka_ joined #sbcl
2017-10-05T17:43:41Z Jesin quit (Quit: Leaving)
2017-10-05T17:52:53Z Jesin joined #sbcl
2017-10-05T17:54:02Z milanj quit (Quit: This computer has gone to sleep)
2017-10-05T18:22:02Z ym quit (Quit: Leaving)
2017-10-05T19:11:39Z foom2 is now known as foom
2017-10-05T19:23:01Z eschatologist quit (Ping timeout: 240 seconds)
2017-10-05T19:23:11Z eschatologist joined #sbcl
2017-10-05T19:35:24Z Bike quit (Remote host closed the connection)
2017-10-05T19:36:25Z Bike joined #sbcl
2017-10-05T20:05:22Z stassats joined #sbcl
2017-10-05T20:13:32Z shka_ quit (Ping timeout: 260 seconds)
2017-10-05T20:17:00Z fortitude joined #sbcl
2017-10-05T20:22:00Z fortitude: if I don't specify --lose-on-corruption, how likely is it that I might get a hang without a warning having been issued first?
2017-10-05T20:22:20Z fortitude: I'm thinking of Windows specifically, but a general answer would be a good start
2017-10-05T20:23:55Z |3b|: windows sbcl is pretty each to lock up (though fairly easy to not lock up too)
2017-10-05T20:24:28Z |3b|: still can't get through the tests without locking up on windows, but they are probably unusually stressful
2017-10-05T20:25:02Z |3b| doesn't usually have any problems, but i don't usually do heavily threaded things
2017-10-05T20:25:54Z |3b|: also depends on what you are doing... bad FFI or safety 0 code can arbitrarily break things without being detected
2017-10-05T20:25:58Z fortitude: I had noticed the issues with tests before, so I guess that's not too surprising
2017-10-05T20:26:22Z fortitude: I've got a multi-threaded tcp server, no special compilation settings, and I just discovered a hang
2017-10-05T20:27:02Z fortitude: but unfortunately I don't know whether it was dropping into ldb, the sort of hang described by --lose-on-corruption, or (unlikely) an application issue
2017-10-05T20:28:59Z fortitude: perhaps relatedly: are either of Anton Kovalenko's or Dmitry Kalyanov's notes on windows threading still current, or has that all been merged by now?
2017-10-05T20:30:03Z |3b| doesn't think there has been any major work on windows sbcl outside the main repo since last big update, so probably merged
2017-10-05T20:30:56Z |3b| might be misremembering though, don't remember much windows-related happening lately in general
2017-10-05T20:32:16Z fortitude: the most recent dates on any of that stuff are a few years old now, so you're probably right
2017-10-05T20:35:38Z ym joined #sbcl
2017-10-05T20:49:35Z sjl joined #sbcl
2017-10-05T20:57:42Z ym quit (Quit: Leaving)
2017-10-05T21:01:18Z ym joined #sbcl
2017-10-05T21:07:45Z Bike quit (Ping timeout: 248 seconds)
2017-10-05T21:39:55Z Bike joined #sbcl
2017-10-05T21:39:58Z phoe quit (Ping timeout: 264 seconds)
2017-10-05T22:25:32Z Jesin quit (Quit: Leaving)
2017-10-05T22:32:31Z stassats quit (Ping timeout: 248 seconds)
2017-10-05T22:34:41Z igajsin joined #sbcl
2017-10-05T22:34:45Z igajsin quit (Client Quit)
2017-10-05T22:34:59Z dougk_: fortitude: given an unexplained hang, you're probably going to need to attach with gdb and inquire where the threads are waiting
2017-10-05T22:36:31Z fortitude: dougk_: if only I could. The hang isn't reproducible (so far), I just found it in the field and had to restart the service to get things running again
2017-10-05T22:37:30Z fortitude: does gdb support that sort of attachment on windows?
2017-10-05T22:37:41Z |3b| thinks it worked last time i tried
2017-10-05T22:41:24Z |3b|: yeah, looks like it
2017-10-05T22:42:56Z phoe_ joined #sbcl
2017-10-05T22:47:05Z fortitude: |3b|: guess I just have to wait for my test system to fall over, then
2017-10-05T22:50:53Z |3b|: possibly should try once before, just to make sure you have things set up correctly and can get the pid etc
2017-10-05T22:54:53Z fortitude: probably a good idea
2017-10-05T22:55:21Z fortitude: I've checked thread stacks (in procexp, not a real debugger) on a few occasions, and they mostly seem to be variations on a theme of "WaitForMultipleObjects"
2017-10-05T22:55:42Z fortitude: I should see if gdb can produce a real stack trace
2017-10-05T22:56:16Z fortitude: guess I'll head home and try it out
2017-10-05T22:56:22Z fortitude quit (Quit: Leaving)
2017-10-05T23:05:31Z attila_lendvai quit (Quit: Leaving.)
2017-10-05T23:09:26Z Jesin joined #sbcl
2017-10-06T00:01:01Z cromachina joined #sbcl
2017-10-06T00:01:53Z fortitude joined #sbcl
2017-10-06T00:33:50Z Bicyclidine joined #sbcl
2017-10-06T00:35:25Z fortitude: I'm trying to get a backtrace of sbcl with gdb on windows, and getting a 'cannot access memory at address 0xffffffef' error
2017-10-06T00:35:42Z fortitude: gdb is from msys2, and there are some previous complaints about dll architectures
2017-10-06T00:35:51Z fortitude: is there a better way to get gdb on windows than msys2?
2017-10-06T00:36:08Z Bike quit (Ping timeout: 240 seconds)
2017-10-06T01:13:21Z eschatologist quit (Ping timeout: 240 seconds)
2017-10-06T01:38:27Z milanj joined #sbcl
2017-10-06T02:04:26Z joshe quit (Ping timeout: 255 seconds)
2017-10-06T02:04:26Z jyc quit (Ping timeout: 255 seconds)
2017-10-06T02:04:42Z joshe joined #sbcl
2017-10-06T02:05:19Z jdz quit (Ping timeout: 255 seconds)
2017-10-06T02:05:20Z drmeister quit (Ping timeout: 255 seconds)
2017-10-06T02:05:34Z jdz joined #sbcl
2017-10-06T02:07:17Z drmeister joined #sbcl
2017-10-06T02:11:38Z lacedaemon joined #sbcl
2017-10-06T02:12:38Z p_l quit (Ping timeout: 255 seconds)
2017-10-06T02:12:38Z fe[nl]ix quit (Quit: No Ping reply in 180 seconds.)
2017-10-06T02:13:42Z jyc joined #sbcl
2017-10-06T02:25:12Z |3b|` joined #sbcl
2017-10-06T02:25:53Z |3b| quit (Ping timeout: 246 seconds)
2017-10-06T02:30:32Z Bicyclidine quit (Ping timeout: 258 seconds)
2017-10-06T02:31:21Z |3b|`` joined #sbcl
2017-10-06T02:36:00Z |3b|` quit (Ping timeout: 255 seconds)
2017-10-06T03:19:35Z fortitude quit (Quit: Leaving)
2017-10-06T03:53:43Z eschatologist joined #sbcl
2017-10-06T04:05:01Z eschatologist quit (Ping timeout: 240 seconds)
2017-10-06T04:06:13Z p_l joined #sbcl
2017-10-06T04:06:29Z eschatologist joined #sbcl
2017-10-06T04:15:51Z shka_ joined #sbcl
2017-10-06T04:52:57Z milanj quit (Quit: This computer has gone to sleep)
2017-10-06T05:00:01Z shka_ quit (Ping timeout: 240 seconds)
2017-10-06T05:36:16Z milanj joined #sbcl
2017-10-06T06:01:32Z scymtym quit (Ping timeout: 260 seconds)
2017-10-06T06:30:30Z carenz_ joined #sbcl
2017-10-06T06:38:01Z eschatologist quit (Ping timeout: 240 seconds)
2017-10-06T07:18:00Z milanj quit (Quit: This computer has gone to sleep)
2017-10-06T07:28:14Z hooman quit (Read error: Connection reset by peer)
2017-10-06T07:34:48Z scymtym joined #sbcl
2017-10-06T07:56:45Z edgar-rft quit (Quit: edgar-rft)
2017-10-06T08:13:46Z scymtym: dougk_: i see. for that we could rename the parameter of the GENERIC-FUNCTION-* accessors from STANDARD-GENERIC-FUNCTION to GENERIC-FUNCTION and rely on non-standard generic function implementations defining appropriate methods. maybe we can defined methods on GENERIC-FUNCTION that signal some sort of protocol violation condition. alternative, since http://metamodular.com/CLOS-MOP/generic-function-name.html prescribes a 
2017-10-06T08:13:46Z scymtym: behavior, we could also think about moving the NAME slot into GENERIC-FUNCTION
2017-10-06T08:14:33Z scymtym: dougk_: regarding the test in describe.impure.lisp, it should probably be changed to define a subclass of GENERIC-FUNCTION and DESCRIBE an instance of that class
2017-10-06T09:25:48Z m00natic joined #sbcl
2017-10-06T10:59:13Z Bike joined #sbcl
2017-10-06T11:03:42Z ym quit (Ping timeout: 260 seconds)
2017-10-06T11:05:27Z ym joined #sbcl
2017-10-06T11:33:36Z ym quit (Remote host closed the connection)
2017-10-06T11:35:41Z milanj joined #sbcl
2017-10-06T12:28:29Z Bike quit (Ping timeout: 255 seconds)
2017-10-06T12:57:39Z Jesin quit (Quit: Leaving)
2017-10-06T13:00:57Z Bike joined #sbcl
2017-10-06T13:16:07Z |3b|`` is now known as |3b|
2017-10-06T13:19:28Z sjl quit (Ping timeout: 240 seconds)
2017-10-06T13:25:06Z cromachina quit (Read error: Connection reset by peer)
2017-10-06T14:25:42Z Jesin joined #sbcl
2017-10-06T14:36:14Z stassats joined #sbcl
2017-10-06T14:38:18Z stassats: (handler-bind ((error (lambda (c) (use-value 'list c)))) (cons 1)) => (1)
2017-10-06T14:40:01Z phoe_: wait a second...
2017-10-06T14:40:49Z phoe_: So you want to be able to supply a different function there?
2017-10-06T14:41:10Z stassats: i don't want to be able, i'm already able
2017-10-06T14:41:20Z phoe_: oh wow
2017-10-06T14:41:45Z stassats: that was a live demo
2017-10-06T14:41:54Z phoe_: that is very nice
2017-10-06T14:42:35Z stassats: i think all these restarts i've introduced are only really useful when i'm butchering the compiler
2017-10-06T14:42:41Z carenz_ quit (Ping timeout: 255 seconds)
2017-10-06T14:42:51Z stassats: "oops, a mistake, can't recompile it, the compiler itself is hosed"
2017-10-06T14:44:05Z phoe_: So you can supply new functions. Would supplying new values be of any use?
2017-10-06T14:44:21Z stassats: that's another restart, supply anything
2017-10-06T14:44:27Z phoe_: Since it's actually the first thing I thought of - the restart is named USE-VALUE after all.
2017-10-06T14:45:00Z stassats: i renamed it anyway, having two use-value isn't convenient
2017-10-06T14:45:08Z phoe_: Yes, exactly what I thought of.
2017-10-06T14:45:16Z phoe_: USE-FUNCTION or something similar?
2017-10-06T14:46:49Z stassats: replace-function
2017-10-06T14:47:12Z phoe_: Good.
2017-10-06T14:47:18Z stassats: use-function is shorter
2017-10-06T14:47:43Z stassats: but replace-function is visually more distinguishing
2017-10-06T14:47:48Z phoe_: Might form a nice pair of restarts, USE-FUNCTION and USE-VALUE
2017-10-06T14:47:50Z phoe_: I see
2017-10-06T14:48:17Z stassats: but neither are standard
2017-10-06T14:48:21Z stassats: so, can go wild
2017-10-06T14:48:26Z stassats: emojis, anyone?
2017-10-06T14:55:30Z stassats: (handler-bind ((error (lambda (c) (invoke-restart 'call-form 123)))) (cons 1)) => 123
2017-10-06T14:58:51Z phoe_: stassats: you're using implementation-specific extensions, so you can write your Lisp with characters that are not ANSI CL base chars
2017-10-06T14:58:56Z phoe_: so emojis sound good
2017-10-06T14:59:05Z stassats: except emacs
2017-10-06T15:00:13Z scymtym: rename serious-condition to 😱-condition?
2017-10-06T15:00:49Z phoe_: (deftype silly-condition () '(and condition (not serious-condition)))
2017-10-06T15:00:51Z stassats: had to copy somewhere else to see that (emacs again)
2017-10-06T15:01:10Z scymtym: hm, i wrote it in emacs
2017-10-06T15:01:32Z stassats: window-system => ns
2017-10-06T15:02:25Z scymtym: ok, i'm on window-system => x
2017-10-06T15:02:38Z scymtym: anyway, back to PACK-ITERATIVE
2017-10-06T15:02:42Z fortitude joined #sbcl
2017-10-06T15:02:53Z stassats: scymtym: what do you want to do to it?
2017-10-06T15:03:19Z scymtym: make ironclad compilation a little less slow, if possible (without slowing down other things)
2017-10-06T15:03:25Z stassats: that was my guess
2017-10-06T15:03:45Z scymtym: turns out that's pretty hard to do
2017-10-06T15:04:03Z stassats: yeah
2017-10-06T15:04:32Z stassats: right now i'm more annoyed by the actual results, not the performance
2017-10-06T15:04:46Z stassats: but i don't comprehend it enough
2017-10-06T15:07:01Z scymtym: the core of the algorithm is not that complicated. the implicit assumptions about how things have to be sorted to make it perform good are another story
2017-10-06T15:09:02Z stassats: some of it is due to earlier stages or the backend
2017-10-06T15:09:27Z stassats: like no-op SC conversions requiring a new register
2017-10-06T15:10:02Z scymtym: yeah, taking pressure off regalloc would alleviate the problem somewhat
2017-10-06T15:10:17Z stassats: but, how many cycles can be had from tighter regalloc these days, probably not a huge win
2017-10-06T15:10:34Z stassats: but just annoying seeing all the suboptimalities in the code
2017-10-06T15:10:52Z scymtym: better representation selection would probably be more important
2017-10-06T15:10:56Z scymtym: also for code size
2017-10-06T15:11:04Z scymtym: also the tagging and untagging
2017-10-06T15:11:20Z scymtym: s/also/all/
2017-10-06T15:11:41Z scymtym: … and slower than vanilla sbcl again. enough regalloc for today
2017-10-06T15:11:47Z stassats: oh yeah, i've given up on writing specialized VOPs cause they will not be selected anyhow
2017-10-06T15:12:07Z stassats: because it decides to untag and go with a different vop instead, so silly
2017-10-06T15:13:26Z scymtym: the hottest few instructions could be eliminated from PACK-ITERATIVE if ssets had a "disabled" state in addition to the deleted state
2017-10-06T15:13:33Z scymtym: maybe i will try that tomorrow
2017-10-06T15:24:48Z stassats: i wish i could come up with a way to restart type errors
2017-10-06T15:25:28Z stassats: i can do a naive restart, but due to copies and constraint propagation it would lead to problems
2017-10-06T15:28:12Z stassats: right now you can restart (on x86-64): undefined functions, invalid-arg-count, unknown &key, unbound variable
2017-10-06T15:28:35Z Jesin quit (Quit: Leaving)
2017-10-06T15:29:46Z stassats: would be possible to restart invalid-array-index-error
2017-10-06T15:36:18Z stassats: i think i can restart (values-list '(1 2 3 . 4)) everywhere
2017-10-06T15:40:08Z Jesin joined #sbcl
2017-10-06T16:31:43Z milanj quit (Quit: Leaving)
2017-10-06T16:32:14Z Jesin quit (Quit: Leaving)
2017-10-06T16:47:13Z brucem quit (Ping timeout: 276 seconds)
2017-10-06T16:52:39Z stassats quit (Ping timeout: 258 seconds)
2017-10-06T17:07:47Z shka joined #sbcl
2017-10-06T17:20:13Z scymtym: ironclad in 62 instead of 76 seconds if i didn't screw up the timing again
2017-10-06T17:25:10Z karswell joined #sbcl
2017-10-06T17:26:07Z m00natic quit (Remote host closed the connection)
2017-10-06T17:37:43Z scymtym: seems to be a tiny bit faster for some other systems as well
2017-10-06T17:42:00Z scymtym: could anybody try ./run-sbcl.sh --load PATH-TO-QUICKLISP/setup.lisp --eval '(time (asdf:load-system :ironclad :force (quote (:ironclad))))' --quit with 9c854b4f6 and again with http://paste.lisp.org/display/357816 applied?
2017-10-06T17:42:27Z stassats joined #sbcl
2017-10-06T17:44:14Z stassats: scymtym: and greedy?
2017-10-06T17:46:28Z scymtym: stassats: i haven't touched greedy
2017-10-06T17:46:38Z stassats: it's faster
2017-10-06T17:47:18Z scymtym: i know, but ironclad (locally?) declares speed > compilation-speed or somesuch. thus iterative is used
2017-10-06T17:47:33Z stassats: greedy can be forzed
2017-10-06T17:47:35Z stassats: forced
2017-10-06T17:47:46Z Jesin joined #sbcl
2017-10-06T17:47:59Z scymtym: sure
2017-10-06T17:48:03Z scymtym: let me time that
2017-10-06T17:49:49Z Jesin quit (Max SendQ exceeded)
2017-10-06T17:50:06Z scymtym: greedy is 29 seconds for me
2017-10-06T17:50:13Z Jesin joined #sbcl
2017-10-06T17:52:01Z stassats: ironclad: 95 => 46
2017-10-06T17:52:05Z stassats: with the patch
2017-10-06T17:52:46Z scymtym: and does it work?
2017-10-06T17:52:54Z scymtym: that seems a bit much
2017-10-06T17:52:58Z stassats: how do i tell?
2017-10-06T17:53:07Z scymtym: i think there are tests
2017-10-06T17:53:14Z stassats: No tests failed.;
2017-10-06T17:53:40Z dougk_: stassats: the invalid-arg-counts.2 test should also be x86-64 I think. it gets a package lock violation for sb-kernel::call-form otherwise
2017-10-06T17:53:50Z stassats: consing went up from 4,215,816,592 to 6,698,018,928
2017-10-06T17:53:53Z stassats: dougk_: true
2017-10-06T17:54:02Z scymtym: stassats: that's expected
2017-10-06T17:54:20Z scymtym: stassats: will still worth it for the speedup?
2017-10-06T17:54:29Z scymtym: s/will/maybe/
2017-10-06T17:54:38Z stassats: only the speed matters
2017-10-06T17:55:29Z stassats: and i have (declaim (debug 2))
2017-10-06T17:55:46Z scymtym: stassats: i'm still anxious. did you compare the number of forms interpreted and lambdas converted?
2017-10-06T17:56:08Z scymtym: some of my timing mishaps came from asdf compiling different amounts of code
2017-10-06T17:56:24Z stassats: scymtym: the same
2017-10-06T17:56:51Z scymtym: ok, maybe it's real, then. i wonder how your machine can be so different, though
2017-10-06T17:57:14Z stassats: well, (declaim (debug 2)), for one thing
2017-10-06T17:57:32Z scymtym: i did everything with (speed 1) (safety 1) (debug 1)
2017-10-06T17:58:04Z stassats: debug increases register pressure
2017-10-06T17:58:16Z Jesin quit (Quit: Leaving)
2017-10-06T17:58:59Z scymtym: yes. so it help the worst case even more than my test indicates
2017-10-06T17:59:46Z scymtym: i couldn't find things that get slower so far, but that's a big concern, of course. wouldn't want to slow other things down just for ironclad
2017-10-06T17:59:53Z stassats: greedy is 11 seconds
2017-10-06T18:00:19Z scymtym: yeah
2017-10-06T18:01:18Z scymtym: i will try to do something with the outer loop as well. currently, it spills one tn and restarts from scratch with that tn removed
2017-10-06T18:06:40Z scymtym: ok, THREEFISH1024, for example, spills 140 tns. that means it starts over that many times. BLAKE2S-ROUNDS: 499 tns
2017-10-06T18:14:03Z Jesin joined #sbcl
2017-10-06T18:23:18Z eschatologist joined #sbcl
2017-10-06T18:30:06Z dougk_: stassats: i think i know how to remove LRAs
2017-10-06T18:30:36Z stassats: i was thinking about that a couple of hours ago, so, how?
2017-10-06T18:30:47Z stassats: don't tell "precise backtraces"
2017-10-06T18:33:18Z dougk_: no, i have completely segregated code and data. if any word on the stack (or in register) lacks a lowtag and points to a page with CODE_PAGE_FLAG, just pin that page, or even better, find that code object and pin it.
2017-10-06T18:33:41Z dougk_: All our former problems surrounding pinning are gone. It's efficient, it doesn't touch neighboring objects.
2017-10-06T18:34:37Z dougk_: well, obviously this doesn't work on cheneygc
2017-10-06T18:35:06Z stassats: let's also remove cheneygc, while nyef is not looking
2017-10-06T18:35:18Z dougk_: i was about to push the change that segregates code, but it randomly exposes a pre-existing bug in map-allocated-objects. It avers that the start and end don't cross, but they do.
2017-10-06T18:35:31Z dougk_: yes to that!
2017-10-06T18:36:29Z Jesin quit (Quit: Leaving)
2017-10-06T18:36:38Z dougk_: i've seen this M-A-O bug before, but I see it more now.  The thing I'm doing is relaxing the allocator's bias toward always starting new pages at every possible opportunity.
2017-10-06T18:37:07Z stassats: although supporting gencgc is quite trivial, what's not trivial is getting a hold of alpha or hppa
2017-10-06T18:37:48Z stassats: but also updating the VOPs to not issue LRAs and the whole LRA mess
2017-10-06T18:38:13Z stassats: so, probably should be introduced gradually
2017-10-06T18:38:31Z stassats: (i'd settle with just supporting it on ARM64, because I don't care about anything else)
2017-10-06T18:46:25Z |3b|: is returning multiple values through funcall slow?
2017-10-06T18:46:47Z stassats: |3b|: no
2017-10-06T18:46:57Z stassats: as much as multiple values itself are slow
2017-10-06T18:47:38Z |3b|: ok, so slow in general but not funcall specifically?
2017-10-06T18:47:51Z |3b| is seeing sb-c:return-multiple at top of profile
2017-10-06T18:47:54Z stassats: yes, but not really slow
2017-10-06T18:48:35Z |3b| wonders if my sprof hack is reporting it correctly
2017-10-06T18:49:03Z attila_lendvai joined #sbcl
2017-10-06T18:49:03Z attila_lendvai quit (Changing host)
2017-10-06T18:49:03Z attila_lendvai joined #sbcl
2017-10-06T18:49:09Z stassats: return-multiple is when you are not tail calling something
2017-10-06T18:50:45Z |3b|: does that include tail call to funcall?
2017-10-06T18:51:03Z stassats: there's no difference between any other call and funcall
2017-10-06T18:51:10Z |3b|: ok
2017-10-06T18:51:49Z |3b| will have to look at the code closer then
2017-10-06T18:51:51Z stassats: i wonder if simple cleanups like unbind-special could preserve the return values
2017-10-06T18:52:06Z stassats: (yes, they could, but how hard to convinced them to?)
2017-10-06T18:53:59Z |3b|: ah, looks like it expands to multiple-value-prog1, is that it?
2017-10-06T18:54:06Z stassats: that would be it, yes
2017-10-06T18:54:34Z |3b|: ok, looks like there isn't anything after it in that particular case, so should be able to get rid of it
2017-10-06T18:54:37Z stassats: but i can't sprinkle any compiler magic over m-v-prog1
2017-10-06T18:57:46Z stassats: but handling %special-unbind is intriguing
2017-10-06T19:13:30Z eschatologist quit (Remote host closed the connection)
2017-10-06T19:29:03Z scymtym quit (Ping timeout: 258 seconds)
2017-10-06T19:37:56Z ym joined #sbcl
2017-10-06T19:46:47Z stassats: ok, i don't think we can return multiple values without tail calls currently
2017-10-06T20:14:15Z scymtym joined #sbcl
2017-10-06T20:41:18Z stassats: but default-unknown-values is written with 6 registers in mind
2017-10-06T20:48:01Z shka quit (Ping timeout: 240 seconds)
2017-10-06T21:10:34Z Bike quit (Ping timeout: 264 seconds)
2017-10-06T21:11:05Z Bike joined #sbcl
2017-10-06T21:15:45Z Bike quit (Ping timeout: 248 seconds)
2017-10-06T21:17:57Z angavrilov quit (Remote host closed the connection)
2017-10-06T21:28:38Z milanj joined #sbcl
2017-10-06T21:28:47Z sjl joined #sbcl
2017-10-06T21:30:17Z milanj quit (Client Quit)
2017-10-06T21:31:37Z stassats: we should make call-arguments-limit 32-bit
2017-10-06T21:32:03Z stassats: save on REX prefixes
2017-10-06T21:49:04Z eschatologist joined #sbcl
2017-10-06T21:53:29Z Bike joined #sbcl
2017-10-06T22:47:19Z fortitude quit (Quit: Leaving)
2017-10-06T22:58:20Z cromachina joined #sbcl
2017-10-06T23:04:18Z brucem joined #sbcl
2017-10-06T23:05:23Z brucem quit (Changing host)
2017-10-06T23:05:23Z brucem joined #sbcl
2017-10-06T23:31:45Z attila_lendvai quit (Ping timeout: 252 seconds)
2017-10-07T00:00:15Z whoman joined #sbcl
2017-10-07T00:20:14Z karswell quit (Read error: Connection reset by peer)
2017-10-07T00:21:44Z karswell joined #sbcl
2017-10-07T01:04:35Z stassats: avoiding rex prefixes is a stupid waste of time
2017-10-07T01:04:40Z stassats: and i've just wasted three hours
2017-10-07T01:42:02Z sjl quit (Ping timeout: 258 seconds)
2017-10-07T01:54:33Z sjl joined #sbcl
2017-10-07T02:40:01Z Bike quit (Ping timeout: 248 seconds)
2017-10-07T02:59:47Z stassats quit (Ping timeout: 255 seconds)
2017-10-07T05:35:33Z shka joined #sbcl
2017-10-07T07:13:01Z shka quit (Ping timeout: 240 seconds)
2017-10-07T07:31:40Z angavrilov joined #sbcl
2017-10-07T09:34:57Z edgar-rft joined #sbcl
2017-10-07T09:55:48Z attila_lendvai joined #sbcl
2017-10-07T10:08:32Z scymtym quit (Ping timeout: 260 seconds)
2017-10-07T10:20:15Z Bike joined #sbcl
2017-10-07T10:32:53Z attila_lendvai quit (Quit: Leaving.)
2017-10-07T10:34:34Z scymtym joined #sbcl
2017-10-07T10:47:11Z carenz_ joined #sbcl
2017-10-07T10:57:26Z edgar-rft quit (Quit: edgar-rft)
2017-10-07T11:54:42Z carenz_ quit (Ping timeout: 260 seconds)
2017-10-07T12:42:07Z panji joined #sbcl
2017-10-07T13:46:11Z panji quit (Read error: Connection reset by peer)
2017-10-07T13:58:51Z shka joined #sbcl
2017-10-07T14:02:11Z panji joined #sbcl
2017-10-07T14:08:55Z stassats joined #sbcl
2017-10-07T14:11:05Z stassats: working on making NTH-VALUE less silly, 10000000 times x (nth-value 10 (values 1))  0.512 seconds => 0.063
2017-10-07T14:11:25Z stassats: code size 141 bytes => 48
2017-10-07T14:11:35Z scymtym: dougk_: build failure on macos with sb-unicode disabled: https://ci.cor-lab.org/job/sbcl-master/2515/featureset=3,label=MAC_OS_mavericks_64bit/console
2017-10-07T14:11:52Z stassats: and multiple-value-bind in general, especially if it has ignored variables
2017-10-07T14:12:45Z stassats: it used to copy all the arguments with NIL default to the same locations, i.e. using stack for more than 3, and then read them
2017-10-07T14:13:07Z stassats: i can copy to the requested locations directly, without using additional stack (if enough registers are present)
2017-10-07T14:13:57Z stassats: scymtym: sure not sb-thread?
2017-10-07T14:14:24Z scymtym: let me check
2017-10-07T14:15:13Z stassats: well, there's no way the thing that is not used on sb-thread would be affect by sb-unicode
2017-10-07T14:17:41Z scymtym: stassats: this is the build with unicode explicitly disabled. is sb-thread maybe still disabled by default on macos?
2017-10-07T14:17:54Z stassats: i think so
2017-10-07T14:17:55Z scymtym: anyway other configuration on macos seem to fail as well
2017-10-07T14:18:36Z stassats: hm, i need a way to tell at IR2 conversion time to not allocate a register
2017-10-07T14:18:38Z scymtym: the complete picture will reveal itself over the next ~ 5 hours :)
2017-10-07T14:19:33Z stassats: but that's getting ahead of myself, i need to package my changes to CALL in a non hacky way
2017-10-07T14:19:51Z scymtym: stassats: initial benchmark with multiple arguments: https://techfak.de/~jmoringe/benchmark-4.png . this is biased in favor of fastgf: 1) most methods do not share the respective first specializer (best case for decision trees) 2) fastgf only does the dispatch but does not invoke the effective method (which would take more time than the dispatch itself, probably because PCL's method body walking optimizations dot not kic
2017-10-07T14:19:51Z scymtym: methods)
2017-10-07T14:20:01Z stassats: speaking of hacky, i'm tired of breaking the compiler, i think i'll work on using the cross compiler to do hacking on it, not the live compiler
2017-10-07T14:20:35Z scymtym: (setf *evaluator-mode* :interpret) while modifying the compiler?
2017-10-07T14:21:37Z stassats: maybe
2017-10-07T14:22:03Z stassats: but using the xc i can also work on other targets
2017-10-07T14:22:19Z scymtym: true
2017-10-07T14:23:09Z stassats: and *evaluator-mode* doesn't work with C-c C-c in slime
2017-10-07T14:23:42Z scymtym: that tripped me up a few times as well. have to use C-x C-e instead
2017-10-07T14:23:59Z stassats: but that's where i should voice my opposition to removing ! from SB!C?
2017-10-07T14:24:08Z stassats: scymtym: C-M-x is better
2017-10-07T14:24:26Z scymtym: does it work reliably from inside the tlf?
2017-10-07T14:24:55Z stassats: as reliably as C-c C-c doe
2017-10-07T14:25:06Z scymtym: better indeed, then
2017-10-07T14:29:58Z stassats: huh, i may be closer to "don't do any work for unused result registers"
2017-10-07T14:30:48Z stassats: since that's what i'm doing now, but for mv calls only
2017-10-07T14:31:31Z stassats: but also have a hunch on more register control in VOPs
2017-10-07T14:45:26Z stassats` joined #sbcl
2017-10-07T14:49:59Z stassats: and also, some modifications to the compiler require a lot of changes, with xc i can just load it into slime and work on it
2017-10-07T14:50:23Z stassats: i've already loaded xc into slime in the past, shouldn't be hard
2017-10-07T14:51:06Z stassats: getting it to produce real core objects a not genesis stuff would be really  nice
2017-10-07T14:54:55Z stassats quit (Read error: Connection reset by peer)
2017-10-07T15:38:03Z stassats`: pushed load-xc.lisp
2017-10-07T16:09:41Z scymtym quit (Ping timeout: 246 seconds)
2017-10-07T16:19:16Z Jesin joined #sbcl
2017-10-07T16:24:01Z stassats` quit (Ping timeout: 240 seconds)
2017-10-07T16:44:47Z panji quit (Ping timeout: 248 seconds)
2017-10-07T17:03:04Z Bike left #sbcl
2017-10-07T17:04:54Z Bike_ joined #sbcl
2017-10-07T17:05:21Z Bike_ is now known as Bike
2017-10-07T17:40:26Z panji joined #sbcl
2017-10-07T17:52:18Z stassats joined #sbcl
2017-10-07T18:55:12Z stassats: cross compiling from within slime turns out to be really useful, i was to debug several problems much faster and easier
2017-10-07T19:04:33Z Jesin quit (Quit: Leaving)
2017-10-07T19:13:55Z stassats: it's going to revolutionize the way i do large-scale compiler surgeries
2017-10-07T19:22:51Z phoe_: vivisections ftw
2017-10-07T19:42:17Z stassats: nth-value is not embarrassing on x86-64 anymore, porting to other backends is not involving, but a large chunk of assembly code needs to be rewritten
2017-10-07T20:10:25Z edgar-rft joined #sbcl
2017-10-07T20:52:38Z angavrilov quit (Remote host closed the connection)
2017-10-07T21:35:05Z stassats quit (Ping timeout: 240 seconds)
2017-10-07T21:44:47Z shka quit (Ping timeout: 255 seconds)
2017-10-07T21:49:00Z whoman quit (Ping timeout: 240 seconds)
2017-10-07T22:06:50Z ym quit (Quit: Leaving)
2017-10-07T22:31:06Z panji quit (Remote host closed the connection)
2017-10-07T22:32:13Z whoman joined #sbcl
2017-10-07T22:32:50Z ym joined #sbcl
2017-10-07T22:52:51Z stassats joined #sbcl
2017-10-07T23:12:43Z karlos joined #sbcl
2017-10-07T23:17:52Z karlos: stassats: in the case that the ! in sb!c gets removed, couldnt you still load the xc into the image just as conveniently as now, if you rename all the host packages to host-sb-c etc first?
2017-10-07T23:25:12Z ym quit (Quit: Leaving)
2017-10-07T23:25:55Z ym joined #sbcl
2017-10-08T00:01:05Z Fare joined #sbcl
2017-10-08T00:01:57Z Fare: ahem, can some kind developer upgrade ASDF in SBCL for 2-year-old 3.1.5 to months-tested 3.3.0? https://bugs.launchpad.net/sbcl/+bug/1681201
2017-10-08T00:22:02Z stassats: so that it breaks the "2 years without dealing with asdf problems" cycle?
2017-10-08T00:43:45Z stassats quit (Ping timeout: 248 seconds)
2017-10-08T01:08:06Z karlos quit (Ping timeout: 258 seconds)
2017-10-08T01:15:20Z Fare: maybe you should downgrade to 1.369, then
2017-10-08T01:16:13Z Fare: asdf is quite stable these days, but 3.3.0 still has fixes for corner cases on sbcl compared to 3.2.1
2017-10-08T01:16:32Z Fare: which also has a few compared to 3.1.5
2017-10-08T01:18:37Z Fare: especially on Windows
2017-10-08T01:19:51Z Fare: last you refused to upgrade asdf, you cited lack of testing on Windows as an argument -- even though at the time 3.1.6 or 3.1.7 hadn't been *less* tested on Windows than 3.1.5. Well, since then, 3.2.x and 3.3.0 have been extensively tested on Windows, with many fixes indeed for run-program
2017-10-08T01:20:41Z Fare: (dammit he quit a few minutes ago, so who am I talking to?)
2017-10-08T03:23:51Z joshe: he timed out so just the crickets
2017-10-08T05:18:57Z Bike quit (Quit: Lost terminal)
2017-10-08T05:30:45Z angavrilov joined #sbcl
2017-10-08T07:17:56Z karswell_ joined #sbcl
2017-10-08T07:20:33Z karswell quit (Ping timeout: 248 seconds)
2017-10-08T07:49:32Z whoman quit (Quit: Leaving)
2017-10-08T08:48:48Z jibanes quit (Ping timeout: 240 seconds)
2017-10-08T08:50:58Z jibanes joined #sbcl
2017-10-08T09:01:19Z jibanes quit (Ping timeout: 248 seconds)
2017-10-08T09:03:25Z jibanes joined #sbcl
2017-10-08T09:08:07Z shka joined #sbcl
2017-10-08T10:01:13Z scymtym joined #sbcl
2017-10-08T10:08:55Z scymtym_ joined #sbcl
2017-10-08T10:13:02Z scymtym quit (Ping timeout: 255 seconds)
2017-10-08T10:23:26Z scymtym_: Fare: i would like to upgrade asdf as well. if you are willing to provide assistance, i will prepare the patch. maybe we can convince stassats if the work is already done
2017-10-08T10:27:06Z stassats joined #sbcl
2017-10-08T10:27:50Z stassats: scymtym_: i know we'll have to update asdf eventually, i just like to delay it as much as possible
2017-10-08T10:28:00Z stassats: (it's no secret how i feel about asdf)
2017-10-08T10:28:50Z stassats: early next release cycle (to get most testing) could be the time
2017-10-08T10:31:50Z scymtym_: stassats: sure, nobody likes dealing with asdf issues. how about this: i prepare the update in a branch now and try to address any issues with Fare. if everything works out, we can merge the change early in the next cycle
2017-10-08T10:32:54Z stassats: whichever cycle it is, it's preferable to be early
2017-10-08T10:34:39Z scymtym_: yeah, that's a good idea
2017-10-08T10:35:33Z stassats` joined #sbcl
2017-10-08T10:36:44Z stassats: i wonder if some part of the compiler can be simplified, we're always extending it
2017-10-08T10:36:55Z stassats: and it's becoming a mess
2017-10-08T10:37:19Z stassats: i bet it was all clean and comprehensible at the start
2017-10-08T10:41:13Z scymtym_: i think a more powerful type system would help could do lots of the work that currently require special purpose code
2017-10-08T10:48:28Z stassats quit (Quit: ERC (IRC client for Emacs 25.1.2))
2017-10-08T10:50:33Z stassats`: one of the problems is breaking IR invariants
2017-10-08T10:50:57Z stassats`: it'd be nice to know which modification specifically is responsible
2017-10-08T10:54:33Z stassats`: time to put some more complexity in (deleting unused &optional entry points)
2017-10-08T10:57:06Z carenz_ joined #sbcl
2017-10-08T11:00:31Z scymtym_: Fare: pull-asdf.sh seems to expect the "release" branch in the asdf repository to point to the current release. is this no longer the case?
2017-10-08T11:03:30Z carenz_ quit (Ping timeout: 240 seconds)
2017-10-08T11:13:15Z forcefaction joined #sbcl
2017-10-08T11:29:11Z Posterdati quit (Remote host closed the connection)
2017-10-08T11:35:37Z Posterdati joined #sbcl
2017-10-08T11:52:04Z scymtym_ is rewriting contrib asds
2017-10-08T11:52:48Z scymtym_: lots of cruft to remove
2017-10-08T11:52:59Z edgar-rft quit (Quit: edgar-rft)
2017-10-08T12:11:35Z Bike joined #sbcl
2017-10-08T12:21:06Z scymtym_: Fare: draft: https://github.com/scymtym/sbcl/compare/asdf-3.3.0~5...asdf-3.3.0
2017-10-08T12:26:30Z scymtym_: is gc.impure.lisp bug-529014 failing for anybody else?
2017-10-08T12:29:12Z stassats`: i've seen some gc failures yesterday
2017-10-08T12:35:54Z scymtym_: going by our CI server, the failures started with "Switch on strict separation of pages containing code from non-code"
2017-10-08T12:56:07Z stassats` quit (Ping timeout: 258 seconds)
2017-10-08T13:01:13Z whoman joined #sbcl
2017-10-08T14:01:23Z josemanuel joined #sbcl
2017-10-08T14:48:40Z forcefaction quit (Remote host closed the connection)
2017-10-08T15:05:51Z Fare: scymtym_, maybe we failed to push the release branch... wait...
2017-10-08T15:07:22Z Fare: oops, thanks, indeed, we did
2017-10-08T15:09:26Z Fare: scymtym_, that patch looks generally good
2017-10-08T15:09:48Z scymtym_: Fare: no problem. i "manually" used the 3.3.0 branch in the patch
2017-10-08T15:09:54Z Fare: did you read my asdf "best practices" document, BTW? https://github.com/fare/asdf/blob/master/doc/best_practices.md
2017-10-08T15:10:09Z scymtym_: Fare: no, didn't know about that. reading …
2017-10-08T15:11:57Z scymtym_: Fare: using strings for system names is one thing i wasn't aware of. is this recommended now?
2017-10-08T15:14:19Z attila_lendvai joined #sbcl
2017-10-08T15:41:45Z Fare: yes, that is what I now recommend
2017-10-08T15:42:08Z Fare: not a high priority change, but still the official style guide.
2017-10-08T15:42:22Z scymtym_: Fare: would that change how users can interact with contrib systems in any way?
2017-10-08T15:42:30Z Fare: i.e. there is no plan for symbols to ever stop working as designators for lower-case strings.
2017-10-08T15:42:44Z Fare: no change whatsoever.
2017-10-08T15:42:44Z scymtym_: ok, i see
2017-10-08T15:43:01Z Fare: note the *lower-case* above
2017-10-08T15:43:28Z scymtym_: yes, i saw that in the guide. i can change that as well, then. while i'm at it
2017-10-08T15:43:33Z Fare: which is the confusion I want to resolve by recommending use of lower-case strings as the canonical name.
2017-10-08T15:43:49Z Fare: if you're making changes, then yes I recommend following that practice, too.
2017-10-08T15:44:03Z scymtym_: will do
2017-10-08T15:44:07Z Fare: thanks!
2017-10-08T15:44:17Z scymtym_: i have to go now. i check the logs later
2017-10-08T15:44:22Z Fare: ttyl
2017-10-08T15:48:44Z scymtym_ quit (Ping timeout: 255 seconds)
2017-10-08T16:00:00Z Fare: stassats, BTW, I hate ASDF and its bugs even more than you do, more than anyone else on earth --- that's why I started hacking on it, I suppose
2017-10-08T16:00:18Z Fare: (and wrote two alternatives to it)
2017-10-08T16:17:07Z scymtym joined #sbcl
2017-10-08T16:17:58Z stassats joined #sbcl
2017-10-08T16:46:16Z slyrus: morning
2017-10-08T16:46:37Z slyrus: sb-introspect is failing to build for me
2017-10-08T16:52:49Z sjl quit (Quit: WeeChat 1.3)
2017-10-08T16:56:12Z flip214: stassats: ISTR you talking about having no -O for most archs; is that fixed already?
2017-10-08T16:56:27Z edgar-rft joined #sbcl
2017-10-08T16:56:33Z flip214: I'd be interested in better performance on ARM, so that sounds interesting.
2017-10-08T16:56:41Z flip214: thanks in advance!
2017-10-08T17:01:45Z slyrus: dougk_: the call to sb-c:allocate-code-object fails to add the first (boolean) arg when #+immobile-code
2017-10-08T17:08:27Z Jesin joined #sbcl
2017-10-08T17:11:42Z Fare quit (Remote host closed the connection)
2017-10-08T17:21:06Z stassats: flip214: of course it's fixed
2017-10-08T17:22:08Z sixbitslacker joined #sbcl
2017-10-08T17:23:05Z sixbitslacker: hi.  just checking before i report a bug -- does 1.4.0 pass tests with --with-sb-fastevel for everyone?
2017-10-08T17:23:45Z stassats: no idea, it's no the default
2017-10-08T17:58:30Z slyrus quit (Remote host closed the connection)
2017-10-08T18:07:00Z slyrus joined #sbcl
2017-10-08T18:10:59Z jibanes quit (Ping timeout: 240 seconds)
2017-10-08T18:13:06Z jibanes joined #sbcl
2017-10-08T18:47:00Z Jesin quit (Quit: Leaving)
2017-10-08T18:52:33Z Jesin joined #sbcl
2017-10-08T19:16:10Z flip214: stassats: thanks... didn't see a matching commit, but that's my fault for sure.
2017-10-08T19:16:23Z Guest43159 quit (Ping timeout: 258 seconds)
2017-10-08T19:21:03Z PuercoPop joined #sbcl
2017-10-08T19:21:26Z PuercoPop is now known as Guest18570
2017-10-08T19:33:51Z attila_lendvai quit (Quit: Leaving.)
2017-10-08T20:15:49Z knobo joined #sbcl
2017-10-08T20:17:05Z knobo: Does the new "Don't bother filling ignored variables." patch fill ignored variables if (optimize (debug 3))? ref Commit [a7b92f]
2017-10-08T20:28:57Z sjl joined #sbcl
2017-10-08T20:51:34Z stassats: knobo: what are you asking?
2017-10-08T20:54:19Z panji joined #sbcl
2017-10-08T20:55:59Z stylewarning: I’m guessing he wants to see local variables even if they’re ignored
2017-10-08T20:56:21Z stassats: that never happened before
2017-10-08T21:02:05Z stassats: and the change doesn't even deal with local variables
2017-10-08T21:27:55Z scymtym: sixbitslacker: the --with-sb-fastevel build fails a couple of tests. it is unfortunate but expected
2017-10-08T21:28:17Z stassats: mark it as expected?
2017-10-08T21:29:16Z scymtym: i investigated once, and iirc, it's mostly non-consing assertions that fail due to consing in a testing macro or similar
2017-10-08T21:29:21Z scymtym: so not expected per se
2017-10-08T21:32:27Z sjl quit (Quit: WeeChat 1.9.1)
2017-10-08T21:34:25Z shka quit (Ping timeout: 248 seconds)
2017-10-08T21:42:25Z sjl joined #sbcl
2017-10-08T21:46:41Z scymtym: stassats: do you know about any angles for making SIMPLE-VECTORs with fill pointers? i managed to speed up PACK-ITERATIVE some more by emulating that kind of vector (instead of incrementally consing and discarding lists). that got me thinking about the lack of an efficient vector that can do VECTOR-PUSH
2017-10-08T21:48:38Z stassats: a new widetag
2017-10-08T21:48:40Z stassats: fist
2017-10-08T21:49:09Z stassats: changing make-array and all the make-array transforms
2017-10-08T21:49:26Z scymtym: that last one is the one i feared
2017-10-08T21:49:29Z stassats: and data-vector-ref
2017-10-08T21:49:50Z stassats: and new primitive object...
2017-10-08T21:50:25Z stassats: with gc accommodations
2017-10-08T21:50:29Z scymtym: and always putting fill pointers in SIMPLE-VECTORs? too wasteful?
2017-10-08T21:51:36Z stassats: i'd say yes
2017-10-08T21:51:45Z stassats: for small vectors
2017-10-08T21:52:01Z stassats: can end up costing two words
2017-10-08T21:52:18Z scymtym: i can see that, especially for small element types
2017-10-08T21:52:32Z stassats: simple-vector has only one element type
2017-10-08T21:53:19Z stassats: and adding fill pointers for all vectors is not feasible — not enough widetags to go around
2017-10-08T21:54:05Z scymtym: yeah, didn't think that through
2017-10-08T21:54:20Z scymtym: i guess simple-vector is the only candidate
2017-10-08T21:54:21Z stassats: well, though, if you always put it, no extra widetag needed
2017-10-08T21:54:41Z scymtym: child woke up
2017-10-08T21:55:01Z stassats: but the best candidate is simple-vector-with-fp
2017-10-08T21:55:19Z sixbitslacker: ok, thanks.  it's something i hadn't run into in the past, which is why i asked (and reported it).
2017-10-08T21:55:27Z stassats: well, actually, you go the array header route, it only needs one extra widetag, but costs more, one indirection
2017-10-08T21:55:34Z stassats: but no need to check for displacement
2017-10-08T21:56:18Z josemanuel quit (Quit: leaving)
2017-10-08T21:57:49Z stassats: there are some unused widetags, but do they fit the array widetag magic?
2017-10-08T21:58:46Z sixbitslacker left #sbcl
2017-10-08T21:59:13Z scymtym: i haven't looked into that. do we exploit array widetag ranges to handle multiple widetags with bitwise operations?
2017-10-08T21:59:23Z stassats: yes
2017-10-08T22:00:37Z stassats: (typep x 'array) is CMP BYTE PTR [RCX-15], -127 JB FALSE
2017-10-08T22:00:43Z scymtym: that's something i still have to add to the fastgf decision tree generator. it is smart about intervals with a fixed stride but not arbitrary bitwise operations
2017-10-08T22:01:34Z scymtym: relevant for fixnum lowtags as well
2017-10-08T22:02:35Z stassats: but, it's not a bit test, you can just compare to 125, not 129
2017-10-08T22:02:37Z stassats: so there's space
2017-10-08T22:03:24Z stassats: although right now all arrays are 8-bit tagged
2017-10-08T22:06:05Z Xof quit (Ping timeout: 240 seconds)
2017-10-08T22:08:01Z karswell_ quit (Read error: Connection reset by peer)
2017-10-08T22:08:06Z scymtym: stassats: thanks for your thoughts, i will keep them in mind for later. for now, i will try to finish and push the regalloc changes. the final form halved ironclad compilation time for me
2017-10-08T22:08:54Z stassats: you can just roll your own vectors with fill pointer
2017-10-08T22:09:09Z scymtym: i did that
2017-10-08T22:09:33Z scymtym: well, i made the simple-vector and the "fill pointer" two slots in the containing structure
2017-10-08T22:09:51Z stassats: could put it at 0
2017-10-08T22:10:32Z stassats: unless there's already a structure
2017-10-08T22:10:40Z scymtym: there is
2017-10-08T22:11:02Z scymtym: but it's kind of ugly either way. i wouldn't have done it hadn't it improved performance so much
2017-10-08T22:11:59Z stassats: but general improvements to unoptimized code is in my interests
2017-10-08T22:12:30Z stassats: poor complex vector performance was always an annoyance
2017-10-08T22:12:35Z scymtym: i find that important as well
2017-10-08T22:12:41Z scymtym: same for displaced arrays
2017-10-08T22:12:54Z scymtym: i would like to have them usable in practice
2017-10-08T22:13:10Z scymtym: oh, speaking of that
2017-10-08T22:13:11Z stassats: i converted a lot of sequence functions to use with-array-data, but they're still bad for user code
2017-10-08T22:13:46Z stassats: the standard did a blunder on making displaced arrays chainable
2017-10-08T22:14:16Z scymtym: that will only matter after sbcl backpointer things is solved, though
2017-10-08T22:14:44Z stassats: i think i had an angle on that, resolving at creating time but then updating, but i forgot how by now
2017-10-08T22:15:24Z scymtym: for that, shouldn't both %SAVE-DISPLACED-ARRAY-BACKPOINTER calls be conditional on the displaced-to array being adjustable? one doesn't seem to be
2017-10-08T22:17:05Z stassats: not getting into that now
2017-10-08T22:17:28Z stassats: still trying to delete optional entries (and just stumbled on a tangent of my own)
2017-10-08T22:18:03Z scymtym: last time we talked about this regarding the general case, i suggested using a hash-table to improve the O(N) backpointer removal and you suggested getting rid of the current backpointer mechanism and handling things in gc
2017-10-08T22:18:26Z stassats: yeah, i remember something about that, but not whether it was actually right
2017-10-08T22:18:34Z scymtym: ok, i'm not distracting you further. thanks again for the chat
2017-10-08T22:22:21Z stassats: the tangent was, if we have to check min and max arguments, and both are powers of two it can be done with a single mask
2017-10-08T22:22:42Z scymtym: nice
2017-10-08T22:25:25Z angavrilov quit (Remote host closed the connection)
2017-10-08T22:29:35Z stassats: or maybe it only works when min is zero
2017-10-08T22:35:00Z scymtym: couldn't you shift otherwise? would still save a branch
2017-10-08T22:35:22Z stassats: the register is to be preserved
2017-10-08T22:35:33Z stassats: could do a sub in that case
2017-10-08T22:36:52Z scymtym now really tries to finish the regalloc stuff before bedtime
2017-10-08T22:37:16Z stassats: i have a rule: don't work on sbcl close to bedtime
2017-10-08T22:37:31Z stassats: the "one more build to try the fix" can turn into hours
2017-10-08T22:37:39Z stassats: (i regularly break it)
2017-10-08T22:40:09Z DeadTrickster_ joined #sbcl
2017-10-08T22:42:59Z DeadTrickster quit (Ping timeout: 240 seconds)
2017-10-08T22:49:41Z scymtym: trying to sleep with things not finished can be worse, especially when under the impression of having thought of a working solution
2017-10-08T22:50:24Z scymtym: only doing finishing touches and a final test run with the regalloc thing
2017-10-08T22:50:29Z scymtym: (famous last words)
2017-10-08T22:51:10Z stassats: ARM64 can compare and shift at the same time, but the comparison values has to be loaded into a register first
2017-10-08T22:52:47Z stassats: doesn't matter anyway
2017-10-08T23:16:32Z panji quit (Quit: Leaving)
2017-10-09T00:39:05Z karswell_ joined #sbcl
2017-10-09T00:54:07Z stassats: got (defun foo (&rest args) (declare (ignore args)) 10) to produce the same code as (defun foo () (declare (optimize (safety 0))) 10)
2017-10-09T01:01:07Z stassats: much more compact and if there's a lot of arguments, faster
2017-10-09T01:26:08Z stassats quit (Ping timeout: 240 seconds)
2017-10-09T01:30:23Z stassats joined #sbcl
2017-10-09T01:52:59Z stassats: got the input stream hosed, can't print a backtrace => attach gdb, call ldb_monitor()
2017-10-09T01:53:01Z stassats: ther p PRINT-BACKTRACE
2017-10-09T01:53:16Z stassats: get its address, exit ldb and do  call funcall0(addr)
2017-10-09T01:55:07Z stassats: while typing this, i realised, --disable-debugger, doh
2017-10-09T01:55:20Z stassats: but still useful if unable to restart
2017-10-09T02:05:27Z stassats: and the build on ppc is broken
2017-10-09T02:05:38Z stassats:   The value 0 is not of type LIST, no backtrace...
2017-10-09T02:25:03Z slyrus quit (Quit: Client Quit)
2017-10-09T02:26:41Z sjl quit (Ping timeout: 248 seconds)
2017-10-09T02:42:22Z sjl joined #sbcl
2017-10-09T03:05:32Z slyrus joined #sbcl
2017-10-09T03:24:04Z stassats quit (Ping timeout: 255 seconds)
2017-10-09T03:49:45Z Bike quit (Quit: Lost terminal)
2017-10-09T04:01:47Z sjl quit (Ping timeout: 255 seconds)
2017-10-09T04:40:16Z sjl joined #sbcl
2017-10-09T04:49:14Z shka joined #sbcl
2017-10-09T05:01:21Z scymtym quit (Ping timeout: 248 seconds)
2017-10-09T05:18:07Z stylewarning: Does SBCL support SPARC64
2017-10-09T05:35:23Z dougk_ quit (Ping timeout: 255 seconds)
2017-10-09T05:45:01Z eschatologist quit (Ping timeout: 240 seconds)
2017-10-09T05:48:11Z dougk_ joined #sbcl
2017-10-09T05:51:41Z sjl quit (Ping timeout: 255 seconds)
2017-10-09T05:58:07Z angavrilov joined #sbcl
2017-10-09T06:24:31Z shka quit (Ping timeout: 248 seconds)
2017-10-09T07:23:22Z scymtym joined #sbcl
2017-10-09T08:03:32Z knobo quit (Ping timeout: 255 seconds)
2017-10-09T08:36:43Z DeadTrickster_ is now known as DeadTrickster
2017-10-09T08:45:32Z knobo joined #sbcl
2017-10-09T08:51:24Z eschatologist joined #sbcl
2017-10-09T09:55:51Z m00natic joined #sbcl
2017-10-09T11:18:22Z Bike joined #sbcl
2017-10-09T11:25:56Z foom2 joined #sbcl
2017-10-09T11:28:26Z foom quit (Ping timeout: 246 seconds)
2017-10-09T11:34:29Z scymtym quit (Ping timeout: 255 seconds)
2017-10-09T11:34:47Z stassats joined #sbcl
2017-10-09T11:37:10Z scymtym joined #sbcl
2017-10-09T11:37:43Z carenz_ joined #sbcl
2017-10-09T12:32:05Z Bike quit (Ping timeout: 255 seconds)
2017-10-09T12:38:12Z stassats: settling for only ignoring trailing optionals, can't reliably detect usage in default forms
2017-10-09T12:38:17Z stassats: need to go up a level, maybe next time
2017-10-09T12:52:52Z Jesin quit (Quit: Leaving)
2017-10-09T12:57:24Z Bike joined #sbcl
2017-10-09T13:29:07Z panji joined #sbcl
2017-10-09T13:40:46Z scymtym_ joined #sbcl
2017-10-09T13:44:59Z scymtym quit (Ping timeout: 258 seconds)
2017-10-09T13:46:37Z cromachina quit (Read error: Connection reset by peer)
2017-10-09T14:12:03Z Jesin joined #sbcl
2017-10-09T14:14:10Z Jesin quit (Remote host closed the connection)
2017-10-09T14:20:18Z Jesin joined #sbcl
2017-10-09T14:25:17Z panji quit (Read error: Connection reset by peer)
2017-10-09T14:35:23Z scymtym_ quit (Ping timeout: 255 seconds)
2017-10-09T14:42:34Z panji joined #sbcl
2017-10-09T14:59:04Z sjl joined #sbcl
2017-10-09T15:01:12Z rgrau joined #sbcl
2017-10-09T15:18:18Z stassats: got rid of the weird  JNE L1 JMP L5 L1: ...
2017-10-09T15:18:29Z stassats: which come in a lot from optional dispatch
2017-10-09T15:21:54Z sjl_ joined #sbcl
2017-10-09T15:23:59Z sjl quit (Ping timeout: 246 seconds)
2017-10-09T15:27:30Z stylewarning: stassats: do you see any value in peephole optimization for SBCL
2017-10-09T15:27:44Z stylewarning: I know it’s been thought about before
2017-10-09T15:27:52Z stassats: stylewarning: sure, but not any road to it
2017-10-09T15:27:58Z knobo quit (Ping timeout: 240 seconds)
2017-10-09T15:30:21Z stassats: but shouldn't the stupidity of representation selection behind a peephole optimizer in any case
2017-10-09T15:31:03Z panji quit (Quit: Leaving)
2017-10-09T15:31:12Z stassats: which is the source for the most "damn, i could write better assembly by hand"
2017-10-09T15:36:01Z stassats: in other news, i overtook nikodemus by the number of commits, now sitting behind dougk_
2017-10-09T15:41:46Z stylewarning: stassats: But will you overtake nikodemus by the number of cool features?
2017-10-09T15:43:54Z stassats: is the no cost (declare (ignore &rest-arg)) a cool feature?
2017-10-09T15:46:10Z stylewarning: Pretty cool
2017-10-09T15:50:28Z stylewarning: CONSTANTLY should be so fast now
2017-10-09T16:02:05Z sjl_ is now known as sjl
2017-10-09T16:05:38Z stassats: stylewarning: it is faster, but only if you pass to it a lot of arguments
2017-10-09T16:06:38Z stassats: faster either way, but that case is noticeable
2017-10-09T16:09:50Z stassats: constantly could be improved further, by not issuing any stack allocation code and just returning the closed-over value
2017-10-09T16:11:11Z stassats: done in a proper fashion would bring me closer to specialized entry points, for floats or words
2017-10-09T16:12:09Z stassats: but already have a pretty good angle on that, might just do it one day
2017-10-09T16:16:47Z stassats quit (Ping timeout: 258 seconds)
2017-10-09T16:21:37Z stylewarning: That sounds useful
2017-10-09T16:24:55Z slyrus: sb-introspect appears to be broken again
2017-10-09T16:26:11Z slyrus: only this time it's a test failure, instead of a compilation failure
2017-10-09T16:33:17Z Posterdati quit (Ping timeout: 246 seconds)
2017-10-09T16:46:35Z Posterdati joined #sbcl
2017-10-09T16:49:19Z slyrus: dougk_: gc bignum fix broke sb-introspect
2017-10-09T16:57:03Z m00natic quit (Remote host closed the connection)
2017-10-09T17:03:01Z scymtym joined #sbcl
2017-10-09T17:20:02Z knobo joined #sbcl
2017-10-09T17:24:22Z rgrau quit (Ping timeout: 264 seconds)
2017-10-09T17:32:35Z stassats joined #sbcl
2017-10-09T17:43:18Z shka_ joined #sbcl
2017-10-09T17:44:33Z scymtym quit (Ping timeout: 248 seconds)
2017-10-09T18:01:21Z sjl quit (Quit: WeeChat 1.9.1)
2017-10-09T18:02:15Z sjl joined #sbcl
2017-10-09T18:03:15Z sjl quit (Client Quit)
2017-10-09T18:03:41Z sjl joined #sbcl
2017-10-09T18:06:59Z _death: hmm, latest commit has an sb-introspect test, allocation-information.7, which fails.. also seems to have silly code: (and (let ... (eq ...) (eq ...)) ...) .. and since it fails make install skips the module silently, may want to warn instead
2017-10-09T18:08:43Z carenz_ quit (Ping timeout: 258 seconds)
2017-10-09T18:12:22Z Posterdati quit (Ping timeout: 255 seconds)
2017-10-09T18:25:28Z rgrau joined #sbcl
2017-10-09T18:26:00Z Posterdati joined #sbcl
2017-10-09T18:30:55Z Posterdati quit (Read error: Connection reset by peer)
2017-10-09T18:47:28Z Posterdati joined #sbcl
2017-10-09T18:51:11Z rgrau quit (Ping timeout: 248 seconds)
2017-10-09T19:08:37Z scymtym joined #sbcl
2017-10-09T19:24:04Z knobo quit (Quit: WeeChat 1.7)
2017-10-09T19:30:35Z Bicyclidine joined #sbcl
2017-10-09T19:31:31Z Bike quit (Ping timeout: 258 seconds)
2017-10-09T19:42:15Z Bicyclidine quit (Ping timeout: 258 seconds)
2017-10-09T19:45:05Z shka_ quit (Ping timeout: 255 seconds)
2017-10-09T19:47:39Z Bike joined #sbcl
2017-10-09T19:52:46Z josemanuel joined #sbcl
2017-10-09T19:56:28Z leah2 quit (Ping timeout: 240 seconds)
2017-10-09T20:04:34Z Bike quit (Ping timeout: 264 seconds)
2017-10-09T20:05:40Z rgrau joined #sbcl
2017-10-09T20:12:19Z leah2 joined #sbcl
2017-10-09T20:15:40Z Bike joined #sbcl
2017-10-09T20:16:45Z sjl quit (Quit: WeeChat 1.9.1)
2017-10-09T20:19:56Z lacedaemon is now known as fe[nl]ix
2017-10-09T20:27:57Z angavrilov quit (Remote host closed the connection)
2017-10-09T20:50:30Z sjl joined #sbcl
2017-10-09T21:09:14Z Bike quit (Ping timeout: 255 seconds)
2017-10-09T21:10:32Z Posterdati quit (Ping timeout: 260 seconds)
2017-10-09T21:10:40Z slyrus: thanks dougk_!
2017-10-09T21:14:02Z sjl quit (Ping timeout: 260 seconds)
2017-10-09T21:14:23Z Posterdati joined #sbcl
2017-10-09T21:32:08Z Bike joined #sbcl
2017-10-09T21:48:06Z Jesin quit (Quit: Leaving)
2017-10-09T21:52:37Z sjl joined #sbcl
2017-10-09T21:53:30Z whoman quit (Quit: Leaving)
2017-10-09T22:19:04Z josemanuel quit (Quit: leaving)
2017-10-09T22:27:35Z sjl quit (Ping timeout: 240 seconds)
2017-10-09T22:58:00Z sjl joined #sbcl
2017-10-10T00:17:09Z rgrau quit (Ping timeout: 240 seconds)
2017-10-10T00:53:58Z stassats quit (Ping timeout: 240 seconds)
2017-10-10T00:54:11Z whoman joined #sbcl
2017-10-10T02:16:45Z Guest18570 quit (Remote host closed the connection)
2017-10-10T03:55:12Z shka_ joined #sbcl
2017-10-10T03:59:10Z Bike quit (Quit: leaving)
2017-10-10T04:51:01Z eschatologist quit (Ping timeout: 240 seconds)
2017-10-10T04:58:37Z eschatologist joined #sbcl
2017-10-10T06:05:24Z angavrilov joined #sbcl
2017-10-10T06:08:31Z scymtym quit (Ping timeout: 248 seconds)
2017-10-10T06:18:10Z jrm quit (Read error: Connection reset by peer)
2017-10-10T06:18:40Z jrm joined #sbcl
2017-10-10T06:51:23Z shka_ quit (Quit: Konversation terminated!)
2017-10-10T07:11:52Z shka_ joined #sbcl
2017-10-10T07:21:40Z shka_ quit (Ping timeout: 255 seconds)
2017-10-10T07:31:26Z scymtym joined #sbcl
2017-10-10T08:30:39Z rgrau joined #sbcl
2017-10-10T08:38:44Z panji joined #sbcl
2017-10-10T08:50:19Z attila_lendvai joined #sbcl
2017-10-10T09:25:44Z attila_lendvai quit (Quit: Leaving.)
2017-10-10T09:29:36Z m00natic joined #sbcl
2017-10-10T10:41:42Z stassats joined #sbcl
2017-10-10T10:53:00Z scymtym: this seems to be clisp crashing: https://ci.cor-lab.org/job/sbcl-master/2526/featureset=6,label=ubuntu_trusty_64bit/console
2017-10-10T10:54:15Z scymtym: i wonder what "Free space: 84 bytes." means
2017-10-10T10:57:14Z flip214: there's enough space left for the calculations needed to put 3 men on the moon.
2017-10-10T10:57:26Z flip214: or women, of course. although that didn't happen.
2017-10-10T10:59:14Z stassats: ppc is inexplicably broken
2017-10-10T10:59:33Z stassats:   The value 67109569 is not of type LIST
2017-10-10T10:59:38Z stassats: the value keeps changing
2017-10-10T11:18:07Z flip214 quit (Ping timeout: 260 seconds)
2017-10-10T11:18:19Z flip214 joined #sbcl
2017-10-10T11:45:44Z Bike joined #sbcl
2017-10-10T11:57:40Z stassats: apparently, fails in the top-level loop in braid.lisp
2017-10-10T12:30:27Z Bike quit (Ping timeout: 260 seconds)
2017-10-10T12:43:34Z akovalenko joined #sbcl
2017-10-10T12:50:48Z m00natic quit (Remote host closed the connection)
2017-10-10T13:00:30Z Bike joined #sbcl
2017-10-10T13:05:27Z sjl quit (Ping timeout: 260 seconds)
2017-10-10T13:06:37Z akovalenko quit (Ping timeout: 260 seconds)
2017-10-10T14:03:31Z foom2 is now known as foom
2017-10-10T14:29:27Z panji quit (Ping timeout: 260 seconds)
2017-10-10T14:41:34Z panji joined #sbcl
2017-10-10T14:41:51Z panji quit (Client Quit)
2017-10-10T14:57:44Z m00natic joined #sbcl
2017-10-10T15:40:28Z carenz_ joined #sbcl
2017-10-10T15:49:40Z edgar-rft quit (Quit: edgar-rft)
2017-10-10T15:52:10Z minion quit (Remote host closed the connection)
2017-10-10T15:52:16Z minion joined #sbcl
2017-10-10T15:56:28Z carenz_ quit (Ping timeout: 255 seconds)
2017-10-10T16:43:27Z stassats: scymtym: regalloc: Faster hiding/showing of vertex neighbors seems to be breaking ppc
2017-10-10T16:45:08Z scymtym: stassats: that will probably be difficult to figure out. should i just revert for now?
2017-10-10T16:45:30Z stassats: it is really difficult to figure out indeed
2017-10-10T16:46:23Z scymtym: what breaks? miscompiled target image?
2017-10-10T16:46:33Z stassats: yes
2017-10-10T16:48:14Z scymtym: do you know which functions gets miscompiled?
2017-10-10T16:48:21Z stassats: no idea
2017-10-10T16:48:27Z stassats: (do-all-symbols (s)) breaks
2017-10-10T16:48:34Z stassats: whether itself or something earlier
2017-10-10T16:48:37Z stassats: or some data
2017-10-10T16:48:59Z stassats: well, Memory fault at 0x187e5998 (pc=0x187e5998)
2017-10-10T16:49:02Z stassats: that's interesting
2017-10-10T16:49:44Z scymtym: i made a mistake in any case. i intended no change in behavior
2017-10-10T16:50:27Z stassats: i'll debug it for some time
2017-10-10T16:51:14Z scymtym: thanks, but don't waste too much time on it, if you don't want to
2017-10-10T16:51:17Z scymtym: i will take a look at the change, think hard and then revert unless i can figure it out
2017-10-10T16:51:23Z stassats: well, iterative is not on by default, so it must be something called by do-all-symbols
2017-10-10T16:55:45Z attila_lendvai joined #sbcl
2017-10-10T16:55:45Z attila_lendvai quit (Changing host)
2017-10-10T16:55:45Z attila_lendvai joined #sbcl
2017-10-10T16:56:55Z stassats: it does iterate over some symbols
2017-10-10T16:57:03Z stassats: breaks on BOOLEAN
2017-10-10T16:57:27Z stassats: (do-symbols (s :cl) (print s)) too
2017-10-10T16:58:51Z stassats: that's the last symbol
2017-10-10T17:03:27Z attila_lendvai quit (Quit: Leaving.)
2017-10-10T17:09:49Z scymtym: stassats: can you try this http://paste.lisp.org/display/358231 ?
2017-10-10T17:12:54Z stassats: ok
2017-10-10T17:14:20Z Jesin joined #sbcl
2017-10-10T17:14:29Z shka_ joined #sbcl
2017-10-10T17:19:11Z stassats: scymtym: that succeeds
2017-10-10T17:19:38Z scymtym: i see some scary differences on x86_64 as well
2017-10-10T17:20:08Z scymtym: it might have just spilled everything with the broken change
2017-10-10T17:20:46Z scymtym: i wish we had a way to track changes in compiler output
2017-10-10T17:20:57Z stassats: i have an idea how to do that
2017-10-10T17:22:00Z stassats: first prerequisite, serializing IR
2017-10-10T17:22:18Z scymtym: i'm going to test some more and then push the potential fix
2017-10-10T17:22:29Z stassats: then, load up the compiler, go through all functions in SB-C, and instrument
2017-10-10T17:22:35Z stassats: record the changes with each function call
2017-10-10T17:22:36Z scymtym: so we can run later phases on canned ir1 examples?
2017-10-10T17:22:50Z stassats: then a tool that visualizes everything (the hardest part)
2017-10-10T17:23:17Z Posterdati quit (Ping timeout: 260 seconds)
2017-10-10T17:23:21Z stassats: scymtym: that could be arranged as well
2017-10-10T17:24:14Z scymtym: i had some success with mcclim for visualizing compilery things
2017-10-10T17:24:27Z stassats: first, a simple IR visualizer needs to be written (i'm thinking about doing one in JS), then a time-travelling one
2017-10-10T17:25:50Z m00natic quit (Remote host closed the connection)
2017-10-10T17:25:54Z scymtym: will the time-travelling one require non-destructive phases? or would you make snapshots?
2017-10-10T17:26:26Z stassats: on each function call go through the component and find changes to the IR
2017-10-10T17:26:51Z stassats: record them and the function that made them
2017-10-10T17:27:17Z stassats: with a call tree as well
2017-10-10T17:27:26Z stassats: so, basically, a profiler
2017-10-10T17:29:14Z scymtym: could also add tracing (in the model transformation sense): from which other ir objects has a given object been derived and by which operation. but you probably thought of that
2017-10-10T17:29:34Z Posterdati joined #sbcl
2017-10-10T17:43:36Z attila_lendvai joined #sbcl
2017-10-10T17:43:36Z attila_lendvai quit (Changing host)
2017-10-10T17:43:36Z attila_lendvai joined #sbcl
2017-10-10T19:01:59Z gko quit (Quit: ZNC - http://znc.in)
2017-10-10T19:12:50Z Posterdati|2 joined #sbcl
2017-10-10T19:13:08Z Posterdati quit (Ping timeout: 255 seconds)
2017-10-10T19:25:48Z stassats: looks like i'm able to fix a stack-analyze problem not by deleting an aver, but by doing better dead code elimination
2017-10-10T19:26:23Z stassats: (although semantically the aver is not right)
2017-10-10T19:30:58Z scymtym quit (Ping timeout: 264 seconds)
2017-10-10T19:31:22Z Posterdati|2 quit (Quit: KVIrc 4.2.0 Equilibrium http://www.kvirc.net/)
2017-10-10T19:31:44Z Posterdati joined #sbcl
2017-10-10T19:33:05Z stassats: but optimizing and fixing a bug in one go is nicer
2017-10-10T19:34:59Z gko joined #sbcl
2017-10-10T19:37:30Z angavrilov quit (Remote host closed the connection)
2017-10-10T19:38:46Z malm quit (Ping timeout: 264 seconds)
2017-10-10T19:43:20Z malm joined #sbcl
2017-10-10T19:47:05Z Bike quit (Ping timeout: 240 seconds)
2017-10-10T19:57:16Z Bike joined #sbcl
2017-10-10T20:44:51Z scymtym joined #sbcl
2017-10-10T20:49:31Z edgar-rft joined #sbcl
2017-10-10T20:58:27Z shka_ quit (Ping timeout: 240 seconds)
2017-10-10T21:08:58Z sjl joined #sbcl
2017-10-10T21:13:57Z Bike quit (Ping timeout: 240 seconds)
2017-10-10T21:50:06Z Bike joined #sbcl
2017-10-10T22:11:22Z attila_lendvai quit (Quit: Leaving.)
2017-10-10T22:13:52Z attila_lendvai joined #sbcl
2017-10-10T22:20:23Z attila_lendvai quit (Quit: Leaving.)
2017-10-10T23:21:07Z cromachina joined #sbcl
2017-10-10T23:28:49Z anticrisis joined #sbcl
2017-10-11T00:09:41Z stassats quit (Ping timeout: 240 seconds)
2017-10-11T00:25:57Z rgrau quit (Ping timeout: 240 seconds)
2017-10-11T01:00:42Z whoman quit (Read error: Connection reset by peer)
2017-10-11T01:17:19Z libre-man quit (Ping timeout: 248 seconds)
2017-10-11T01:57:54Z karswell_ quit (Read error: Connection reset by peer)
2017-10-11T02:43:35Z |3b| quit (Ping timeout: 246 seconds)
2017-10-11T03:05:46Z anticris` joined #sbcl
2017-10-11T03:08:57Z anticrisis quit (Ping timeout: 260 seconds)
2017-10-11T03:12:01Z anticris` quit (Ping timeout: 248 seconds)
2017-10-11T03:23:58Z shka_ joined #sbcl
2017-10-11T03:27:30Z Bike quit (Quit: Lost terminal)
2017-10-11T06:22:05Z scymtym quit (Ping timeout: 240 seconds)
2017-10-11T07:04:01Z shka_ quit (Ping timeout: 248 seconds)
2017-10-11T07:09:14Z angavrilov joined #sbcl
2017-10-11T07:27:40Z scymtym joined #sbcl
2017-10-11T07:39:52Z scymtym_ joined #sbcl
2017-10-11T07:43:48Z scymtym quit (Ping timeout: 246 seconds)
2017-10-11T07:53:24Z panji joined #sbcl
2017-10-11T09:02:15Z ym quit (Quit: Leaving)
2017-10-11T09:05:30Z m00natic joined #sbcl
2017-10-11T09:06:31Z rgrau joined #sbcl
2017-10-11T10:19:52Z libre-man joined #sbcl
2017-10-11T10:29:47Z dustinm` quit (Quit: Leaving)
2017-10-11T10:33:45Z dustinm` joined #sbcl
2017-10-11T10:35:09Z stassats joined #sbcl
2017-10-11T10:55:48Z Bike joined #sbcl
2017-10-11T11:02:55Z angular_mike___ quit (Ping timeout: 248 seconds)
2017-10-11T11:06:37Z angular_mike___ joined #sbcl
2017-10-11T11:35:30Z salva0 joined #sbcl
2017-10-11T11:48:14Z salva0 quit (Quit: Leaving)
2017-10-11T11:48:27Z panji quit (Ping timeout: 240 seconds)
2017-10-11T12:01:50Z panji joined #sbcl
2017-10-11T12:02:45Z panji quit (Client Quit)
2017-10-11T12:32:27Z Bike quit (Ping timeout: 240 seconds)
2017-10-11T12:44:56Z stassats: dougk_: the fasteval problem, find-lexical-var declares ENV-PAYLOAD simple-vector, while it's NIL
2017-10-11T12:45:12Z stassats: i'm not sure what the right fix is: check for NIL or put #() into env-payload
2017-10-11T12:53:00Z Jesin quit (Quit: Leaving)
2017-10-11T12:59:58Z edgar-rft quit (Quit: edgar-rft)
2017-10-11T13:03:58Z edgar-rft joined #sbcl
2017-10-11T13:04:29Z Bike joined #sbcl
2017-10-11T13:21:27Z sjl quit (Ping timeout: 260 seconds)
2017-10-11T13:56:45Z sjl joined #sbcl
2017-10-11T14:05:10Z cromachina quit (Read error: Connection reset by peer)
2017-10-11T14:19:09Z sjl quit (Quit: WeeChat 1.9)
2017-10-11T14:19:47Z sjl joined #sbcl
2017-10-11T14:20:53Z Jesin joined #sbcl
2017-10-11T14:36:27Z scymtym_ quit (Ping timeout: 246 seconds)
2017-10-11T15:02:09Z dougk_: stassats: i'll take a look at fasteval
2017-10-11T15:21:05Z rgrau quit (Ping timeout: 248 seconds)
2017-10-11T15:33:24Z edgar-rft quit (Quit: edgar-rft)
2017-10-11T15:41:20Z attila_lendvai joined #sbcl
2017-10-11T15:49:26Z |3b| joined #sbcl
2017-10-11T16:00:41Z carenz_ joined #sbcl
2017-10-11T16:05:48Z carenz_ quit (Ping timeout: 240 seconds)
2017-10-11T16:37:28Z m00natic quit (Read error: Connection reset by peer)
2017-10-11T16:56:57Z attila_lendvai quit (Quit: Leaving.)
2017-10-11T17:12:21Z sjl quit (Quit: WeeChat 1.9)
2017-10-11T17:18:19Z sjl joined #sbcl
2017-10-11T17:38:42Z Jesin quit (Ping timeout: 260 seconds)
2017-10-11T17:42:04Z Jesin joined #sbcl
2017-10-11T18:03:59Z sjl quit (Read error: Connection reset by peer)
2017-10-11T18:04:23Z sjl joined #sbcl
2017-10-11T18:43:48Z edgar-rft joined #sbcl
2017-10-11T18:45:41Z aeth quit (Read error: Connection reset by peer)
2017-10-11T18:46:30Z aeth joined #sbcl
2017-10-11T19:00:24Z akovalenko joined #sbcl
2017-10-11T19:11:30Z sjl_ joined #sbcl
2017-10-11T19:35:29Z sjl_ quit (Ping timeout: 248 seconds)
2017-10-11T19:43:25Z scymtym joined #sbcl
2017-10-11T19:46:20Z les quit (Quit: "")
2017-10-11T19:56:56Z les joined #sbcl
2017-10-11T20:13:22Z akovalenko quit (Remote host closed the connection)
2017-10-11T20:13:45Z akovalenko joined #sbcl
2017-10-11T20:31:26Z dougk_: it looks like cross-building from x86-64 to sparc is broken, but native sparc-on-sparc works. 100% reproducible. Wonder where a host dependency has crept in
2017-10-11T20:32:06Z dougk_: and I think this is affecting ppc. no reason it wouldn't be.
2017-10-11T20:42:07Z akovalenko quit (Ping timeout: 248 seconds)
2017-10-11T20:42:36Z shka_ joined #sbcl
2017-10-11T20:47:10Z Jesin quit (Quit: Leaving)
2017-10-11T20:48:08Z sjl quit (Ping timeout: 240 seconds)
2017-10-11T21:00:08Z sjl joined #sbcl
2017-10-11T21:00:38Z akovalenko joined #sbcl
2017-10-11T21:13:51Z Bike quit (Ping timeout: 258 seconds)
2017-10-11T21:17:21Z shka_ quit (Ping timeout: 248 seconds)
2017-10-11T21:30:58Z Xof joined #sbcl
2017-10-11T21:48:20Z Bike joined #sbcl
2017-10-11T21:48:49Z foom quit (Ping timeout: 255 seconds)
2017-10-11T21:53:02Z sjl quit (Ping timeout: 260 seconds)
2017-10-11T21:58:09Z foom joined #sbcl
2017-10-11T22:04:43Z DeadTrickster_ joined #sbcl
2017-10-11T22:07:35Z DeadTrickster quit (Ping timeout: 240 seconds)
2017-10-11T22:30:51Z attila_lendvai joined #sbcl
2017-10-11T22:30:51Z attila_lendvai quit (Changing host)
2017-10-11T22:30:51Z attila_lendvai joined #sbcl
2017-10-11T23:07:10Z angavrilov quit (Remote host closed the connection)
2017-10-11T23:17:20Z stassats: dougk_: mips is cross-buildable, at least two days ago
2017-10-11T23:19:25Z dougk_: host/target is 64/32 ?
2017-10-11T23:19:33Z stassats: yes, but, mips is le
2017-10-11T23:20:09Z dougk_: ok, but you reported a completely random "0 is not a list", right? and i swear that's what I saw on sparc too. I can't remember
2017-10-11T23:20:23Z stassats: scymtym fixed that
2017-10-11T23:20:32Z dougk_: oh, hm. let me try again then
2017-10-11T23:20:43Z dougk_: but that would have been broken sparc-natively as well
2017-10-11T23:20:46Z stassats: but it was ppc built on ppc
2017-10-11T23:21:12Z stassats: and the non-reproducible arm64 thing Eric reported, i think it was the same thing as well
2017-10-12T00:06:16Z attila_lendvai quit (Quit: Leaving.)
2017-10-12T01:19:35Z stassats quit (Ping timeout: 240 seconds)
2017-10-12T03:24:55Z shka_ joined #sbcl
2017-10-12T03:52:49Z Bike quit (Quit: Lost terminal)
2017-10-12T06:00:27Z scymtym quit (Ping timeout: 240 seconds)
2017-10-12T06:06:34Z angavrilov joined #sbcl
2017-10-12T06:50:41Z shka_ quit (Ping timeout: 248 seconds)
2017-10-12T07:34:51Z carenz_ joined #sbcl
2017-10-12T07:38:36Z scymtym joined #sbcl
2017-10-12T08:41:47Z scymtym: x86 and macos builds are broken (at least on our CI server): https://ci.cor-lab.org/job/sbcl-master/2533/
2017-10-12T09:27:18Z rgrau joined #sbcl
2017-10-12T09:50:09Z stassats joined #sbcl
2017-10-12T09:53:27Z m00natic joined #sbcl
2017-10-12T10:01:27Z rgrau quit (Ping timeout: 240 seconds)
2017-10-12T10:06:06Z josemanuel joined #sbcl
2017-10-12T11:08:26Z Bike joined #sbcl
2017-10-12T11:31:23Z attila_lendvai joined #sbcl
2017-10-12T11:43:31Z stassats: ALLOCATION-INFORMATION still fails, on x86
2017-10-12T11:50:51Z stassats: scymtym: the mac (not really) problems should go away
2017-10-12T11:52:00Z scymtym: stassats: great, i was starting to lose track of what broke when and where
2017-10-12T11:52:26Z stassats: can you switch on sb-thread for darwin?
2017-10-12T11:52:29Z stassats: will test more stuff
2017-10-12T11:52:55Z stassats: or maybe, i should switch them on by default
2017-10-12T11:53:21Z stassats: i'll do just that
2017-10-12T11:56:34Z stassats: done
2017-10-12T12:01:53Z scymtym: thanks. saves me a bit of work
2017-10-12T12:02:17Z stassats: that wasn't just for you
2017-10-12T12:02:47Z scymtym: i know, still better for me :)
2017-10-12T12:02:58Z stassats: but it'll be helpful now to have a -sb-thread test target
2017-10-12T12:09:35Z scymtym quit (Ping timeout: 255 seconds)
2017-10-12T12:12:00Z stassats: why are BIND and UNBOUND so eager to get assigned RDX as a temporary?
2017-10-12T12:15:59Z stassats: that's a rabbit hole
2017-10-12T12:24:58Z stassats: i think it's the old problem of different functional not communicating their register preference, but for unbind, the return wants RDX, why is UNBIND getting it?
2017-10-12T12:25:12Z stassats: did it get it first or something?
2017-10-12T12:30:01Z Bike quit (Ping timeout: 258 seconds)
2017-10-12T12:30:57Z angavrilov quit (Ping timeout: 240 seconds)
2017-10-12T12:32:04Z angavrilov joined #sbcl
2017-10-12T12:38:27Z stassats: saving on constant loads fro BIND produces 49C784246823000017001020, a 12 byte instruction
2017-10-12T12:38:36Z stassats: saves only 1 byte compared to loading NIL into a register first
2017-10-12T12:40:11Z stassats: but saves a register, but who cares, it has other plans on using registers
2017-10-12T12:41:18Z stassats: that's something that would benefit from IR tracing
2017-10-12T12:41:31Z stassats: i should finally do that
2017-10-12T12:41:44Z stassats: will open a door for many more optimizations and fixes
2017-10-12T13:01:04Z Bike joined #sbcl
2017-10-12T13:02:19Z stassats quit (Ping timeout: 255 seconds)
2017-10-12T13:07:17Z scymtym joined #sbcl
2017-10-12T13:19:03Z sjl joined #sbcl
2017-10-12T13:20:43Z sjl_ joined #sbcl
2017-10-12T13:23:34Z josemanuel quit (Quit: leaving)
2017-10-12T13:24:02Z sjl quit (Ping timeout: 260 seconds)
2017-10-12T13:24:52Z attila_lendvai quit (Quit: Leaving.)
2017-10-12T13:59:26Z foom quit (Remote host closed the connection)
2017-10-12T14:06:20Z attila_lendvai joined #sbcl
2017-10-12T14:06:20Z attila_lendvai quit (Changing host)
2017-10-12T14:06:20Z attila_lendvai joined #sbcl
2017-10-12T14:21:33Z stassats joined #sbcl
2017-10-12T14:22:40Z foom joined #sbcl
2017-10-12T14:22:47Z stassats: dougk_: would you advise on always releasing memory back to the OS?
2017-10-12T14:55:38Z Jesin joined #sbcl
2017-10-12T15:16:42Z dougk_: If you always release, then you have the overhead of the system calls plus handling the additional faults on getting the physical memory back. We should release memory back to try to meet a desired resident side, but we're not informing GC of the band in which it should try to constrain that size. That's the first step toward doing anything more intelligently
2017-10-12T15:18:12Z stassats: the other side: the OS will have to free some other pages to make room for these otherwise free pages
2017-10-12T15:26:37Z flip214: how about madvise with        MADV_DONTNEED?
2017-10-12T15:26:44Z dougk_: that is what we do
2017-10-12T15:27:20Z dougk_: but only when we want to release anything at all.  if you do it every time, it's no good because that tells the OS "i need you to zeroize these for me".  We can just do that ourselves.
2017-10-12T15:27:28Z flip214: or perhaps MADV_FREE
2017-10-12T15:27:39Z dougk_: At least that's the linx semantic.  BSD actually means "don't *care*" if paged out.
2017-10-12T15:29:07Z stassats: i want to move alloc_tramp into assembly-routines
2017-10-12T15:29:19Z stassats: and build assem.S with -PIC
2017-10-12T15:29:19Z dougk_: agree
2017-10-12T15:29:53Z stassats: and maybe have a crude sbcl.so
2017-10-12T15:30:49Z foom: there's no -KPIC assembler option for most platforms. You just need to write PIC assembly.
2017-10-12T15:30:51Z stassats: if sbcl.so is single threaded there's no much worry about stop-the-world
2017-10-12T15:32:16Z stassats: only with gc write protection
2017-10-12T15:33:48Z dougk_: might want to see whether the #define i added for ENABLE__PAGE_PROTECTION can be turned back into a settable thing. if sigsegv hasn't been intercepted, then intercept and enable protection; if it has been then run without page protection
2017-10-12T15:34:12Z dougk_: i removed the variable because I was afraid people thought that you could actually touch it dynamically and have it be meaningful, but you can not.
2017-10-12T15:34:35Z stassats: well, it might just work for some applications
2017-10-12T15:34:46Z stassats: better to have sbcl.so that sometimes work than no sbcl.so
2017-10-12T16:08:54Z stassats: would registerizing NIL help with relocating static space?
2017-10-12T16:10:08Z stassats: too few registers as it is
2017-10-12T16:10:25Z sjl joined #sbcl
2017-10-12T16:10:47Z sjl_ quit (Ping timeout: 255 seconds)
2017-10-12T16:12:37Z dougk_: NIL register would help. x86-64 has enough. x86 not enough
2017-10-12T16:13:15Z stassats: enough as in technically possible, but it'll impact performance
2017-10-12T16:15:53Z dougk_: in which direction do you predict impact? i don't know.  Comparisons to nil encode shorter. Loading nil and T into a register is shorter.
2017-10-12T16:16:12Z stassats: an always wired register
2017-10-12T16:16:14Z stassats: more spills
2017-10-12T16:21:08Z tonton quit (Ping timeout: 240 seconds)
2017-10-12T16:22:49Z lonjil quit (Quit: No Ping reply in 180 seconds.)
2017-10-12T16:23:07Z lonjil joined #sbcl
2017-10-12T16:23:14Z tonton joined #sbcl
2017-10-12T16:38:58Z m00natic quit (Remote host closed the connection)
2017-10-12T16:48:15Z rgrau joined #sbcl
2017-10-12T16:50:45Z carenz_ quit (Ping timeout: 246 seconds)
2017-10-12T17:11:01Z stassats: dougk_: i don't think anybody is really running sbcl without threads so the warning is largely academical
2017-10-12T17:11:12Z stassats: especially since the test wouldn't work at all without threads
2017-10-12T17:11:27Z stassats: and sb-sprof is broken on darwin
2017-10-12T17:12:34Z stassats: and is it still reproducible?
2017-10-12T17:14:24Z foom: You could allocate a spot for nil in tls instead of a dedicated register.
2017-10-12T17:14:43Z stassats: nothing's happening here, but then again, sb-sprof is broken
2017-10-12T17:14:56Z shka_ joined #sbcl
2017-10-12T17:21:40Z dougk_: absolutely it's reproducible, it happens to me 9 times out of 10 when running the test suite
2017-10-12T17:21:55Z dougk_: failure is in sb-concurrency, not sb-sprof
2017-10-12T17:22:30Z stassats: with that failure mode?
2017-10-12T17:22:41Z stassats: sb-concurrency tests are broken, but rarely on x86oids
2017-10-12T17:22:49Z dougk_: yes, the Feh message about page without the write_protected bit getting a wp fault
2017-10-12T17:23:09Z stassats: well, we have different darwins then, never seen that
2017-10-12T17:36:57Z easieste joined #sbcl
2017-10-12T17:40:20Z easieste quit (Client Quit)
2017-10-12T17:43:01Z stassats: i don't think i've seen the Feh thing since 2a6ce7f3621c7d3d218ec08abf83b9da50c7a22f
2017-10-12T17:43:23Z easieste joined #sbcl
2017-10-12T17:47:21Z panji joined #sbcl
2017-10-12T17:48:35Z easieste quit (Quit: easieste)
2017-10-12T18:01:50Z dougk_: i don't know what to tell you other than that I seldom see sb-concurrency passing all its tests
2017-10-12T18:02:35Z stassats: i rarely see it pass on ppc or arm64
2017-10-12T18:26:48Z edgar-rft quit (Quit: edgar-rft)
2017-10-12T19:00:24Z sjl_ joined #sbcl
2017-10-12T19:02:25Z sjl quit (Ping timeout: 248 seconds)
2017-10-12T19:03:27Z rgrau quit (Ping timeout: 248 seconds)
2017-10-12T19:26:10Z rgrau joined #sbcl
2017-10-12T19:53:27Z scymtym quit (Ping timeout: 246 seconds)
2017-10-12T20:04:35Z DR2 joined #sbcl
2017-10-12T20:07:01Z DR2: Hi folks, please forgive the FAQ.  Where is SBCL_HOME supposed to be set?  I compiled from github (using a previous version of sbcl), placing the source in /opt/src/sbcl.  Should I set it to the output directory?  I ran `sudo sh install.sh`, and it seemed to work.  But, I'm still getting an error about needing to set SBCL_HOME when running sbcl as part of a Jenkins job.
2017-10-12T20:07:40Z stassats: DR2: consult INSTALL
2017-10-12T20:12:12Z DR2: ok, figured out it needs to be set to /usr/local/bin/sbcl
2017-10-12T20:12:39Z stassats: that's the default location, so you don't need to set anything?
2017-10-12T20:12:41Z DR2: I see now that that's specified in 1.2.1
2017-10-12T20:13:00Z stassats: SBCL_HOME is only needed when you move sbcl
2017-10-12T20:13:36Z DR2: that's the real reason I'm lost (since all I did was run install.sh); maybe I've got two versions of SBCL installed somehow
2017-10-12T20:14:10Z stassats: or you have a rogue SBCL_HOME?
2017-10-12T20:15:19Z DR2: maybe my tool's build process somehow unset's SBCL_HOME
2017-10-12T20:15:37Z DR2: "((LAMBDA NIL :IN "/var/lib/jenkins/jobs/acl2-multi-arith2/workspace/ACL2_PAR/no_p/ACL2_REAL/r/LISP/sbcl/acl2-init.lisp"))    source: (ERROR "Please set environment variable SBCL_HOME.")"
2017-10-12T20:15:52Z DR2: oh well, I think my change stands a good shot atl working, thanks
2017-10-12T20:16:02Z stassats: sbcl produces no such warning
2017-10-12T20:16:20Z DR2: Ah, that's good to know
2017-10-12T20:16:39Z stassats: https://github.com/acl2/acl2/blob/master/acl2-init.lisp#L1785
2017-10-12T20:16:44Z stassats: it wants to know for some reason
2017-10-12T20:20:22Z fe[nl]ix: stassats: if nobody is really running sbcl without threads, why not make it official and drop non-thread support ?
2017-10-12T20:21:01Z stassats: fe[nl]ix: because people run it without threads where is no threads support
2017-10-12T20:21:21Z stassats: and i do run it without threads where threads are supported
2017-10-12T20:39:15Z scymtym joined #sbcl
2017-10-12T20:39:36Z rgrau quit (Read error: Connection reset by peer)
2017-10-12T20:43:10Z edgar-rft joined #sbcl
2017-10-12T20:59:54Z attila_lendvai quit (Quit: Leaving.)
2017-10-12T21:00:42Z jack_rabbit joined #sbcl
2017-10-12T21:09:19Z shka_ quit (Ping timeout: 248 seconds)
2017-10-12T21:11:16Z stassats: no more alloc_tramp in x86-64-assem.S (just running tests and pushing)
2017-10-12T21:24:57Z Bike quit (Ping timeout: 240 seconds)
2017-10-12T21:33:31Z stassats: % file sbcl.so => sbcl.so: ELF 64-bit LSB shared object, x86-64, version 1 (SYSV), dynamically linked, BuildID[sha1]=4b8936153d245007c82993d6a7129ac20dd7f17f, not stripped
2017-10-12T21:34:51Z stassats: with main()
2017-10-12T21:35:59Z stassats: coming up after the break, loading sbcl.so
2017-10-12T21:36:07Z stassats: (not into sbcl)
2017-10-12T21:43:27Z sjl_ quit (Ping timeout: 248 seconds)
2017-10-12T21:43:34Z jdz quit (Ping timeout: 264 seconds)
2017-10-12T21:49:17Z jdz joined #sbcl
2017-10-12T21:59:13Z Bike joined #sbcl
2017-10-12T22:01:57Z DR2 quit (Remote host closed the connection)
2017-10-12T22:13:50Z stassats: falls somewhere around FIND-DYNAMIC-FOREIGN-SYMBOL-ADDRESS
2017-10-12T22:14:50Z DR2 joined #sbcl
2017-10-12T22:16:22Z stassats: probably using a wrong handle
2017-10-12T22:24:22Z stassats: looks like, but don't really need to update the linkage table
2017-10-12T22:25:51Z DR2 left #sbcl
2017-10-12T22:26:06Z stassats: right
2017-10-12T22:28:47Z stassats: http://paste.lisp.org/display/358439
2017-10-12T22:34:08Z stassats: pushed to https://github.com/stassats/sbcl/tree/sbcl.so
2017-10-12T22:36:11Z stassats: now to load into something more interesting, like an iOS app
2017-10-12T22:36:26Z stassats: (not _now_ now)
2017-10-12T22:37:33Z scymtym: nice
2017-10-12T22:40:54Z stassats: i've always wanted to run SBCL on these apple A chips
2017-10-12T22:41:27Z stassats: the fastest i have is A9, but should still be faster than something in my samsung phone
2017-10-12T22:41:48Z stassats: i wonder how it will like mprotect and stuff
2017-10-12T22:42:56Z stassats: probably will have to hook up into elf (mach-o) for true delivery
2017-10-12T22:44:09Z stassats: wikipedia says mach-o is somehow related to sbcl, huh
2017-10-12T22:46:55Z panji quit (Ping timeout: 248 seconds)
2017-10-12T23:01:05Z panji joined #sbcl
2017-10-12T23:18:17Z cromachina joined #sbcl
2017-10-13T00:47:29Z panji quit (Ping timeout: 248 seconds)
2017-10-13T01:26:32Z sjl joined #sbcl
2017-10-13T01:32:55Z stassats quit (Ping timeout: 255 seconds)
2017-10-13T02:47:08Z DGASAU quit (Ping timeout: 240 seconds)
2017-10-13T03:38:02Z sjl_ joined #sbcl
2017-10-13T03:39:13Z sjl_ quit (Read error: Connection reset by peer)
2017-10-13T04:09:01Z Bike quit (Quit: Lost terminal)
2017-10-13T05:49:24Z shka_ joined #sbcl
2017-10-13T06:04:41Z scymtym quit (Ping timeout: 240 seconds)
2017-10-13T06:25:05Z slyrus quit (Ping timeout: 248 seconds)
2017-10-13T06:31:57Z shka_ quit (Ping timeout: 240 seconds)
2017-10-13T06:47:52Z shka_ joined #sbcl
2017-10-13T06:52:56Z cromachina quit (Read error: Connection reset by peer)
2017-10-13T06:53:20Z cromachina joined #sbcl
2017-10-13T06:59:12Z carenz_ joined #sbcl
2017-10-13T07:19:57Z shka_ quit (Ping timeout: 240 seconds)
2017-10-13T07:33:36Z scymtym joined #sbcl
2017-10-13T09:37:13Z rgrau joined #sbcl
2017-10-13T10:29:49Z m00natic joined #sbcl
2017-10-13T10:50:34Z stassats joined #sbcl
2017-10-13T11:07:35Z stassats: building the runtime for iOS doesn't appear to be trivial, maybe i should try sbcl.so for android first
2017-10-13T11:08:33Z stassats: since it already runs on android, but as an executable
2017-10-13T11:12:30Z stassats: but where's fun in that
2017-10-13T11:22:50Z Bike joined #sbcl
2017-10-13T11:31:34Z Bicyclidine joined #sbcl
2017-10-13T11:34:34Z Bike quit (Ping timeout: 264 seconds)
2017-10-13T12:30:58Z Bicyclidine quit (Ping timeout: 258 seconds)
2017-10-13T12:43:00Z Jesin quit (Quit: Leaving)
2017-10-13T12:53:09Z DGASAU joined #sbcl
2017-10-13T12:54:08Z Bike joined #sbcl
2017-10-13T12:55:28Z cromachina quit (Read error: Connection reset by peer)
2017-10-13T13:09:34Z edgar-rft quit (Quit: edgar-rft)
2017-10-13T13:28:41Z m00natic quit (Remote host closed the connection)
2017-10-13T13:30:47Z sjl_ joined #sbcl
2017-10-13T13:32:49Z sjl quit (Ping timeout: 248 seconds)
2017-10-13T13:41:18Z foom: stassats: building for ios will require no dynamic codegen.
2017-10-13T13:41:41Z foom: stassats: you can't ever mark a page exec, they all have to be mapped from the executable.
2017-10-13T13:42:06Z foom: so it also requires that sbcl write out code to a proper object file.
2017-10-13T14:20:56Z Shinmera: Don't Apple's terms disallow compilers & interpreters anyway, making it pretty much a no-go?
2017-10-13T14:21:22Z scymtym quit (Ping timeout: 258 seconds)
2017-10-13T14:21:59Z foom: only if you want to do those things
2017-10-13T14:28:04Z Jesin joined #sbcl
2017-10-13T15:12:47Z carenz_ quit (Ping timeout: 260 seconds)
2017-10-13T15:21:36Z slyrus joined #sbcl
2017-10-13T15:43:15Z m00natic joined #sbcl
2017-10-13T15:43:35Z scymtym joined #sbcl
2017-10-13T15:51:45Z edgar-rft joined #sbcl
2017-10-13T16:02:13Z sjl_ quit (Quit: WeeChat 1.9)
2017-10-13T16:05:40Z sjl joined #sbcl
2017-10-13T16:08:42Z sjl_ joined #sbcl
2017-10-13T16:10:09Z sjl quit (Ping timeout: 248 seconds)
2017-10-13T16:13:11Z sjl_ is now known as sjl
2017-10-13T16:50:28Z Jesin quit (Ping timeout: 240 seconds)
2017-10-13T17:15:21Z shka_ joined #sbcl
2017-10-13T17:17:16Z m00natic quit (Read error: Connection reset by peer)
2017-10-13T17:40:13Z Jesin joined #sbcl
2017-10-13T17:42:13Z rgrau quit (Ping timeout: 255 seconds)
2017-10-13T17:57:12Z eschatologist quit (Remote host closed the connection)
2017-10-13T17:57:55Z eschatologist joined #sbcl
2017-10-13T18:05:10Z Jesin quit (Ping timeout: 264 seconds)
2017-10-13T18:16:24Z Jesin joined #sbcl
2017-10-13T18:40:24Z slyrus: stassats: around?
2017-10-13T18:40:35Z slyrus: you forgot a ! in the commit before last
2017-10-13T18:40:41Z ym joined #sbcl
2017-10-13T18:48:57Z stassats: huh
2017-10-13T18:51:58Z stassats: how did it build?
2017-10-13T18:53:12Z slyrus: it didn't
2017-10-13T18:53:34Z stassats: well, got built multiple times here
2017-10-13T19:10:47Z Jesin quit (Ping timeout: 260 seconds)
2017-10-13T19:12:54Z Jesin joined #sbcl
2017-10-13T19:36:21Z slyrus: fixed now. thanks!
2017-10-13T19:41:39Z Bicyclidine joined #sbcl
2017-10-13T19:42:05Z Bike quit (Ping timeout: 240 seconds)
2017-10-13T19:53:04Z igajsin joined #sbcl
2017-10-13T19:56:04Z igajsin left #sbcl
2017-10-13T21:01:51Z jack_rabbit quit (Ping timeout: 248 seconds)
2017-10-13T21:03:08Z Bicyclidine quit (Ping timeout: 240 seconds)
2017-10-13T21:15:59Z jack_rabbit joined #sbcl
2017-10-13T21:40:05Z Bike joined #sbcl
2017-10-13T21:51:08Z libre-man quit (Quit: WeeChat 1.8)
2017-10-13T21:51:20Z sjl_ joined #sbcl
2017-10-13T21:51:33Z libre-man joined #sbcl
2017-10-13T21:51:40Z libre-man quit (Client Quit)
2017-10-13T21:53:12Z libre-man joined #sbcl
2017-10-13T21:53:46Z sjl quit (Ping timeout: 255 seconds)
2017-10-13T21:59:08Z foom quit (Remote host closed the connection)
2017-10-13T22:08:37Z sjl joined #sbcl
2017-10-13T22:10:26Z sjl_ quit (Ping timeout: 255 seconds)
2017-10-13T23:17:09Z Jesin quit (Quit: Leaving)
2017-10-13T23:26:57Z shka_ quit (Ping timeout: 240 seconds)
2017-10-13T23:35:38Z cromachina joined #sbcl
2017-10-13T23:50:55Z jack_rabbit quit (Ping timeout: 248 seconds)
2017-10-14T00:05:20Z jack_rabbit joined #sbcl
2017-10-14T01:06:17Z Jesin joined #sbcl
2017-10-14T01:06:40Z angavrilov quit (Remote host closed the connection)
2017-10-14T01:07:34Z dougk_ quit (Ping timeout: 264 seconds)
2017-10-14T01:12:27Z stassats quit (Ping timeout: 240 seconds)
2017-10-14T02:11:10Z myrkraverk quit (Ping timeout: 264 seconds)
2017-10-14T02:34:42Z DGASAU quit (Ping timeout: 260 seconds)
2017-10-14T03:20:24Z myrkraverk joined #sbcl
2017-10-14T03:35:09Z crashtestdummy joined #sbcl
2017-10-14T05:13:36Z Bike quit (Quit: Lost terminal)
2017-10-14T08:30:37Z angavrilov joined #sbcl
2017-10-14T09:51:04Z sjl_ joined #sbcl
2017-10-14T09:55:05Z sjl_ quit (Ping timeout: 240 seconds)
2017-10-14T10:39:31Z shka_ joined #sbcl
2017-10-14T11:17:39Z stassats joined #sbcl
2017-10-14T11:24:21Z scymtym quit (Remote host closed the connection)
2017-10-14T11:40:33Z scymtym joined #sbcl
2017-10-14T11:47:13Z scymtym quit (Ping timeout: 248 seconds)
2017-10-14T12:41:50Z scymtym joined #sbcl
2017-10-14T13:13:09Z scymtym: i have a design decision to make regarding a new macro for test-utils: (checked-compile-and-assert (&rest CHECKED-COMPILE-KEYWORDS &key optimize) FORM (ARGS1 VALUES1) (ARGS2 VALUES2) …). now, OPTIMIZE is a specification of optimization settings to map over, compiling and asserting for each one. what should the default behavior be if OPTIMIZE is not supplied?
2017-10-14T13:16:11Z scymtym: stassats: experimenting with the above, i noticed that compiler-2.pure.lisp / :shift-right-transform-nil-type basically does (assert (= (funcall …))). shouldn't this assert a particular return value instead?
2017-10-14T13:16:43Z stassats: that would be more helpful
2017-10-14T13:18:47Z scymtym: do you have an opinion regarding default behavior w.r.t. potentially mapping over multiple optimization settings?
2017-10-14T13:19:16Z stassats: all of them
2017-10-14T13:20:07Z stassats: as in, cross product
2017-10-14T13:21:21Z scymtym: yes, that's one of the options
2017-10-14T13:21:21Z stassats: all all for extra credit, as in, all **policy-dependent-qualities**
2017-10-14T13:21:38Z scymtym: tests take quite a bit longer with that, though
2017-10-14T13:21:52Z Bike joined #sbcl
2017-10-14T13:22:10Z stassats: let's start with crossing debug-safety-speed
2017-10-14T13:22:16Z stassats: see where that leads
2017-10-14T13:22:27Z scymtym: and some tests need :optimize :safe to filter out unsafe combinations
2017-10-14T13:22:55Z stassats: clamp safety at 1
2017-10-14T13:22:59Z scymtym: in general?
2017-10-14T13:23:07Z stassats: default
2017-10-14T13:23:12Z scymtym: ok
2017-10-14T13:25:14Z stassats: but we're not checking for assert-error here, so (safety 0) shouldn't change the result
2017-10-14T13:28:07Z scymtym: yes, this is intended for cases in which the compiled function can be called and the return value checked
2017-10-14T13:28:43Z scymtym: one danger is crashing the test runner but i think we would detect that
2017-10-14T13:29:17Z stassats: can already crash it
2017-10-14T13:47:27Z scymtym: any suggestions for the error reporting? http://paste.lisp.org/display/358615
2017-10-14T13:47:56Z stassats: optimizations on one line
2017-10-14T13:48:00Z scymtym: looking at the paste, i see the error myself, of course
2017-10-14T13:49:39Z scymtym: i put arguments and values on separate lines because i thought that would be easier to parse. did the same for optimizations, but i guess it's not necessary in that case
2017-10-14T14:50:02Z stassats: testing CMOV conversion on arm64, if the branch is taken it's slower, if the branch is not taken CSEL is faster
2017-10-14T14:50:05Z stassats: more compact, though
2017-10-14T14:50:10Z stassats: not sure if that's enough of a reason
2017-10-14T14:51:01Z stassats: CSEL performance is the same independent of the test result, but branching is sometimes slower, sometimes faster
2017-10-14T14:51:05Z stassats: maybe that'll average out
2017-10-14T14:51:28Z stassats: i wonder what gcc does
2017-10-14T14:53:38Z stassats: clang uses csel
2017-10-14T14:54:41Z stassats: and gcc too
2017-10-14T15:08:43Z stassats: the difference on x86-64 is more dramatic
2017-10-14T15:09:24Z stassats: saved the changes, not committing, since i'll also have to optimize branch-if-zero/non-zero due to the changes required to IF-EQ
2017-10-14T15:43:46Z scymtym: huh, #<SB-C:TN-REF :TN #<SB-C:TN t1 :UNUSED> :WRITE-P T :VOP SB-C:CALL-NAMED> not found in reads/writes for its TN
2017-10-14T15:44:16Z scymtym: while converting dynamic-extent.impure.lisp to CHECKED-COMPILE
2017-10-14T15:47:25Z stassats: scymtym: that's my thing
2017-10-14T15:47:53Z stassats: which test?
2017-10-14T15:51:53Z scymtym: status i added (nth-value 4 …) at one point. uploading the modified file
2017-10-14T15:54:04Z scymtym: stassats: replace test file with https://techfak.de/~jmoringe/dynamic-extent.impure-with-error.lisp and run only that test
2017-10-14T15:54:08Z scymtym: does that reproduce it?
2017-10-14T15:56:35Z stassats: yeah
2017-10-14T16:01:37Z stassats: that's sb-c::*check-consistency*, i reckon
2017-10-14T16:06:32Z stassats: scymtym: fixed
2017-10-14T16:16:58Z scymtym: stassats: thanks
2017-10-14T16:17:33Z scymtym: i'm almost done with CHECKED-COMPILE-AND-ASSERT (but it took longer than i wanted it to, of course)
2017-10-14T16:48:02Z scymtym: stassats: http://paste.lisp.org/display/358632
2017-10-14T16:49:11Z stassats: sounds like you're calling it with (((SPEED 0) (SAFETY 3) (DEBUG 0) (COMPILATION-SPEED 1) (SPACE 1)))
2017-10-14T16:50:20Z stassats: Calling (LAMBDA (&REST ARGS) (ELT ARGS 10)) compiled with (SPEED 0) (SAFETY 3) (DEBUG 0) (COMPILATION-SPEED 1) (SPACE 1)
2017-10-14T16:50:30Z scymtym: yes
2017-10-14T16:50:32Z stassats: or (funcall (LAMBDA (&REST ARGS) (ELT ARGS 10)) arguments) ...
2017-10-14T16:51:29Z scymtym: basically (funcall (compile nil '(LAMBDA (&REST ARGS) (declare (optimize (SPEED 0) (SAFETY 3) (DEBUG 0) (COMPILATION-SPEED 1) (SPACE 1))) (ELT ARGS 10))))
2017-10-14T16:52:02Z stassats: with proper line breaks, i guess that would be the clearest reporting
2017-10-14T16:52:06Z stassats: without any human words
2017-10-14T16:52:35Z scymtym: i wanted to preserve the source form and the spliced-in policy, though
2017-10-14T16:53:01Z scymtym: erm, preserve source form, print policy separately, i mean
2017-10-14T16:53:14Z stassats: how do you add the policy?
2017-10-14T16:53:44Z scymtym: by splicing it into the form as written in the message "basically …"
2017-10-14T16:54:10Z stassats: so, it is the actual compiled form, i'd say just display it
2017-10-14T16:54:21Z stassats: or i can't test it with just copy-and-paste
2017-10-14T16:54:50Z scymtym: maybe it could print a form that reproduces the error at the bottom
2017-10-14T16:55:17Z stassats: too much text, hard to parse visually
2017-10-14T16:55:57Z scymtym: i feared to many parens and unclear boundaries between form, policy, arguments and values otherwise
2017-10-14T16:56:00Z scymtym: *too
2017-10-14T16:56:25Z scymtym: would adding something copy-pastable at the end help?
2017-10-14T16:56:54Z stassats: just overloading with information
2017-10-14T16:57:53Z scymtym: i like it, but i will see what i can do
2017-10-14T16:58:05Z scymtym: later, after putting child to bed
2017-10-14T16:58:27Z scymtym: if i can, i would like to finish and push dynamic-extent.impure.lisp
2017-10-14T16:58:48Z scymtym: but now i get "caught WARNING: dead code detected at the end of SB-C::IR1-PHASES"
2017-10-14T17:01:08Z panji joined #sbcl
2017-10-14T17:01:48Z stassats: *check-consistency* again
2017-10-14T17:02:47Z scymtym: want the file again?
2017-10-14T17:02:53Z stassats: yes
2017-10-14T17:03:22Z scymtym: i haven't rebuilt since a few commits, so may not happen in HEAD
2017-10-14T17:03:35Z stassats: doubt it
2017-10-14T17:04:18Z scymtym: https://techfak.de/~jmoringe/dynamic-extent.impure.lisp
2017-10-14T17:04:24Z scymtym: checkout 06383fe6c and run only that test
2017-10-14T17:07:23Z stassats: not clear from the warning which optimization settings have caused it
2017-10-14T17:07:48Z scymtym: the ir1-phases thing is from compiling the test code, i think
2017-10-14T17:08:15Z stassats: i'll *break-on-signal* it
2017-10-14T17:12:59Z stassats: that's from (MULTIPLE-VALUE-LIST NIL)
2017-10-14T17:13:19Z stassats: what's deleting code so late?
2017-10-14T17:14:46Z scymtym: i have to go now, will read the logs later
2017-10-14T17:16:18Z stassats: deleting an optional processor
2017-10-14T17:19:33Z stassats: i'd say that consistency check is just bogus
2017-10-14T17:21:14Z stassats: interestingly, it only happens with symbol arguments
2017-10-14T17:21:21Z scymtym quit (Ping timeout: 246 seconds)
2017-10-14T17:23:06Z stassats: symbols are propagated, but not number, that's curious
2017-10-14T17:39:22Z stassats: i think it should always flush-dead-code but not report if any has been flushed
2017-10-14T17:43:24Z stassats: or i can go earlier and see why it's not deleted
2017-10-14T18:34:21Z panji quit (Quit: Leaving)
2017-10-14T19:00:52Z cromachina quit (Read error: Connection reset by peer)
2017-10-14T19:02:50Z AndreasO joined #sbcl
2017-10-14T19:26:11Z AndreasO quit (Quit: Found more important stuff than irc!)
2017-10-14T19:27:54Z AndreasO joined #sbcl
2017-10-14T20:01:37Z scymtym joined #sbcl
2017-10-14T20:18:58Z stassats: re ir1-phases warning, i think i got rid of by doing better optimization
2017-10-14T20:19:47Z stassats: and making code slightly more general, i.e. propagate-let-args replaces ir1-optimize-mv-bind after a little bit of modification
2017-10-14T20:20:16Z scymtym: great
2017-10-14T20:20:40Z stassats: although the resulting code is the same, but the IR coming out of ir1-phases is better
2017-10-14T20:20:41Z scymtym: i will have to look at the elt &rest thing before pushing the test changes
2017-10-14T20:20:51Z stassats: scymtym: it's just a safety 3 difference
2017-10-14T20:21:22Z stassats: just remove the test
2017-10-14T20:22:02Z scymtym: i can run it with a reduce range of optimization settings
2017-10-14T20:22:15Z stassats: it's just bogus
2017-10-14T20:22:15Z scymtym: but it's still a bug, isn't it?
2017-10-14T20:22:29Z stassats: no, it works as expected
2017-10-14T20:23:08Z stassats: we convert elt to nth with safety below 3
2017-10-14T20:23:30Z scymtym: but why is the warning for ELT correct?
2017-10-14T20:23:38Z stassats: and of course, my ir1-optimize-mv-bind change doesn't survive cold-init
2017-10-14T20:23:43Z stassats: scymtym: which warning? it's an error
2017-10-14T20:24:17Z scymtym: oh, right, i read my own error message wrong :)
2017-10-14T20:24:24Z scymtym: it's signalling at runtime
2017-10-14T20:24:39Z scymtym: no problem then indeed
2017-10-14T20:25:09Z scymtym: i will just filter (safety 3) for that test
2017-10-14T20:25:26Z scymtym: or remove the case
2017-10-14T20:25:34Z stassats: just remove it
2017-10-14T20:25:45Z stassats: the test author wouldn't mind, i'm sure
2017-10-14T20:26:01Z scymtym: maybe i respect our existing tests too much
2017-10-14T20:34:01Z crashtestdummy quit (Read error: Connection reset by peer)
2017-10-14T20:35:44Z crashtestdummy joined #sbcl
2017-10-14T20:37:14Z stassats: cold-init is as helpful as ever
2017-10-14T20:56:44Z stassats: found the problem
2017-10-14T21:08:57Z Jesin quit (Quit: Leaving)
2017-10-14T21:10:00Z sjl_ joined #sbcl
2017-10-14T21:10:26Z stassats: what would really simplify the compiler: only having multiple value combinations
2017-10-14T21:11:02Z sjl quit (Ping timeout: 255 seconds)
2017-10-14T21:11:13Z stassats: or more equal treatment between the two
2017-10-14T21:14:43Z stassats: not there yet, an ansi-test fails
2017-10-14T21:42:21Z stassats: not entirely happy with the change, could do more
2017-10-14T21:42:38Z angavrilov quit (Remote host closed the connection)
2017-10-14T21:51:18Z AndreasO quit (Quit: Found more important stuff than irc!)
2017-10-14T21:56:23Z sjl joined #sbcl
2017-10-14T22:29:57Z shka_ quit (Ping timeout: 260 seconds)
2017-10-14T22:38:48Z stassats quit (Ping timeout: 240 seconds)
2017-10-14T23:07:09Z sjl quit (Quit: WeeChat 1.9)
2017-10-14T23:49:47Z Jesin joined #sbcl
2017-10-15T03:20:39Z Jesin quit (Remote host closed the connection)
2017-10-15T05:04:00Z Bike quit (Quit: Lost terminal)
2017-10-15T05:58:37Z panji joined #sbcl
2017-10-15T06:30:41Z eschatologist quit (Quit: ZNC 1.6.4+deb1 - http://znc.in)
2017-10-15T06:31:27Z eschatologist joined #sbcl
2017-10-15T07:53:35Z panji quit (Ping timeout: 240 seconds)
2017-10-15T08:07:45Z panji joined #sbcl
2017-10-15T08:50:20Z panji quit (Quit: Leaving)
2017-10-15T08:52:55Z shka_ joined #sbcl
2017-10-15T09:31:16Z stassats joined #sbcl
2017-10-15T09:34:57Z scymtym quit (Ping timeout: 248 seconds)
2017-10-15T09:53:51Z scymtym joined #sbcl
2017-10-15T10:37:41Z stassats: optimizing tail (m-v-call x y), if there are no VOPs between multiple-call-* and tail-call-variable, there's no need to copy the register around, just adjust the stack arguments
2017-10-15T10:38:14Z stassats: the case of four return values, it's 5 times faster, but largely because i'm not using REP MOVS
2017-10-15T10:38:28Z stassats: but even without stacks, it's a bit faster and much more compact
2017-10-15T10:38:59Z stassats: can do the same when tehre's multiple-call-*+unbind+multiple-return, since unbind doesn't clobber any registers or flags
2017-10-15T10:39:33Z stassats: and special bindings is a common way of stopping tail calls, this should help
2017-10-15T10:41:08Z stassats: i think i can even inline the new tail-call-variable-simple in the same space it took to call tail-call-variable
2017-10-15T10:41:23Z stassats: maybe a use case for SPACE 0, or SPACE 1
2017-10-15T10:45:04Z stassats: could also use SIMD for copying, shouldn't be much of a problem if i copy one more stack place
2017-10-15T11:16:41Z stassats quit (Ping timeout: 240 seconds)
2017-10-15T11:41:08Z Bike joined #sbcl
2017-10-15T11:45:57Z DGASAU joined #sbcl
2017-10-15T14:46:41Z shka_ quit (Ping timeout: 240 seconds)
2017-10-15T15:13:48Z cromachina joined #sbcl
2017-10-15T15:43:10Z attila_lendvai joined #sbcl
2017-10-15T15:47:53Z scymtym quit (Ping timeout: 246 seconds)
2017-10-15T16:03:34Z scymtym joined #sbcl
2017-10-15T16:19:39Z attila_lendvai quit (Ping timeout: 252 seconds)
2017-10-15T16:29:17Z stassats joined #sbcl
2017-10-15T16:55:10Z pkhuong: rep movsb can be pretty fast nowadays.
2017-10-15T16:55:49Z pkhuong: it somewhat depends on alignment of the destination and of the count.
2017-10-15T17:55:35Z stassats: not for small arguments
2017-10-15T17:55:46Z stassats: small number
2017-10-15T17:56:31Z stassats: the crossover is somewhere around 180 arguments for scalar registers, even higher with SIMD
2017-10-15T18:01:20Z attila_lendvai joined #sbcl
2017-10-15T18:01:20Z attila_lendvai quit (Changing host)
2017-10-15T18:01:20Z attila_lendvai joined #sbcl
2017-10-15T18:02:41Z DeadTrickster_ quit (Ping timeout: 240 seconds)
2017-10-15T18:05:12Z DeadTrickster joined #sbcl
2017-10-15T18:05:48Z attila_lendvai quit (Ping timeout: 252 seconds)
2017-10-15T18:08:22Z attila_lendvai joined #sbcl
2017-10-15T18:08:22Z attila_lendvai quit (Changing host)
2017-10-15T18:08:22Z attila_lendvai joined #sbcl
2017-10-15T18:17:43Z zulu_inuoe joined #sbcl
2017-10-15T18:17:54Z Jesin joined #sbcl
2017-10-15T18:25:05Z Colleen quit (Ping timeout: 240 seconds)
2017-10-15T19:05:00Z ym quit (Quit: Leaving)
2017-10-15T19:14:33Z ym joined #sbcl
2017-10-15T19:18:23Z sjl_ quit (Ping timeout: 248 seconds)
2017-10-15T19:23:55Z stassats: huh, in one code combination a register is restored using move-operand, in another, the same register to the same vop gets loaded by the vop itself
2017-10-15T19:24:26Z stassats: and the move-operand case, of course, moves it into the wrong place, requiring more subsequent moves
2017-10-15T19:25:34Z stassats: and breaks my multiple-call+tail-call-variable detection
2017-10-15T19:29:53Z stassats: i can just do multiple-call+any-vop-that-doesn't-trash-registers+tail-call-variable
2017-10-15T19:30:26Z stassats: would be harder with flags, maybe not general, just a white list, move, move-operand, unbind
2017-10-15T19:31:24Z stassats: but it'd be better to be proactive and stop the register from being trashed, when possible
2017-10-15T19:31:33Z stassats: this is getting quite complicated
2017-10-15T19:48:18Z shka_ joined #sbcl
2017-10-15T20:43:43Z Jesin quit (Ping timeout: 248 seconds)
2017-10-15T21:05:38Z Jesin joined #sbcl
2017-10-15T21:16:39Z attila_lendvai quit (Ping timeout: 252 seconds)
2017-10-15T21:17:22Z attila_lendvai joined #sbcl
2017-10-15T21:55:16Z shka_ quit (Ping timeout: 252 seconds)
2017-10-15T22:16:43Z attila_lendvai quit (Quit: Leaving.)
2017-10-15T22:55:25Z Xof quit (Ping timeout: 255 seconds)
2017-10-16T00:03:13Z stassats: i guess i could avoid all the register problems by merging multiple-call vops with tail-call-variable, not just changing their behavior based on what vop precedes/follows
2017-10-16T00:03:23Z stassats: but that's an explosion of VOPs to handle each case
2017-10-16T00:05:30Z stassats: or i could go through all MOVE-OPERANDs and see if the register is only read by one vop in the same basic block, and change that read to the stack location
2017-10-16T01:17:20Z stassats quit (Ping timeout: 246 seconds)
2017-10-16T01:37:05Z myrkraverk quit (Ping timeout: 248 seconds)
2017-10-16T01:39:46Z myrkraverk joined #sbcl
2017-10-16T04:23:38Z eschatologist quit (Quit: ZNC 1.6.4+deb1 - http://znc.in)
2017-10-16T04:24:17Z eschatologist joined #sbcl
2017-10-16T04:41:35Z Bike quit (Quit: Lost terminal)
2017-10-16T05:10:27Z angavrilov joined #sbcl
2017-10-16T05:43:32Z shka_ joined #sbcl
2017-10-16T06:24:38Z angular_mike___ quit (Ping timeout: 246 seconds)
2017-10-16T06:27:57Z malm quit (Ping timeout: 240 seconds)
2017-10-16T06:28:08Z angular_mike___ joined #sbcl
2017-10-16T06:28:32Z sigjuice quit (Ping timeout: 260 seconds)
2017-10-16T06:29:20Z sigjuice joined #sbcl
2017-10-16T06:29:27Z malm joined #sbcl
2017-10-16T06:30:49Z scymtym quit (Ping timeout: 255 seconds)
2017-10-16T06:47:27Z shka_ quit (Ping timeout: 240 seconds)
2017-10-16T06:58:48Z shka_ joined #sbcl
2017-10-16T07:28:23Z shka_ quit (Ping timeout: 252 seconds)
2017-10-16T07:32:42Z scymtym joined #sbcl
2017-10-16T07:50:05Z carenz_ joined #sbcl
2017-10-16T08:00:14Z brucem quit (Quit: ZNC - http://znc.sourceforge.net)
2017-10-16T08:08:35Z rgrau joined #sbcl
2017-10-16T08:42:35Z scymtym quit (Ping timeout: 255 seconds)
2017-10-16T08:49:45Z scymtym joined #sbcl
2017-10-16T09:12:46Z mayuresh joined #sbcl
2017-10-16T09:13:02Z mayuresh left #sbcl
2017-10-16T09:18:52Z rgrau quit (Ping timeout: 260 seconds)
2017-10-16T10:22:20Z m00natic joined #sbcl
2017-10-16T11:50:29Z attila_lendvai joined #sbcl
2017-10-16T13:17:30Z stassats joined #sbcl
2017-10-16T13:25:14Z Bike joined #sbcl
2017-10-16T13:39:03Z cromachina quit (Read error: Connection reset by peer)
2017-10-16T14:01:57Z Xof joined #sbcl
2017-10-16T14:50:22Z foom joined #sbcl
2017-10-16T15:07:49Z attila_lendvai quit (Quit: Leaving.)
2017-10-16T15:09:25Z scymtym_ joined #sbcl
2017-10-16T15:13:24Z scymtym quit (Ping timeout: 246 seconds)
2017-10-16T15:41:28Z joshe quit (Quit: brb)
2017-10-16T15:42:32Z stassats quit (Ping timeout: 246 seconds)
2017-10-16T15:43:07Z joshe joined #sbcl
2017-10-16T16:02:04Z stassats joined #sbcl
2017-10-16T16:33:57Z sjl joined #sbcl
2017-10-16T16:53:57Z carenz_ quit (Ping timeout: 240 seconds)
2017-10-16T17:08:24Z dougk joined #sbcl
2017-10-16T17:10:33Z shka_ joined #sbcl
2017-10-16T17:17:39Z m00natic quit (Remote host closed the connection)
2017-10-16T18:01:49Z sjl_ joined #sbcl
2017-10-16T18:02:42Z sjl quit (Ping timeout: 260 seconds)
2017-10-16T18:30:25Z eschatologist quit (Remote host closed the connection)
2017-10-16T18:31:50Z eschatologist joined #sbcl
2017-10-16T18:48:22Z panji joined #sbcl
2017-10-16T18:56:23Z sjl joined #sbcl
2017-10-16T18:58:11Z sjl_ quit (Ping timeout: 255 seconds)
2017-10-16T19:04:48Z panji quit (Quit: Leaving)
2017-10-16T19:25:50Z scymtym_ quit (Ping timeout: 246 seconds)
2017-10-16T19:35:57Z shka_ quit (Ping timeout: 240 seconds)
2017-10-16T20:12:32Z scymtym joined #sbcl
2017-10-16T20:18:45Z Jesin quit (Quit: Leaving)
2017-10-16T20:20:17Z Jesin joined #sbcl
2017-10-16T20:20:27Z sjl quit (Ping timeout: 240 seconds)
2017-10-16T21:07:34Z sjl joined #sbcl
2017-10-16T21:08:50Z sjl_ joined #sbcl
2017-10-16T21:12:01Z sjl quit (Ping timeout: 248 seconds)
2017-10-16T21:19:22Z Bike quit (Ping timeout: 258 seconds)
2017-10-16T21:39:37Z Bike joined #sbcl
2017-10-16T21:45:17Z angavrilov quit (Remote host closed the connection)
2017-10-16T22:03:43Z Bike quit (Ping timeout: 248 seconds)
2017-10-16T22:03:50Z Bike_ joined #sbcl
2017-10-16T22:50:50Z milanj joined #sbcl
2017-10-16T23:03:19Z cromachina joined #sbcl
2017-10-16T23:20:33Z hyuman joined #sbcl
2017-10-16T23:22:23Z hyuman is now known as whoman
2017-10-17T00:52:47Z stassats quit (Ping timeout: 260 seconds)
2017-10-17T00:59:59Z sjl_ quit (Ping timeout: 255 seconds)
2017-10-17T01:12:44Z Jesin quit (Quit: Leaving)
2017-10-17T01:20:15Z Jesin joined #sbcl
2017-10-17T01:38:19Z Jesin quit (Ping timeout: 255 seconds)
2017-10-17T01:43:52Z Jesin joined #sbcl
2017-10-17T01:45:55Z milanj quit (Quit: This computer has gone to sleep)
2017-10-17T02:37:12Z copec quit (Ping timeout: 260 seconds)
2017-10-17T02:38:30Z copec joined #sbcl
2017-10-17T03:18:02Z sjl_ joined #sbcl
2017-10-17T03:52:48Z Bike_ quit (Quit: Lost terminal)
2017-10-17T04:33:48Z shka_ joined #sbcl
2017-10-17T04:57:39Z selfishman joined #sbcl
2017-10-17T04:58:59Z selfishman quit (Read error: Connection reset by peer)
2017-10-17T06:25:42Z angavrilov joined #sbcl
2017-10-17T06:33:50Z scymtym quit (Ping timeout: 252 seconds)
2017-10-17T06:58:51Z milanj joined #sbcl
2017-10-17T07:05:57Z sjl_ quit (Ping timeout: 240 seconds)
2017-10-17T07:18:55Z shka_ quit (Ping timeout: 248 seconds)
2017-10-17T07:32:05Z scymtym joined #sbcl
2017-10-17T08:09:31Z milanj quit (Quit: This computer has gone to sleep)
2017-10-17T08:14:46Z milanj joined #sbcl
2017-10-17T08:25:44Z carenz_ joined #sbcl
2017-10-17T08:42:38Z Colleen joined #sbcl
2017-10-17T09:28:19Z AndreasO joined #sbcl
2017-10-17T09:45:26Z AndreasO quit (Quit: Found more important stuff than irc!)
2017-10-17T09:48:02Z AndreasO joined #sbcl
2017-10-17T09:57:02Z AndreasO quit (Quit: Found more important stuff than irc!)
2017-10-17T10:26:36Z AndreasO joined #sbcl
2017-10-17T10:27:14Z AndreasO quit (Client Quit)
2017-10-17T10:30:11Z DeadTrickster quit (Ping timeout: 258 seconds)
2017-10-17T10:36:35Z m00natic joined #sbcl
2017-10-17T10:53:25Z stassats joined #sbcl
2017-10-17T11:39:47Z Bike joined #sbcl
2017-10-17T11:40:12Z Bike quit (Remote host closed the connection)
2017-10-17T13:00:41Z Bike joined #sbcl
2017-10-17T13:37:27Z stassats quit (Ping timeout: 240 seconds)
2017-10-17T13:41:24Z zulu_inuoe_ joined #sbcl
2017-10-17T13:43:24Z cromachina quit (Read error: Connection reset by peer)
2017-10-17T13:43:27Z zulu_inuoe quit (Ping timeout: 248 seconds)
2017-10-17T14:38:38Z Bike quit (Ping timeout: 255 seconds)
2017-10-17T14:39:16Z Bike joined #sbcl
2017-10-17T14:39:28Z scymtym quit (Ping timeout: 252 seconds)
2017-10-17T15:27:56Z stassats joined #sbcl
2017-10-17T15:28:59Z shiyaz joined #sbcl
2017-10-17T15:32:05Z carenz_ quit (Ping timeout: 240 seconds)
2017-10-17T16:08:23Z scymtym joined #sbcl
2017-10-17T16:27:20Z milanj quit (Quit: This computer has gone to sleep)
2017-10-17T16:31:30Z sjl_ joined #sbcl
2017-10-17T16:31:31Z sjl_ quit (Client Quit)
2017-10-17T16:31:57Z sjl joined #sbcl
2017-10-17T16:32:48Z Bicyclidine joined #sbcl
2017-10-17T16:35:17Z Bike quit (Ping timeout: 246 seconds)
2017-10-17T16:49:30Z rgrau joined #sbcl
2017-10-17T16:55:18Z milanj joined #sbcl
2017-10-17T17:00:34Z m00natic quit (Remote host closed the connection)
2017-10-17T17:29:53Z shka_ joined #sbcl
2017-10-17T17:56:10Z jdz quit (Ping timeout: 255 seconds)
2017-10-17T17:56:34Z jdz joined #sbcl
2017-10-17T18:55:34Z shka_ quit (Ping timeout: 255 seconds)
2017-10-17T19:00:15Z rgrau quit (Ping timeout: 248 seconds)
2017-10-17T19:07:40Z zulu_inuoe joined #sbcl
2017-10-17T19:09:51Z zulu_inuoe_ quit (Ping timeout: 248 seconds)
2017-10-17T19:26:55Z milanj quit (Quit: This computer has gone to sleep)
2017-10-17T19:29:05Z DeadTrickster joined #sbcl
2017-10-17T19:35:05Z milanj joined #sbcl
2017-10-17T20:10:12Z Jesin quit (Quit: Leaving)
2017-10-17T20:15:37Z Jesin joined #sbcl
2017-10-17T20:34:27Z Bicyclidine quit (Ping timeout: 240 seconds)
2017-10-17T20:54:27Z Bike joined #sbcl
2017-10-17T21:10:42Z angavrilov quit (Remote host closed the connection)
2017-10-17T22:03:45Z Bike quit (Ping timeout: 248 seconds)
2017-10-17T22:20:30Z milanj quit (Quit: This computer has gone to sleep)
2017-10-17T22:21:44Z hyuman joined #sbcl
2017-10-17T22:23:08Z whoman quit (Ping timeout: 252 seconds)
2017-10-17T22:36:55Z Bike joined #sbcl
2017-10-17T22:39:49Z milanj joined #sbcl
2017-10-17T22:47:19Z shka_ joined #sbcl
2017-10-17T22:50:47Z cromachina joined #sbcl
2017-10-18T00:12:15Z Bike_ joined #sbcl
2017-10-18T00:26:02Z sjl quit (Ping timeout: 246 seconds)
2017-10-18T00:40:34Z Bike quit (Disconnected by services)
2017-10-18T00:40:36Z Bike_ is now known as Bike
2017-10-18T00:40:54Z Bicyclidine joined #sbcl
2017-10-18T00:41:22Z Bicyclidine quit (Client Quit)
2017-10-18T00:52:59Z stassats quit (Ping timeout: 246 seconds)
2017-10-18T01:23:12Z milanj quit (Read error: Connection reset by peer)
2017-10-18T01:23:48Z milanj joined #sbcl
2017-10-18T01:24:01Z milanj quit (Client Quit)
2017-10-18T01:26:28Z milanj joined #sbcl
2017-10-18T01:37:14Z milanj quit (Quit: This computer has gone to sleep)
2017-10-18T01:59:33Z Jesin quit (Quit: Leaving)
2017-10-18T02:02:39Z Jesin joined #sbcl
2017-10-18T02:30:05Z slyrus quit (Ping timeout: 240 seconds)
2017-10-18T03:46:40Z Bike quit (Quit: Lost terminal)
2017-10-18T03:47:05Z myrkraverk quit (Ping timeout: 252 seconds)
2017-10-18T04:13:41Z Jesin quit (Ping timeout: 240 seconds)
2017-10-18T04:55:04Z Jesin joined #sbcl
2017-10-18T06:01:17Z scymtym quit (Ping timeout: 252 seconds)
2017-10-18T06:42:59Z attila_lendvai joined #sbcl
2017-10-18T06:53:40Z myrkraverk joined #sbcl
2017-10-18T07:01:43Z angavrilov joined #sbcl
2017-10-18T07:30:48Z attila_lendvai quit (Ping timeout: 252 seconds)
2017-10-18T07:32:20Z attila_lendvai joined #sbcl
2017-10-18T07:33:54Z attila_lendvai quit (Client Quit)
2017-10-18T07:38:54Z scymtym joined #sbcl
2017-10-18T07:40:27Z shka_ quit (Ping timeout: 240 seconds)
2017-10-18T07:48:33Z carenz_ joined #sbcl
2017-10-18T08:30:58Z Xof quit (Ping timeout: 240 seconds)
2017-10-18T08:37:11Z DeadTrickster quit (Ping timeout: 255 seconds)
2017-10-18T08:37:51Z DeadTrickster joined #sbcl
2017-10-18T08:56:42Z milanj joined #sbcl
2017-10-18T08:57:16Z milanj quit (Client Quit)
2017-10-18T09:11:24Z adlai quit (Ping timeout: 258 seconds)
2017-10-18T09:47:11Z rgrau joined #sbcl
2017-10-18T09:59:34Z m00natic joined #sbcl
2017-10-18T10:31:48Z rgrau quit (Ping timeout: 240 seconds)
2017-10-18T10:49:21Z adlai joined #sbcl
2017-10-18T11:14:14Z stassats joined #sbcl
2017-10-18T11:22:26Z leah2 quit (Ping timeout: 255 seconds)
2017-10-18T11:37:20Z leah2 joined #sbcl
2017-10-18T12:55:18Z Jesin quit (Quit: Leaving)
2017-10-18T12:56:52Z Bike joined #sbcl
2017-10-18T13:31:23Z shiyas joined #sbcl
2017-10-18T13:36:28Z cromachina quit (Read error: Connection reset by peer)
2017-10-18T14:19:13Z leah2 quit (Ping timeout: 248 seconds)
2017-10-18T14:30:11Z leah2 joined #sbcl
2017-10-18T14:33:31Z shiyaz quit (Remote host closed the connection)
2017-10-18T14:33:37Z shiyas quit (Remote host closed the connection)
2017-10-18T14:33:59Z shiyaz joined #sbcl
2017-10-18T14:34:04Z shiyas joined #sbcl
2017-10-18T14:46:36Z karswell joined #sbcl
2017-10-18T14:53:21Z shiyas quit (Quit: Leaving)
2017-10-18T15:23:22Z shiyaz quit (Ping timeout: 260 seconds)
2017-10-18T15:26:41Z carenz_ quit (Ping timeout: 240 seconds)
2017-10-18T15:28:36Z edgar-rft quit (Quit: edgar-rft)
2017-10-18T15:30:21Z hyuman quit (Quit: Leaving)
2017-10-18T15:30:57Z stassats quit (Ping timeout: 240 seconds)
2017-10-18T15:57:02Z Jesin joined #sbcl
2017-10-18T15:58:05Z attila_lendvai joined #sbcl
2017-10-18T16:02:18Z attila_lendvai quit (Ping timeout: 252 seconds)
2017-10-18T16:09:29Z attila_lendvai joined #sbcl
2017-10-18T16:15:58Z stassats joined #sbcl
2017-10-18T16:32:26Z shiyaz joined #sbcl
2017-10-18T16:32:26Z attila_lendvai quit (Read error: Connection reset by peer)
2017-10-18T16:33:03Z attila_lendvai joined #sbcl
2017-10-18T16:33:10Z attila_lendvai quit (Remote host closed the connection)
2017-10-18T16:34:27Z myrkraverk quit (Ping timeout: 240 seconds)
2017-10-18T16:34:53Z myrkraverk joined #sbcl
2017-10-18T16:50:00Z milanj joined #sbcl
2017-10-18T16:50:53Z shiyas joined #sbcl
2017-10-18T16:54:09Z m00natic quit (Remote host closed the connection)
2017-10-18T16:54:35Z shiyaz quit (Ping timeout: 240 seconds)
2017-10-18T16:59:37Z sjl joined #sbcl
2017-10-18T17:11:03Z Jesin quit (Quit: Leaving)
2017-10-18T17:15:17Z Jesin joined #sbcl
2017-10-18T17:20:21Z shka_ joined #sbcl
2017-10-18T17:20:30Z Jesin quit (Quit: Leaving)
2017-10-18T17:21:02Z Jesin joined #sbcl
2017-10-18T17:21:44Z shiyas quit (Remote host closed the connection)
2017-10-18T17:53:28Z stassats quit (Ping timeout: 240 seconds)
2017-10-18T18:00:51Z Jesin quit (Quit: Leaving)
2017-10-18T18:04:47Z Jesin joined #sbcl
2017-10-18T18:19:11Z milanj quit (Quit: This computer has gone to sleep)
2017-10-18T18:25:25Z hyuman joined #sbcl
2017-10-18T18:26:37Z hyuman is now known as whoman
2017-10-18T18:27:01Z Jesin quit (Quit: Leaving)
2017-10-18T18:30:46Z Jesin joined #sbcl
2017-10-18T18:38:18Z milanj joined #sbcl
2017-10-18T18:46:21Z karswell_ joined #sbcl
2017-10-18T18:46:55Z Bike quit (Ping timeout: 248 seconds)
2017-10-18T18:46:57Z rgrau joined #sbcl
2017-10-18T18:47:04Z Bike joined #sbcl
2017-10-18T18:48:01Z karswell quit (Read error: Connection reset by peer)
2017-10-18T18:52:12Z Bicyclidine joined #sbcl
2017-10-18T18:53:41Z Bike quit (Ping timeout: 240 seconds)
2017-10-18T19:29:03Z scymtym quit (Ping timeout: 246 seconds)
2017-10-18T19:43:29Z shka_ quit (Ping timeout: 246 seconds)
2017-10-18T19:46:18Z stassats joined #sbcl
2017-10-18T19:52:44Z whoman quit (Read error: Connection reset by peer)
2017-10-18T20:00:33Z leah2 quit (Ping timeout: 248 seconds)
2017-10-18T20:18:33Z leah2 joined #sbcl
2017-10-18T20:22:38Z stassats: process handling is slightly bonkers on windows
2017-10-18T20:22:43Z stassats: PID is not actually pid
2017-10-18T20:22:52Z edgar-rft joined #sbcl
2017-10-18T20:22:54Z stassats: doesn't close handles
2017-10-18T20:22:59Z stassats: process-close crashes
2017-10-18T20:30:53Z alandipert: stassats i could be mistaken but i think in win10+ handles are closed
2017-10-18T20:52:49Z angavrilov quit (Remote host closed the connection)
2017-10-18T21:10:13Z scymtym joined #sbcl
2017-10-18T21:14:09Z scymtym: stassats: i think there is a problem with continuable VALUES-LIST errors
2017-10-18T21:15:40Z scymtym: i noticed that the :RESTART-BOGUS-ARG-TO-VALUES-LIST-ERROR test is wrong (compares lists using =), but after fixing that, the restart didn't actually work
2017-10-18T21:18:02Z scymtym: manually defining BOGUS-ARG-TO-VALUES-LIST-ERROR and putting it into the internal errors handlers array makes the restart work again
2017-10-18T21:19:14Z Jesin quit (Quit: Leaving)
2017-10-18T21:26:58Z stassats: huh
2017-10-18T21:28:14Z scymtym: hm, now it seems to work in the fixed test
2017-10-18T21:28:24Z scymtym: i rebuilt since then. maybe something was hosed
2017-10-18T21:29:02Z stassats: (values-list '(1 . 2)) works here
2017-10-18T21:29:43Z scymtym: yes, everything i tried now worked as well
2017-10-18T21:29:56Z scymtym: so i'm just fixing the test
2017-10-18T21:30:02Z stassats: = gets continued
2017-10-18T21:30:08Z scymtym: apparently
2017-10-18T21:30:10Z stassats: that doesn't sound good
2017-10-18T21:30:17Z scymtym: otherwise i would fail
2017-10-18T21:30:21Z scymtym: *it
2017-10-18T21:31:01Z stassats: handler-bind should go inside the assert
2017-10-18T21:31:35Z scymtym: yes, i'm placing it just around the FUNCALL as to not affect anything else
2017-10-18T21:32:04Z scymtym: it didn't show the same behavior (continuing silently besides TYPE-ERROR from =) outside the test harness when i tried that
2017-10-18T21:32:19Z scymtym: that may be something to investigate as well
2017-10-18T21:33:15Z stassats: glibc renames __reserved in the arm mcontext to some other combination of __reserved, yet there is float simd state there
2017-10-18T21:33:22Z stassats: and no better name given
2017-10-18T21:33:23Z stassats: sigh
2017-10-18T21:34:29Z stassats: i think i cheat and go from the previous structure slot
2017-10-18T21:34:32Z stassats: ugh, so bad
2017-10-18T21:35:14Z scymtym: what could possibly go wrong
2017-10-18T21:36:06Z stassats: can i #ifdef on a structure slot or something?
2017-10-18T21:36:42Z scymtym: without compiling a test program?
2017-10-18T21:37:09Z scymtym: ok, when executing inside the test harness, there are a bunch of CONTINUE restarts available
2017-10-18T21:37:10Z joshe: fwiw, I did something similarly horrible on openbsd, in tools-for-build/openbsd-sigcontext.sh and the bottom of src/runtime/x86-bsd-os.c
2017-10-18T21:37:28Z stassats: there are now two different names __reserved and __glibc_reserved1
2017-10-18T21:37:55Z stassats: scymtym: yeah, i didn't want to name the condition
2017-10-18T21:38:04Z stassats: well, i want to have it named, but don't want to name it
2017-10-18T21:38:23Z scymtym: stassats: did you intend (= (funcall ...) 1) or (equal (funcall ...) '(1 ...))?
2017-10-18T21:38:38Z stassats: the latter
2017-10-18T21:39:24Z scymtym: ok, i will also think about a name for the condition. adding condition classes is one of the constants in my sbcl activity
2017-10-18T21:41:41Z scymtym: given the problem Xach discovered with ASDF 3.3.0, we should probably postpone updating until that is resolved
2017-10-18T21:43:32Z stassats: i've been fighting all these warnings as well, which got declared "asdf will work this way now"
2017-10-18T21:46:03Z stassats: i guess i was using mcontext_t from the kernel, ad now libc decided to add its own
2017-10-18T21:46:25Z scymtym: i like the progress ASDF has been making, but getting dozens of warnings because not everybody can or wants to keep up with the current ASDF best practices is indeed annoying
2017-10-18T21:47:45Z stassats: asdf is doing too many things
2017-10-18T21:47:49Z stassats: too complicated
2017-10-18T21:48:32Z stassats: i'm already predisposed to disliking it, changes in behavior just reinforce this
2017-10-18T21:49:57Z stassats: i can just hard-code __reserved
2017-10-18T21:50:39Z stassats: i don't want a shell script, an offset from the previous slot (since it's declared as aligned)
2017-10-18T21:50:55Z stassats: hard-coding is bad but does not turn the code into a mess
2017-10-18T22:03:32Z scymtym: stassats: does the COND in interr.lisp:233 have a hidden purpose?
2017-10-18T22:05:59Z Bicyclidine quit (Ping timeout: 252 seconds)
2017-10-18T22:06:05Z stassats: no
2017-10-18T22:06:19Z stassats: it's from restart-undefined
2017-10-18T22:07:55Z scymtym: ok
2017-10-18T22:22:29Z rgrau quit (Ping timeout: 252 seconds)
2017-10-18T22:24:33Z karswell_ quit (Ping timeout: 248 seconds)
2017-10-18T22:28:55Z stassats: trying to test whether this uc_mcontext even does what it's supposed to do
2017-10-18T22:29:05Z stassats: since arms do not have float traps, a bit hard to test
2017-10-18T22:29:30Z stassats: i need an error that refers to a simd register, but the compiler is eager to cons up a float
2017-10-18T22:38:31Z stassats: %primitive time...
2017-10-18T22:44:59Z stassats: hard coding struct fields turns out to be not as easy
2017-10-18T22:45:07Z stassats: with all the casts
2017-10-18T22:55:17Z cromachina joined #sbcl
2017-10-18T23:06:05Z Bike joined #sbcl
2017-10-18T23:06:50Z stassats: scymtym: where is still-bad defined?
2017-10-18T23:10:57Z Jesin joined #sbcl
2017-10-18T23:38:19Z scymtym: in the same respective LABELS
2017-10-18T23:38:43Z scymtym: sorry, gotta sleep now, i can revert tomorrow if something broke
2017-10-18T23:56:17Z sjl quit (Ping timeout: 248 seconds)
2017-10-19T00:08:22Z sjl joined #sbcl
2017-10-19T00:12:35Z sjl quit (Ping timeout: 240 seconds)
2017-10-19T00:37:25Z milanj quit (Quit: This computer has gone to sleep)
2017-10-19T00:56:33Z anticrisis joined #sbcl
2017-10-19T01:05:37Z stassats quit (Ping timeout: 248 seconds)
2017-10-19T01:05:45Z Ionic quit (Ping timeout: 246 seconds)
2017-10-19T01:10:55Z jack_rabbit quit (Ping timeout: 248 seconds)
2017-10-19T01:12:00Z milanj joined #sbcl
2017-10-19T01:26:01Z jack_rabbit joined #sbcl
2017-10-19T01:38:40Z Jesin quit (Quit: Leaving)
2017-10-19T02:08:17Z Jesin joined #sbcl
2017-10-19T02:50:06Z whoman joined #sbcl
2017-10-19T02:54:06Z shka_ joined #sbcl
2017-10-19T02:54:21Z myrkraverk quit (Remote host closed the connection)
2017-10-19T03:32:36Z Ionic joined #sbcl
2017-10-19T03:36:15Z myrkraverk joined #sbcl
2017-10-19T03:52:21Z Bike quit (Quit: Lost terminal)
2017-10-19T04:59:38Z brucem joined #sbcl
2017-10-19T06:06:03Z scymtym quit (Ping timeout: 258 seconds)
2017-10-19T06:44:39Z adlai uses dvorak instead of asdfg :P
2017-10-19T06:53:27Z shka_ quit (Ping timeout: 240 seconds)
2017-10-19T06:54:32Z carenz_ joined #sbcl
2017-10-19T07:02:23Z shka_ joined #sbcl
2017-10-19T07:25:05Z scymtym joined #sbcl
2017-10-19T07:34:27Z shka_ quit (Ping timeout: 240 seconds)
2017-10-19T07:55:05Z angavrilov joined #sbcl
2017-10-19T07:59:38Z milanj quit (Quit: This computer has gone to sleep)
2017-10-19T08:18:49Z rgrau joined #sbcl
2017-10-19T08:38:13Z scymtym_ joined #sbcl
2017-10-19T08:42:08Z scymtym quit (Ping timeout: 255 seconds)
2017-10-19T09:06:10Z anticrisis quit (Quit: sleep is overrated)
2017-10-19T09:51:24Z m00natic joined #sbcl
2017-10-19T10:08:50Z whoman: but qwerty is easier to spell on qwerty
2017-10-19T11:12:09Z sixbitslacker joined #sbcl
2017-10-19T11:37:15Z scymtym__ joined #sbcl
2017-10-19T11:41:14Z scymtym_ quit (Ping timeout: 255 seconds)
2017-10-19T12:01:18Z stassats joined #sbcl
2017-10-19T12:41:12Z Bike joined #sbcl
2017-10-19T13:04:29Z stassats quit (Ping timeout: 255 seconds)
2017-10-19T13:10:14Z stassats joined #sbcl
2017-10-19T13:46:55Z oleo joined #sbcl
2017-10-19T13:47:23Z oleo quit (Client Quit)
2017-10-19T13:48:29Z oleo joined #sbcl
2017-10-19T13:56:40Z shiyaz joined #sbcl
2017-10-19T14:10:19Z cromachina quit (Read error: Connection reset by peer)
2017-10-19T14:25:11Z hlavaty joined #sbcl
2017-10-19T14:38:59Z scymtym__ quit (Ping timeout: 255 seconds)
2017-10-19T15:37:43Z milanj joined #sbcl
2017-10-19T15:46:27Z carenz_ quit (Ping timeout: 240 seconds)
2017-10-19T15:55:55Z scymtym joined #sbcl
2017-10-19T16:06:05Z milanj quit (Quit: This computer has gone to sleep)
2017-10-19T16:07:26Z milanj joined #sbcl
2017-10-19T16:07:50Z Jesin quit (Quit: Leaving)
2017-10-19T16:10:37Z milanj quit (Client Quit)
2017-10-19T16:21:35Z tonton quit (Ping timeout: 240 seconds)
2017-10-19T16:23:41Z tonton joined #sbcl
2017-10-19T16:44:34Z sjl joined #sbcl
2017-10-19T16:50:07Z stassats quit (Ping timeout: 248 seconds)
2017-10-19T16:58:29Z oleo quit (Ping timeout: 258 seconds)
2017-10-19T16:58:35Z m00natic quit (Remote host closed the connection)
2017-10-19T17:00:06Z milanj joined #sbcl
2017-10-19T17:02:38Z oleo joined #sbcl
2017-10-19T17:10:18Z shka_ joined #sbcl
2017-10-19T17:40:15Z rgrau quit (Ping timeout: 248 seconds)
2017-10-19T17:44:46Z josemanuel[m] quit (Ping timeout: 264 seconds)
2017-10-19T18:26:38Z carenz_ joined #sbcl
2017-10-19T18:35:47Z hlavaty quit (Remote host closed the connection)
2017-10-19T18:43:36Z stassats joined #sbcl
2017-10-19T19:12:32Z milanj quit (Quit: This computer has gone to sleep)
2017-10-19T19:24:51Z sjl quit (Ping timeout: 255 seconds)
2017-10-19T19:44:48Z carenz_ quit (Ping timeout: 240 seconds)
2017-10-19T20:03:09Z sjl joined #sbcl
2017-10-19T20:28:44Z oleo quit (Quit: Leaving)
2017-10-19T20:48:37Z milanj joined #sbcl
2017-10-19T20:50:36Z angavrilov quit (Remote host closed the connection)
2017-10-19T21:25:51Z oleo joined #sbcl
2017-10-19T21:42:47Z sjl quit (Ping timeout: 252 seconds)
2017-10-19T21:50:25Z Bike quit (Ping timeout: 248 seconds)
2017-10-19T21:50:33Z karswell_ joined #sbcl
2017-10-19T21:57:58Z josemanuel[m] joined #sbcl
2017-10-19T21:58:47Z karswell_ is now known as karswell
2017-10-19T22:20:46Z Posterdati quit (Ping timeout: 255 seconds)
2017-10-19T22:21:36Z Posterdati joined #sbcl
2017-10-19T22:22:36Z Bike joined #sbcl
2017-10-19T22:41:50Z sjl joined #sbcl
2017-10-19T23:23:41Z shka_ quit (Ping timeout: 240 seconds)
2017-10-19T23:32:39Z cromachina joined #sbcl
2017-10-19T23:49:07Z Jesin joined #sbcl
2017-10-19T23:51:05Z scymtym: stassats: the fixed :RESTART-BOGUS-ARG-TO-VALUES-LIST-ERROR test with CHECK-COMPILE-AND-ASSERT iterating through policies causes a memory fault on x86: https://ci.cor-lab.org/job/sbcl-master/2555/featureset=1,label=ubuntu_trusty_32bit/consoleFull (search for "MEMORY-FAULT"). is it only supposed to work with certain policies?
2017-10-20T00:21:08Z stassats: scymtym: well, it's a type error
2017-10-20T00:21:48Z scymtym: so only test safe policies?
2017-10-20T00:22:35Z stassats: yes
2017-10-20T00:23:33Z scymtym: does x86_64 generate safe code despite the policy?
2017-10-20T00:26:23Z scymtym: no, i confused myself. the test only compiles with one policy, the default, i.e. safe one
2017-10-20T00:26:49Z scymtym: so adding VALUE-LIST-ARGUMENT-ERROR and using that in the caused the change
2017-10-20T00:27:54Z scymtym: either introducing the condition broke it or introducing the condition unmasked an error that was CONTINUEd over previously
2017-10-20T00:28:31Z scymtym: i will experiment on the x86 slave tomorrow
2017-10-20T00:28:34Z stassats: arch_skip_instruction can be broken
2017-10-20T00:32:31Z sjl quit (Ping timeout: 248 seconds)
2017-10-20T00:32:59Z crashtestdummy quit (Read error: Connection reset by peer)
2017-10-20T00:33:07Z Jesin quit (Quit: Leaving)
2017-10-20T00:35:33Z stassats: right
2017-10-20T00:35:55Z Jesin joined #sbcl
2017-10-20T00:37:11Z scymtym goes to bed
2017-10-20T00:38:40Z stassats: still using the old format, ugh
2017-10-20T00:39:27Z stassats: but shouldn't be doing anything but calling skip_internal_error
2017-10-20T00:41:43Z stassats: scymtym: fixed
2017-10-20T00:43:25Z stassats: i fear arch_skip_instruction will be broken everywhere but x86 and arm64
2017-10-20T00:44:40Z stassats: if the error is encoded in more than one instruction
2017-10-20T00:45:02Z Jesin quit (Quit: Leaving)
2017-10-20T00:47:19Z Jesin joined #sbcl
2017-10-20T00:47:56Z stassats: their loss
2017-10-20T00:48:06Z milanj quit (Quit: This computer has gone to sleep)
2017-10-20T00:50:49Z stassats: an easy fix could be calling skip_internal_error after arch_skip_instruction in interrupt_internal_error
2017-10-20T00:54:21Z jack_rabbit quit (Read error: Connection timed out)
2017-10-20T00:56:38Z stassats quit (Ping timeout: 246 seconds)
2017-10-20T00:58:29Z Bicyclidine joined #sbcl
2017-10-20T00:58:37Z Bicyclidine quit (Client Quit)
2017-10-20T01:11:48Z jack_rabbit joined #sbcl
2017-10-20T01:37:02Z oleo quit (Read error: Connection reset by peer)
2017-10-20T01:44:18Z milanj joined #sbcl
2017-10-20T01:49:56Z oleo joined #sbcl
2017-10-20T01:51:48Z live__ joined #sbcl
2017-10-20T01:52:02Z live__ quit (Remote host closed the connection)
2017-10-20T01:52:10Z oleo quit (Remote host closed the connection)
2017-10-20T01:52:55Z oleo joined #sbcl
2017-10-20T01:56:39Z milanj quit (Quit: This computer has gone to sleep)
2017-10-20T03:02:50Z oleo quit (Ping timeout: 255 seconds)
2017-10-20T03:38:27Z Bike quit (Quit: Lost terminal)
2017-10-20T04:36:25Z crashtestdummy joined #sbcl
2017-10-20T05:00:51Z shka_ joined #sbcl
2017-10-20T05:29:16Z angavrilov joined #sbcl
2017-10-20T05:52:12Z milanj joined #sbcl
2017-10-20T06:06:12Z scymtym quit (Ping timeout: 260 seconds)
2017-10-20T06:08:58Z carenz_ joined #sbcl
2017-10-20T07:03:46Z shka_ quit (Ping timeout: 252 seconds)
2017-10-20T07:29:43Z whoman quit (Quit: Leaving)
2017-10-20T07:34:31Z karswell quit (Read error: Connection reset by peer)
2017-10-20T07:48:52Z scymtym joined #sbcl
2017-10-20T08:42:05Z salva quit (Ping timeout: 240 seconds)
2017-10-20T08:57:27Z rgrau joined #sbcl
2017-10-20T09:15:48Z shiyaz quit (Quit: Leaving)
2017-10-20T09:16:11Z shiyaz joined #sbcl
2017-10-20T09:41:13Z stassats joined #sbcl
2017-10-20T09:56:46Z shiyas joined #sbcl
2017-10-20T09:58:19Z shiyas left #sbcl
2017-10-20T10:23:23Z milanj quit (Quit: This computer has gone to sleep)
2017-10-20T10:23:29Z oleo joined #sbcl
2017-10-20T10:25:15Z shiyaz quit (Quit: Leaving)
2017-10-20T10:48:08Z m00natic joined #sbcl
2017-10-20T12:33:21Z stassats quit (Ping timeout: 240 seconds)
2017-10-20T12:39:52Z Jesin quit (Quit: Leaving)
2017-10-20T12:41:22Z azrael_ joined #sbcl
2017-10-20T12:41:56Z azrael_ quit (Client Quit)
2017-10-20T12:46:36Z cromachina quit (Read error: Connection reset by peer)
2017-10-20T12:58:46Z Bike joined #sbcl
2017-10-20T13:17:42Z sixbitslacker left #sbcl
2017-10-20T14:20:00Z Jesin joined #sbcl
2017-10-20T14:32:45Z ebdreger joined #sbcl
2017-10-20T14:34:07Z carenz_ quit (Ping timeout: 248 seconds)
2017-10-20T14:34:47Z stassats joined #sbcl
2017-10-20T14:55:17Z whoman joined #sbcl
2017-10-20T15:12:18Z milanj joined #sbcl
2017-10-20T15:42:48Z scymtym: stassats: are you also doing WITH-SLOTS?
2017-10-20T16:11:35Z stassats: scymtym: no
2017-10-20T16:11:46Z stassats: with-accessors is particularly confusing, since it's not like with-slots
2017-10-20T16:12:01Z stassats: otherwise it's a rabbit hole of trying to fix all the macros and special forms
2017-10-20T16:14:36Z scymtym: it requires infrastructure
2017-10-20T16:15:25Z stassats: i want a different kind of infrastructure, "submit an error that confused you"
2017-10-20T16:15:50Z stassats: since what confuses us is a bit different
2017-10-20T16:16:29Z scymtym: that's true for error message
2017-10-20T16:17:17Z stassats: though i use with-accessors approximately never
2017-10-20T16:17:23Z scymtym: same
2017-10-20T16:18:17Z scymtym: but the hundreds of instances of (unless (and (consp form) (eq (first form) ...)) (error ...)) are basically ad-hoc solutions to a pattern matching or parsing problem. so a pattern matcher would be the right infrastructure
2017-10-20T16:18:40Z stassats: typep helps a bit
2017-10-20T16:18:41Z DeadTrickster quit (Ping timeout: 248 seconds)
2017-10-20T16:19:05Z dougk quit (Quit: Leaving)
2017-10-20T16:19:21Z scymtym: yes, but not with error messages and it doesn't bind match variables and it cannot handle repetition
2017-10-20T16:20:36Z stassats: (and the with-accessors fix was a response to https://www.reddit.com/r/lisp/comments/77cvt5/why_cant_i_car_my_accessor/ )
2017-10-20T16:21:13Z scymtym: i see
2017-10-20T16:21:20Z scymtym: i'm doing WITH-SLOTS
2017-10-20T16:21:39Z scymtym: and adding WITH-CURRENT-SOURCE-FORM so slime will highlight the right sub-form
2017-10-20T16:22:10Z scymtym: bummer that WITH-ACCESSORS doesn't work the way the person assumed
2017-10-20T16:22:39Z stassats: i've always found that that a bit strange
2017-10-20T16:25:30Z DeadTrickster joined #sbcl
2017-10-20T16:27:01Z _death: same
2017-10-20T16:28:14Z _death: maybe it was more prominent to use class-property naming style, and the variable name would just be property
2017-10-20T16:32:01Z dougk joined #sbcl
2017-10-20T16:50:12Z m00natic quit (Remote host closed the connection)
2017-10-20T16:57:22Z oleo quit (Remote host closed the connection)
2017-10-20T16:58:05Z oleo joined #sbcl
2017-10-20T17:05:29Z milanj quit (Quit: This computer has gone to sleep)
2017-10-20T17:08:45Z oleo quit (Remote host closed the connection)
2017-10-20T17:09:24Z oleo joined #sbcl
2017-10-20T17:10:45Z sjl joined #sbcl
2017-10-20T17:28:06Z milanj joined #sbcl
2017-10-20T17:28:31Z jack_rabbit quit (Ping timeout: 248 seconds)
2017-10-20T17:41:08Z oleo quit (Ping timeout: 246 seconds)
2017-10-20T17:42:01Z oleo joined #sbcl
2017-10-20T17:55:59Z shka_ joined #sbcl
2017-10-20T18:14:01Z rgrau quit (Ping timeout: 240 seconds)
2017-10-20T18:24:14Z milanj_ joined #sbcl
2017-10-20T18:27:57Z milanj quit (Ping timeout: 246 seconds)
2017-10-20T19:09:35Z jack_rabbit joined #sbcl
2017-10-20T19:15:11Z angavrilov quit (Remote host closed the connection)
2017-10-20T19:25:00Z scymtym quit (Ping timeout: 246 seconds)
2017-10-20T20:14:41Z scymtym joined #sbcl
2017-10-20T21:19:22Z zulu_inuoe quit (Read error: Connection reset by peer)
2017-10-20T21:32:17Z shka_ quit (Ping timeout: 248 seconds)
2017-10-20T21:32:41Z anticrisis joined #sbcl
2017-10-20T21:33:45Z anticrisis quit (Client Quit)
2017-10-20T21:50:57Z Bike quit (Ping timeout: 248 seconds)
2017-10-20T22:05:02Z cromachina joined #sbcl
2017-10-20T22:14:28Z Jesin quit (Quit: Leaving)
2017-10-20T22:26:09Z Bike joined #sbcl
2017-10-20T23:31:54Z milanj_ quit (Quit: This computer has gone to sleep)
2017-10-20T23:33:09Z milanj_ joined #sbcl
2017-10-20T23:42:52Z whoman quit (Read error: Connection reset by peer)
2017-10-20T23:46:01Z whoman joined #sbcl
2017-10-20T23:51:18Z milanj_ quit (Quit: This computer has gone to sleep)
2017-10-21T00:16:43Z milanj_ joined #sbcl
2017-10-21T00:31:46Z whoman quit (Ping timeout: 258 seconds)
2017-10-21T00:35:19Z jibanes quit (Ping timeout: 255 seconds)
2017-10-21T00:36:56Z jibanes joined #sbcl
2017-10-21T00:57:35Z jibanes quit (Ping timeout: 240 seconds)
2017-10-21T00:59:36Z jibanes joined #sbcl
2017-10-21T01:01:17Z stassats quit (Ping timeout: 258 seconds)
2017-10-21T01:02:41Z oleo quit (Ping timeout: 255 seconds)
2017-10-21T01:03:23Z milanj_ quit (Quit: This computer has gone to sleep)
2017-10-21T01:35:31Z milanj_ joined #sbcl
2017-10-21T01:35:59Z Jesin joined #sbcl
2017-10-21T01:54:38Z sjl quit (Ping timeout: 246 seconds)
2017-10-21T03:09:57Z panji joined #sbcl
2017-10-21T03:28:14Z |3b|` joined #sbcl
2017-10-21T03:29:28Z |3b| quit (Ping timeout: 255 seconds)
2017-10-21T03:30:57Z panji quit (Ping timeout: 240 seconds)
2017-10-21T03:34:59Z |3b|`` joined #sbcl
2017-10-21T03:36:08Z |3b|` quit (Ping timeout: 255 seconds)
2017-10-21T03:41:42Z whoman joined #sbcl
2017-10-21T03:43:11Z panji joined #sbcl
2017-10-21T04:23:49Z shka_ joined #sbcl
2017-10-21T04:34:29Z Bike quit (Quit: Lost terminal)
2017-10-21T05:44:29Z sjl joined #sbcl
2017-10-21T06:29:04Z angavrilov joined #sbcl
2017-10-21T06:46:07Z panji quit (Ping timeout: 255 seconds)
2017-10-21T06:53:53Z sjl quit (Ping timeout: 248 seconds)
2017-10-21T07:04:59Z sjl joined #sbcl
2017-10-21T07:09:53Z sjl quit (Ping timeout: 248 seconds)
2017-10-21T07:20:01Z shka_ quit (Ping timeout: 248 seconds)
2017-10-21T08:02:26Z milanj_ quit (Quit: This computer has gone to sleep)
2017-10-21T09:05:49Z sjl joined #sbcl
2017-10-21T09:10:25Z sjl quit (Ping timeout: 248 seconds)
2017-10-21T09:48:58Z scymtym quit (Ping timeout: 264 seconds)
2017-10-21T09:51:28Z sjl joined #sbcl
2017-10-21T09:55:57Z sjl quit (Ping timeout: 240 seconds)
2017-10-21T10:26:52Z scymtym joined #sbcl
2017-10-21T10:38:47Z foom quit (Read error: Connection reset by peer)
2017-10-21T10:39:02Z foom joined #sbcl
2017-10-21T10:48:17Z oleo joined #sbcl
2017-10-21T10:48:27Z DeadTrickster quit (Remote host closed the connection)
2017-10-21T11:19:21Z faraco joined #sbcl
2017-10-21T11:30:17Z DeadTrickster joined #sbcl
2017-10-21T11:31:13Z sjl joined #sbcl
2017-10-21T11:36:01Z sjl quit (Ping timeout: 248 seconds)
2017-10-21T11:38:16Z Bike joined #sbcl
2017-10-21T11:41:47Z faraco quit (Quit: Leaving)
2017-10-21T12:49:05Z aeth quit (Read error: Connection reset by peer)
2017-10-21T12:54:17Z aeth joined #sbcl
2017-10-21T13:22:21Z stassats joined #sbcl
2017-10-21T13:32:05Z sjl joined #sbcl
2017-10-21T13:36:47Z sjl quit (Ping timeout: 260 seconds)
2017-10-21T15:02:35Z shka_ joined #sbcl
2017-10-21T15:06:55Z DeadTrickster quit (Remote host closed the connection)
2017-10-21T15:32:49Z sjl joined #sbcl
2017-10-21T15:37:24Z sjl quit (Ping timeout: 246 seconds)
2017-10-21T16:02:21Z tonton quit (Ping timeout: 240 seconds)
2017-10-21T16:04:28Z tonton joined #sbcl
2017-10-21T16:21:22Z milanj_ joined #sbcl
2017-10-21T16:26:24Z scymtym quit (Ping timeout: 246 seconds)
2017-10-21T16:52:49Z scymtym joined #sbcl
2017-10-21T18:01:06Z xristos joined #sbcl
2017-10-21T18:51:54Z bkst joined #sbcl
2017-10-21T20:08:38Z stassats: what if transform-call doesn't perform full lambda transformation but directly transforms into a let
2017-10-21T20:09:18Z stassats: which would save a lot of intermediate steps for transforms with &key
2017-10-21T20:17:04Z stassats: unknown key args would be a problem, but that rarely happens, could fall back the full convert-lambda
2017-10-21T20:19:27Z angavrilov quit (Remote host closed the connection)
2017-10-21T20:37:53Z stassats: though even LET isn't transformed into a let-call
2017-10-21T21:06:26Z DeadTrickster joined #sbcl
2017-10-21T21:15:59Z DeadTrickster quit (Remote host closed the connection)
2017-10-21T21:23:26Z DeadTrickster joined #sbcl
2017-10-21T21:32:10Z shka_ quit (Ping timeout: 255 seconds)
2017-10-21T21:37:31Z milanj_ quit (Quit: This computer has gone to sleep)
2017-10-21T22:27:54Z scymtym: stassats: couldn't the sb comparison at the end of PACK-WIRED-TN been done as (eq sb (l-t-v (sb-or-loose 'control-stack)))? (saw the STRING= while making sb-sprof flamegraph in mcclim)
2017-10-21T22:33:27Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-21T22:33:51Z DeadTrickster joined #sbcl
2017-10-21T22:33:52Z stassats: what is it even about?
2017-10-21T22:35:35Z stassats: :specified-save is not even used on these backends
2017-10-21T22:36:35Z scymtym: wouldn't that also be an AVER if anything?
2017-10-21T22:37:11Z stassats: avers are mute
2017-10-21T22:39:52Z scymtym: right, a BUG then. i mean, it seems like a consistency check and an internal error if violated
2017-10-21T22:40:52Z stassats: and these correspond to return-pc-passing-offset and old-fp-passing-offset
2017-10-21T22:41:17Z stassats: which are actually sc-offsets, but that was too long to type out
2017-10-21T22:49:02Z stassats: i'll change it
2017-10-21T22:56:01Z eschatologist quit (Ping timeout: 240 seconds)
2017-10-21T23:03:59Z scymtym: just built successfully with sc-locations changed to (simple-array sc-offset 1). that should help in a couple of places
2017-10-21T23:06:43Z milanj joined #sbcl
2017-10-21T23:07:38Z stassats: was string= called a lot?
2017-10-21T23:07:51Z stassats: just running the tests and pushing it out
2017-10-21T23:11:25Z scymtym: stassats: https://techfak.de/~jmoringe/pack-flamegraph.png
2017-10-21T23:11:56Z stassats: enough to fit the name string=*?
2017-10-21T23:12:53Z scymtym: mainly because it's the most expensive thing in PACK-WIRED-TN
2017-10-21T23:13:35Z scymtym: SELECT-LOCATION also seems to have low hanging fruit
2017-10-21T23:14:08Z stassats: scymtym: try now
2017-10-21T23:14:35Z scymtym: building
2017-10-21T23:14:48Z stassats: i think the whole way we ir1-optimize combinations is wrong
2017-10-21T23:14:52Z stassats: and transforms
2017-10-21T23:15:28Z stassats: when a transform is inserted, it's a call, it gets let converted quickly, but no propagate-let-args is called
2017-10-21T23:16:19Z stassats: then ir1-optimize continues optimizing the newly converted code, but the types are not propagated, so it either has to call the transforms again or get transformed into suboptimal code
2017-10-21T23:17:44Z stassats: i think transforms should only be performed after the whole component has all the types propagated
2017-10-21T23:17:55Z stassats: and even after constraintprop
2017-10-21T23:19:08Z stassats: or a lot of them need delay-ir1-transform :constraint
2017-10-21T23:19:58Z stassats: but the issue of transform-call leaving propagated code, i think after anything is ir1 converted, ir1-optimize should just stop and run again
2017-10-21T23:20:35Z stassats: since there are guaranteed to be nodes with reoptimize=t, and it'll will run again, but potentially twice
2017-10-21T23:20:45Z stassats: aborting early can actually reduce the number of ir1-optimize calls
2017-10-21T23:21:33Z stassats: i tried calling ir1-optimize-combination in transform-call, but that didn't go well on deleted calls
2017-10-21T23:22:19Z stassats: i guess i'll need some code to time compilation against
2017-10-21T23:22:25Z stassats: scymtym: what are you using?
2017-10-21T23:24:07Z stassats: i think i need something with lots of blocks and lots of transforms, so, branchy with array access and arithmetic
2017-10-21T23:28:40Z scymtym: stassats: https://techfak.de/~jmoringe/pack-flamegraph-new.png PACK-WIRED-TN is now one of the unlabeled blips which is good
2017-10-21T23:29:07Z stassats: and the overall timing?
2017-10-21T23:29:16Z stassats: and compiling what?
2017-10-21T23:29:48Z scymtym: i didn't measure that. this is just sb-sprof visualization, i.e. relative amounts of time spent
2017-10-21T23:30:00Z scymtym: code was a trivial function
2017-10-21T23:33:04Z scymtym: as far as test cases go, i haven't found a canonical example that exercises everything
2017-10-21T23:36:18Z stassats: scymtym: CDR?
2017-10-21T23:36:46Z stassats: from stable-sort #'cdr, i guess
2017-10-21T23:40:10Z scymtym: stassats: SCHWARTZIAN-STABLE-SORT-{LIST,VECTOR} full calls CDR
2017-10-21T23:41:14Z Jesin quit (Quit: Leaving)
2017-10-21T23:41:22Z scymtym: could try to (declare (inline stable-sort)) in SCHWARTZIAN-STABLE-SORT-*
2017-10-21T23:42:27Z scymtym: oh, no, it is already transformed
2017-10-21T23:44:57Z scymtym: i wonder whether the whole sorting in SELECT-LOCATION can be cached
2017-10-21T23:45:04Z scymtym: or precomputed
2017-10-21T23:45:36Z scymtym: the combinations are SC-OFFSET \times SC
2017-10-21T23:48:06Z eschatologist joined #sbcl
2017-10-22T00:02:17Z brucem quit (Ping timeout: 252 seconds)
2017-10-22T00:12:05Z brucem joined #sbcl
2017-10-22T00:14:19Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-22T00:15:13Z DeadTrickster joined #sbcl
2017-10-22T00:16:35Z brucem quit (Ping timeout: 252 seconds)
2017-10-22T00:19:03Z brucem joined #sbcl
2017-10-22T00:29:14Z brucem quit (Ping timeout: 252 seconds)
2017-10-22T00:43:13Z milanj quit (Quit: This computer has gone to sleep)
2017-10-22T00:43:41Z milanj joined #sbcl
2017-10-22T00:45:35Z brucem joined #sbcl
2017-10-22T00:55:45Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-22T00:56:11Z DeadTrickster joined #sbcl
2017-10-22T01:03:27Z oleo quit (Remote host closed the connection)
2017-10-22T01:04:03Z oleo joined #sbcl
2017-10-22T01:23:12Z stassats quit (Ping timeout: 260 seconds)
2017-10-22T02:08:21Z stassats joined #sbcl
2017-10-22T03:15:05Z stassats quit (Ping timeout: 240 seconds)
2017-10-22T04:39:35Z sjl_ joined #sbcl
2017-10-22T04:48:47Z guicho joined #sbcl
2017-10-22T04:55:53Z Bike quit (Quit: Lost terminal)
2017-10-22T07:37:30Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-22T07:38:12Z DeadTrickster joined #sbcl
2017-10-22T08:21:39Z attila_lendvai joined #sbcl
2017-10-22T08:21:39Z DeadTrickster quit (Remote host closed the connection)
2017-10-22T08:55:39Z attila_lendvai quit (Ping timeout: 252 seconds)
2017-10-22T08:57:21Z leah2 quit (Ping timeout: 240 seconds)
2017-10-22T09:11:56Z leah2 joined #sbcl
2017-10-22T09:12:17Z oleo quit (Ping timeout: 255 seconds)
2017-10-22T09:37:06Z shka_ joined #sbcl
2017-10-22T09:43:48Z scymtym quit (Ping timeout: 240 seconds)
2017-10-22T10:02:52Z scymtym joined #sbcl
2017-10-22T10:03:14Z milanj quit (Quit: This computer has gone to sleep)
2017-10-22T10:23:35Z guicho quit (Ping timeout: 240 seconds)
2017-10-22T10:44:14Z oleo joined #sbcl
2017-10-22T11:11:03Z Bike joined #sbcl
2017-10-22T11:12:16Z angavrilov joined #sbcl
2017-10-22T11:12:32Z fiveop joined #sbcl
2017-10-22T12:25:25Z fiveop quit
2017-10-22T12:55:54Z stassats joined #sbcl
2017-10-22T13:55:32Z scymtym: stassats: (compile :node/lvar :derive-type :misc.4) in compiler.pure.lisp started to fail because a WARNING is signaled now (used to be a STYLE-WARNING)
2017-10-22T14:28:19Z stassats: scymtym: yeah, i have a fix waiting
2017-10-22T14:28:50Z stassats: some editing problems yesterday, was supposed to go in together with the change
2017-10-22T14:29:28Z stassats: in other test news, nlx-causes-alien-stack-leak started failing on -c-stack-is-control-stack
2017-10-22T14:29:49Z stassats: yet the code fails on even older versions
2017-10-22T14:29:52Z stassats: so somehow it was masked
2017-10-22T14:30:42Z stassats: ;; No need for unwind protect here, since allocation involves modifying NSP, and NSP is saved and restored during NLX. And in non-transformed case it performs finalization.
2017-10-22T14:30:55Z stassats: that's not really the case...
2017-10-22T14:32:29Z stassats: lp 645121
2017-10-22T14:32:29Z specbot: https://bugs.launchpad.net/bugs/645121
2017-10-22T14:34:41Z stassats: that's an incorrect assumption, if i THROW from within with-alien, the CATCH will indeed restore NSP, but not a return-from
2017-10-22T14:38:04Z stassats: an unwind-protect slows down the code quite a bit
2017-10-22T14:40:34Z stassats: but this is basically akin to UNBIND, no full blown UWP needed, because of the stated reasons
2017-10-22T14:41:05Z stassats: it'll get cleaned up on unwinds, just need to reset NSP on non-local exits
2017-10-22T14:41:25Z stassats: i just need to introduce a new cleanup form kind, save NSP and clean it up
2017-10-22T14:41:35Z stassats: ILTWIS"just"
2017-10-22T14:44:23Z rgrau joined #sbcl
2017-10-22T14:44:31Z stassats: if done right, even the non NLX case will get faster, since multiple NSP restores can be rolled into one
2017-10-22T16:02:39Z milanj joined #sbcl
2017-10-22T16:17:05Z DeadTrickster joined #sbcl
2017-10-22T16:25:45Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-22T16:26:11Z DeadTrickster joined #sbcl
2017-10-22T16:41:05Z scymtym quit (Ping timeout: 246 seconds)
2017-10-22T16:48:18Z rgrau quit (Ping timeout: 258 seconds)
2017-10-22T16:54:58Z scymtym joined #sbcl
2017-10-22T16:59:44Z stassats: nsp-protect is proving to be difficult
2017-10-22T16:59:57Z stassats: can't easily pass the saved lvar through all the ir stages
2017-10-22T17:14:26Z eschatologist quit (Remote host closed the connection)
2017-10-22T17:21:55Z eschatologist joined #sbcl
2017-10-22T17:47:17Z wheelsucker joined #sbcl
2017-10-22T17:56:34Z carenz_ joined #sbcl
2017-10-22T17:57:09Z stassats: after adjusting ir1-convert-combination-args to use lvar, i think i got it
2017-10-22T18:00:34Z josemanuel joined #sbcl
2017-10-22T18:10:41Z stassats quit (Ping timeout: 240 seconds)
2017-10-22T18:28:29Z eschatologist quit (Remote host closed the connection)
2017-10-22T18:29:52Z eschatologist joined #sbcl
2017-10-22T18:46:57Z carenz_ quit (Ping timeout: 248 seconds)
2017-10-22T18:47:02Z Jesin joined #sbcl
2017-10-22T18:53:38Z josemanuel[m] quit (Ping timeout: 252 seconds)
2017-10-22T19:24:47Z stassats joined #sbcl
2017-10-22T19:35:00Z Jesin quit (Quit: Leaving)
2017-10-22T20:00:51Z milanj quit (Quit: This computer has gone to sleep)
2017-10-22T20:20:51Z stassats: huh, the same clean up may be emitted twice
2017-10-22T20:21:13Z stassats: that breaks my solution, sigh
2017-10-22T20:25:13Z stassats: (defun x (a) (let (*) (if a 10 20))) unbinds twice
2017-10-22T20:26:23Z stassats: not running twice, but two sequences doing the same thing
2017-10-22T20:32:30Z angavrilov quit (Remote host closed the connection)
2017-10-22T21:00:06Z stassats: tricky
2017-10-22T21:05:52Z rgrau joined #sbcl
2017-10-22T21:10:09Z shka_ quit (Ping timeout: 246 seconds)
2017-10-22T21:28:41Z stassats: saved by the code freeze
2017-10-22T21:28:43Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-22T21:29:14Z DeadTrickster joined #sbcl
2017-10-22T21:45:56Z josemanuel quit (Quit: leaving)
2017-10-22T21:46:05Z jackdaniel quit (Ping timeout: 240 seconds)
2017-10-22T21:48:48Z stassats: rethinking the whole approach, instead of trying to make a really specific nsp-protect, maybe i can have simple-unwind-protect, which acts like special bindings, without establishing catch or unwind blocks, just executing code on RETURN-FROM
2017-10-22T21:48:55Z stassats: might be useful for something else
2017-10-22T22:09:55Z scymtym quit (Remote host closed the connection)
2017-10-22T22:16:01Z scymtym joined #sbcl
2017-10-22T22:30:28Z scymtym quit (Remote host closed the connection)
2017-10-22T22:31:03Z stassats: if i could figure out how to do that...
2017-10-22T22:34:29Z scymtym joined #sbcl
2017-10-22T22:36:28Z oleo quit (Remote host closed the connection)
2017-10-22T22:37:00Z oleo joined #sbcl
2017-10-22T23:26:39Z stassats quit (Ping timeout: 246 seconds)
2017-10-23T02:12:35Z brucem quit (Changing host)
2017-10-23T02:12:35Z brucem joined #sbcl
2017-10-23T02:25:05Z rgrau quit (Ping timeout: 240 seconds)
2017-10-23T03:00:03Z Bike quit (Quit: Lost terminal)
2017-10-23T03:35:43Z wheelsucker quit (Quit: Client Quit)
2017-10-23T05:10:43Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-23T05:11:05Z DeadTrickster joined #sbcl
2017-10-23T05:28:05Z shka_ joined #sbcl
2017-10-23T05:34:49Z milanj joined #sbcl
2017-10-23T06:16:48Z scymtym quit (Ping timeout: 240 seconds)
2017-10-23T06:30:11Z salva joined #sbcl
2017-10-23T06:46:22Z angavrilov joined #sbcl
2017-10-23T06:51:55Z bkst quit (Quit: leaving)
2017-10-23T07:19:50Z attila_lendvai joined #sbcl
2017-10-23T07:19:50Z attila_lendvai quit (Changing host)
2017-10-23T07:19:50Z attila_lendvai joined #sbcl
2017-10-23T07:28:32Z oleo quit (Quit: Leaving)
2017-10-23T07:29:12Z scymtym joined #sbcl
2017-10-23T07:32:29Z scymtym: minion: memo for stassats: (assert-error (funcall (checked-compile FORM) ARGS) CONDITION) is (checked-compile-and-assert () FORM (ARGS (condition CONDITION))). the benefit being iteration through policies and the ability to check multiple expected conditions, possibly mixed with expected return values, in one go
2017-10-23T07:32:29Z minion: Remembered. I'll tell stassats when he/she/it next speaks.
2017-10-23T07:52:13Z carenz_ joined #sbcl
2017-10-23T08:31:15Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-23T08:32:43Z DeadTrickster joined #sbcl
2017-10-23T08:33:56Z jackdaniel joined #sbcl
2017-10-23T09:03:38Z jackdaniel quit (Quit: reboot)
2017-10-23T09:25:31Z jackdaniel joined #sbcl
2017-10-23T09:25:34Z m00natic joined #sbcl
2017-10-23T09:46:17Z jackdaniel quit (Remote host closed the connection)
2017-10-23T09:49:15Z jackdaniel joined #sbcl
2017-10-23T10:21:44Z rtmpdavid joined #sbcl
2017-10-23T11:33:41Z scymtym quit (Ping timeout: 246 seconds)
2017-10-23T12:15:54Z scymtym joined #sbcl
2017-10-23T12:42:16Z stassats joined #sbcl
2017-10-23T12:47:58Z sjl joined #sbcl
2017-10-23T12:51:46Z Bike joined #sbcl
2017-10-23T12:58:34Z sjl quit (Ping timeout: 264 seconds)
2017-10-23T13:20:05Z cromachina quit (Read error: Connection reset by peer)
2017-10-23T13:34:13Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-23T13:34:50Z DeadTrickster joined #sbcl
2017-10-23T13:54:45Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-23T13:55:23Z DeadTrickster joined #sbcl
2017-10-23T14:23:15Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-23T14:24:01Z DeadTrickster joined #sbcl
2017-10-23T14:37:05Z scymtym quit (Ping timeout: 246 seconds)
2017-10-23T14:38:53Z oleo joined #sbcl
2017-10-23T14:51:25Z Jesin joined #sbcl
2017-10-23T14:59:44Z stassats: ok
2017-10-23T14:59:44Z minion: stassats, memo from scymtym: (assert-error (funcall (checked-compile FORM) ARGS) CONDITION) is (checked-compile-and-assert () FORM (ARGS (condition CONDITION))). the benefit being iteration through policies and the ability to check multiple expected conditions, possibly mixed with expected return values, in one go
2017-10-23T15:06:14Z gargaml joined #sbcl
2017-10-23T15:08:32Z milanj quit (Quit: This computer has gone to sleep)
2017-10-23T15:10:30Z milanj joined #sbcl
2017-10-23T15:16:09Z attila_lendvai quit (Ping timeout: 252 seconds)
2017-10-23T15:24:45Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-23T15:25:10Z DeadTrickster joined #sbcl
2017-10-23T15:28:02Z milanj quit (Quit: This computer has gone to sleep)
2017-10-23T16:01:35Z crashtestdummy quit (Ping timeout: 248 seconds)
2017-10-23T16:03:15Z scymtym joined #sbcl
2017-10-23T16:34:55Z m00natic quit (Read error: Connection reset by peer)
2017-10-23T16:39:41Z milanj joined #sbcl
2017-10-23T17:07:56Z gargaml quit (Quit: WeeChat 1.9)
2017-10-23T17:21:04Z carenz_ quit (Ping timeout: 255 seconds)
2017-10-23T18:47:06Z akovalenko quit (Remote host closed the connection)
2017-10-23T18:47:30Z akovalenko joined #sbcl
2017-10-23T19:31:32Z sjl joined #sbcl
2017-10-23T19:36:05Z sjl quit (Ping timeout: 255 seconds)
2017-10-23T19:42:59Z DGASAU quit (Read error: Connection reset by peer)
2017-10-23T19:43:23Z DGASAU joined #sbcl
2017-10-23T19:53:10Z jdz quit (Ping timeout: 255 seconds)
2017-10-23T19:53:56Z stassats: after ten tries, i fixed nlx-causes-alien-stack-leak on -c-stack-is-control-stack without using unwind-protect
2017-10-23T19:55:15Z stassats: all the trouble with trying to preserve a variable for code that might be emitted in the future
2017-10-23T19:59:00Z jdz joined #sbcl
2017-10-23T20:11:15Z shka_ quit (Ping timeout: 246 seconds)
2017-10-23T20:27:10Z Jesin quit (Quit: Leaving)
2017-10-23T20:52:47Z angavrilov quit (Remote host closed the connection)
2017-10-23T21:07:59Z milanj quit (Quit: This computer has gone to sleep)
2017-10-23T21:42:25Z myrkraverk_ joined #sbcl
2017-10-23T21:42:26Z milanj joined #sbcl
2017-10-23T21:42:55Z myrkraverk quit (Ping timeout: 248 seconds)
2017-10-23T21:44:13Z myrkraverk_ is now known as myrkraverk
2017-10-23T21:45:58Z Bike quit (Ping timeout: 264 seconds)
2017-10-23T21:57:46Z karswell joined #sbcl
2017-10-23T22:25:08Z Bike joined #sbcl
2017-10-23T22:47:38Z cromachina joined #sbcl
2017-10-23T23:06:41Z stassats quit (Ping timeout: 248 seconds)
2017-10-24T00:25:41Z milanj quit (Quit: This computer has gone to sleep)
2017-10-24T00:25:52Z milanj joined #sbcl
2017-10-24T00:25:57Z milanj quit (Client Quit)
2017-10-24T01:17:39Z edgar-rft quit (Quit: edgar-rft)
2017-10-24T01:19:56Z edgar-rft joined #sbcl
2017-10-24T03:08:57Z Bike quit (Quit: gnight)
2017-10-24T04:21:28Z shka_ joined #sbcl
2017-10-24T04:34:07Z karswell_ joined #sbcl
2017-10-24T04:34:27Z karswell quit (Ping timeout: 240 seconds)
2017-10-24T04:44:54Z oleo quit (Quit: Leaving)
2017-10-24T05:38:41Z sjl_ quit (Ping timeout: 240 seconds)
2017-10-24T06:25:05Z scymtym quit (Ping timeout: 248 seconds)
2017-10-24T07:07:29Z angavrilov joined #sbcl
2017-10-24T07:22:39Z shka_ quit (Ping timeout: 248 seconds)
2017-10-24T07:48:38Z scymtym joined #sbcl
2017-10-24T07:50:33Z carenz_ joined #sbcl
2017-10-24T08:37:55Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-24T08:38:17Z DeadTrickster joined #sbcl
2017-10-24T08:44:20Z milanj joined #sbcl
2017-10-24T09:16:27Z m00natic joined #sbcl
2017-10-24T11:24:43Z Bike joined #sbcl
2017-10-24T11:32:44Z Bicyclidine joined #sbcl
2017-10-24T11:34:02Z Bike quit (Ping timeout: 255 seconds)
2017-10-24T11:51:08Z Bicyclidine quit (Ping timeout: 255 seconds)
2017-10-24T12:05:16Z milanj quit (Quit: This computer has gone to sleep)
2017-10-24T12:25:00Z gargaml joined #sbcl
2017-10-24T12:56:26Z attila_lendvai joined #sbcl
2017-10-24T12:56:26Z attila_lendvai quit (Changing host)
2017-10-24T12:56:26Z attila_lendvai joined #sbcl
2017-10-24T13:08:47Z Bike joined #sbcl
2017-10-24T13:32:14Z sjl joined #sbcl
2017-10-24T13:41:55Z cromachina quit (Read error: Connection reset by peer)
2017-10-24T14:33:22Z DeadTrickster quit (Remote host closed the connection)
2017-10-24T15:16:30Z milanj joined #sbcl
2017-10-24T15:34:13Z oleo joined #sbcl
2017-10-24T15:45:52Z carenz_ quit (Ping timeout: 260 seconds)
2017-10-24T15:59:49Z milanj quit (Read error: Connection reset by peer)
2017-10-24T16:00:12Z milanj joined #sbcl
2017-10-24T16:00:32Z libre-man quit (Ping timeout: 255 seconds)
2017-10-24T16:12:54Z libre-man joined #sbcl
2017-10-24T16:16:18Z milanj quit (Quit: This computer has gone to sleep)
2017-10-24T16:40:47Z DeadTrickster joined #sbcl
2017-10-24T16:51:09Z sjl_ joined #sbcl
2017-10-24T16:51:58Z sjl quit (Ping timeout: 264 seconds)
2017-10-24T16:55:55Z sjl_ is now known as sjl
2017-10-24T17:01:28Z m00natic quit (Read error: Connection reset by peer)
2017-10-24T17:02:01Z DGASAU quit (Remote host closed the connection)
2017-10-24T17:03:02Z DGASAU joined #sbcl
2017-10-24T17:03:31Z gargaml quit (Ping timeout: 255 seconds)
2017-10-24T17:03:36Z myrkraverk_ joined #sbcl
2017-10-24T17:05:14Z myrkraverk quit (Ping timeout: 255 seconds)
2017-10-24T17:06:07Z myrkraverk_ is now known as myrkraverk
2017-10-24T17:07:45Z attila_lendvai quit (Quit: Leaving.)
2017-10-24T17:10:47Z Jesin joined #sbcl
2017-10-24T17:30:05Z milanj joined #sbcl
2017-10-24T17:35:27Z shka_ joined #sbcl
2017-10-24T17:40:36Z stassats joined #sbcl
2017-10-24T17:41:04Z Chream joined #sbcl
2017-10-24T17:53:44Z sjl_ joined #sbcl
2017-10-24T17:56:10Z sjl_ quit (Client Quit)
2017-10-24T18:18:24Z edgar-rft quit (Quit: edgar-rft)
2017-10-24T18:19:53Z milanj quit (Quit: This computer has gone to sleep)
2017-10-24T18:24:44Z angavrilov quit (Remote host closed the connection)
2017-10-24T18:28:17Z karswell_ quit (Ping timeout: 248 seconds)
2017-10-24T19:02:59Z milanj joined #sbcl
2017-10-24T19:19:57Z shka_ quit (Ping timeout: 260 seconds)
2017-10-24T19:31:19Z Jesin quit (Quit: Leaving)
2017-10-24T19:33:25Z scymtym quit (Remote host closed the connection)
2017-10-24T19:33:42Z scymtym joined #sbcl
2017-10-24T19:34:08Z Chream quit (Ping timeout: 246 seconds)
2017-10-24T19:39:23Z scymtym quit (Ping timeout: 246 seconds)
2017-10-24T20:02:53Z Chream joined #sbcl
2017-10-24T20:04:06Z milanj quit (Quit: This computer has gone to sleep)
2017-10-24T20:19:07Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-24T20:19:35Z DeadTrickster joined #sbcl
2017-10-24T20:26:21Z scymtym joined #sbcl
2017-10-24T20:37:34Z edgar-rft joined #sbcl
2017-10-24T20:52:41Z milanj joined #sbcl
2017-10-24T20:55:35Z Jesin joined #sbcl
2017-10-24T21:04:20Z Jesin quit (Quit: Leaving)
2017-10-24T21:08:05Z Bike quit (Ping timeout: 240 seconds)
2017-10-24T21:40:30Z Chream_ joined #sbcl
2017-10-24T21:44:41Z Chream_ quit (Read error: Connection reset by peer)
2017-10-24T21:45:49Z Chream_ joined #sbcl
2017-10-24T21:47:25Z Bike joined #sbcl
2017-10-24T21:51:22Z attila_lendvai joined #sbcl
2017-10-24T21:51:22Z attila_lendvai quit (Changing host)
2017-10-24T21:51:22Z attila_lendvai joined #sbcl
2017-10-24T21:52:57Z Posterdati quit (Ping timeout: 240 seconds)
2017-10-24T21:57:21Z sjl quit (Ping timeout: 248 seconds)
2017-10-24T22:06:13Z Posterdati joined #sbcl
2017-10-24T22:10:47Z stassats quit (Ping timeout: 255 seconds)
2017-10-24T22:15:06Z attila_lendvai quit (Quit: Leaving.)
2017-10-24T22:18:04Z attila_lendvai joined #sbcl
2017-10-24T22:18:04Z attila_lendvai quit (Changing host)
2017-10-24T22:18:04Z attila_lendvai joined #sbcl
2017-10-24T22:22:16Z attila_lendvai quit (Client Quit)
2017-10-24T22:32:47Z Lord_Nightmare quit (Quit: ZNC - http://znc.in)
2017-10-24T22:33:20Z Jesin joined #sbcl
2017-10-24T22:35:18Z Chream quit (Remote host closed the connection)
2017-10-24T22:36:48Z Posterdati quit (Ping timeout: 240 seconds)
2017-10-24T22:56:29Z stassats joined #sbcl
2017-10-24T23:04:19Z Lord_Nightmare joined #sbcl
2017-10-24T23:30:56Z cromachina joined #sbcl
2017-10-24T23:42:43Z milanj quit (Quit: This computer has gone to sleep)
2017-10-25T00:45:42Z milanj joined #sbcl
2017-10-25T00:56:29Z milanj quit (Quit: This computer has gone to sleep)
2017-10-25T00:58:57Z Bike quit (Ping timeout: 240 seconds)
2017-10-25T00:59:14Z Bike joined #sbcl
2017-10-25T01:07:23Z oleo quit (Ping timeout: 258 seconds)
2017-10-25T01:45:05Z stassats quit (Ping timeout: 248 seconds)
2017-10-25T02:14:07Z oleo joined #sbcl
2017-10-25T02:20:33Z DeadTrickster quit (Read error: Connection reset by peer)
2017-10-25T02:20:56Z DeadTrickster joined #sbcl
2017-10-25T03:09:59Z Bike quit (Quit: Lost terminal)
2017-10-25T03:15:17Z sjl joined #sbcl
2017-10-25T03:27:35Z sjl_ joined #sbcl
2017-10-25T03:28:03Z sjl quit (Disconnected by services)
2017-10-25T03:28:18Z sjl_ is now known as sjl
2017-10-25T03:30:21Z Chream joined #sbcl
2017-10-25T03:34:27Z Chream quit (Ping timeout: 240 seconds)
2017-10-25T03:38:30Z Chream_ quit (Read error: Connection reset by peer)
2017-10-25T03:39:14Z Chream_ joined #sbcl
2017-10-25T03:41:38Z Chream_2 joined #sbcl
2017-10-25T03:43:38Z Chream_ quit (Ping timeout: 252 seconds)
2017-10-25T03:50:11Z shka_ joined #sbcl
2017-10-25T04:04:33Z Chream_2 quit (Read error: Connection reset by peer)
2017-10-25T04:06:02Z Chream_ joined #sbcl
2017-10-25T04:08:24Z shka_ quit (Read error: Connection reset by peer)
2017-10-25T04:14:59Z shka_ joined #sbcl
2017-10-25T05:19:57Z Chream_ quit (Ping timeout: 240 seconds)
2017-10-25T05:20:42Z Chream_ joined #sbcl
2017-10-25T05:27:08Z Chream_ quit (Ping timeout: 240 seconds)
2017-10-25T05:30:07Z Chream_ joined #sbcl
2017-10-25T05:38:10Z oleo quit (Quit: Leaving)
2017-10-25T05:47:02Z sjl quit (Ping timeout: 260 seconds)
2017-10-25T06:10:20Z mfiano joined #sbcl
2017-10-25T06:23:48Z Chream_ quit (Ping timeout: 240 seconds)
2017-10-25T06:24:12Z Chream_ joined #sbcl
2017-10-25T06:32:10Z scymtym quit (Ping timeout: 255 seconds)
2017-10-25T06:55:08Z attila_lendvai joined #sbcl
2017-10-25T07:06:19Z angavrilov joined #sbcl
2017-10-25T07:31:52Z scymtym joined #sbcl
2017-10-25T07:35:01Z shka_ quit (Ping timeout: 240 seconds)
2017-10-25T07:38:14Z m00natic joined #sbcl
2017-10-25T07:43:01Z Chream_ quit (Ping timeout: 240 seconds)
2017-10-25T07:47:07Z carenz_ joined #sbcl
2017-10-25T08:17:33Z gargaml joined #sbcl
2017-10-25T08:43:04Z milanj joined #sbcl
2017-10-25T09:00:22Z carenz_ quit (Ping timeout: 264 seconds)
2017-10-25T09:35:32Z attila_lendvai quit (Quit: Leaving.)
2017-10-25T10:27:00Z Chream_ joined #sbcl
2017-10-25T10:32:06Z Bike joined #sbcl
2017-10-25T10:41:03Z Chream_ quit (Ping timeout: 248 seconds)
2017-10-25T10:43:55Z Chream_ joined #sbcl
2017-10-25T11:15:35Z Chream_ quit (Ping timeout: 240 seconds)
2017-10-25T11:18:54Z Chream_ joined #sbcl
2017-10-25T11:34:48Z Colleen quit (Ping timeout: 246 seconds)
2017-10-25T11:41:32Z Colleen joined #sbcl
2017-10-25T12:28:56Z Bike quit (Ping timeout: 255 seconds)
2017-10-25T12:33:05Z Chream_ quit (Ping timeout: 240 seconds)
2017-10-25T12:49:30Z milanj quit (Quit: This computer has gone to sleep)
2017-10-25T12:59:24Z Bike joined #sbcl
2017-10-25T13:29:55Z Chream_ joined #sbcl
2017-10-25T13:33:08Z Chream_ quit (Read error: Connection reset by peer)
2017-10-25T13:33:41Z Chream_2 joined #sbcl
2017-10-25T13:34:53Z oleo joined #sbcl
2017-10-25T13:35:36Z sjl___ joined #sbcl
2017-10-25T13:40:42Z Chream_2 quit (Ping timeout: 260 seconds)
2017-10-25T13:47:27Z cromachina quit (Read error: Connection reset by peer)
2017-10-25T13:48:36Z stassats joined #sbcl
2017-10-25T14:17:22Z sjl___ is now known as sjl
2017-10-25T14:29:25Z Chream_ joined #sbcl
2017-10-25T14:33:40Z Chream_ quit (Ping timeout: 255 seconds)
2017-10-25T14:36:48Z scymtym quit (Ping timeout: 246 seconds)
2017-10-25T14:57:33Z whoman quit (Read error: Connection reset by peer)
2017-10-25T15:10:58Z Jesin quit (Quit: Leaving)
2017-10-25T15:13:14Z Jesin joined #sbcl
2017-10-25T15:23:35Z m00natic quit (Remote host closed the connection)
2017-10-25T15:26:26Z rpg joined #sbcl
2017-10-25T15:26:31Z rpg quit (Client Quit)
2017-10-25T15:26:44Z rpg joined #sbcl
2017-10-25T15:30:18Z milanj joined #sbcl
2017-10-25T15:36:28Z edgar-rft quit (Quit: edgar-rft)
2017-10-25T15:47:12Z scymtym joined #sbcl
2017-10-25T15:57:33Z gargaml quit (Quit: WeeChat 1.9)
2017-10-25T16:18:20Z rtmpdavid quit (Remote host closed the connection)
2017-10-25T16:39:49Z Chream_ joined #sbcl
2017-10-25T16:52:28Z Chream_2 joined #sbcl
2017-10-25T16:53:02Z Chream_ quit (Ping timeout: 246 seconds)
2017-10-25T16:53:12Z milanj quit (Quit: This computer has gone to sleep)
2017-10-25T17:26:36Z shka_ joined #sbcl
2017-10-25T18:17:16Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2017-10-25T18:23:11Z milanj joined #sbcl
2017-10-25T18:35:36Z rpg joined #sbcl
2017-10-25T18:57:00Z edgar-rft joined #sbcl
2017-10-25T19:04:23Z Chream_2 quit (Read error: Connection reset by peer)
2017-10-25T19:06:18Z Chream_2 joined #sbcl
2017-10-25T19:10:01Z Chream_2 quit (Read error: Connection reset by peer)
2017-10-25T19:10:15Z Chream_ joined #sbcl
2017-10-25T19:12:36Z Chream_2 joined #sbcl
2017-10-25T19:13:04Z Chream_ quit (Read error: Connection reset by peer)
2017-10-25T19:13:04Z Chream_2 quit (Read error: Connection reset by peer)
2017-10-25T19:13:16Z Chream_ joined #sbcl
2017-10-25T19:14:24Z Chream_2 joined #sbcl
2017-10-25T19:15:14Z Chream_2 quit (Read error: Connection reset by peer)
2017-10-25T19:15:47Z Chream_2 joined #sbcl
2017-10-25T19:16:44Z Chream_ quit (Read error: Connection reset by peer)
2017-10-25T19:47:46Z stassats quit (Ping timeout: 264 seconds)
2017-10-25T19:48:38Z shka_ quit (Ping timeout: 258 seconds)
2017-10-25T20:21:06Z angavrilov quit (Remote host closed the connection)
2017-10-25T20:21:44Z attila_lendvai joined #sbcl
2017-10-25T20:33:52Z Bike quit (Ping timeout: 258 seconds)
2017-10-25T21:03:40Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2017-10-25T21:14:05Z rpg joined #sbcl
2017-10-25T21:16:39Z whoman joined #sbcl
2017-10-25T21:22:46Z Bike joined #sbcl
2017-10-25T21:36:26Z Chream_2 quit (Read error: Connection reset by peer)
2017-10-25T21:36:32Z Chream_ joined #sbcl
2017-10-25T21:37:03Z Chream_ quit (Read error: Connection reset by peer)
2017-10-25T21:37:07Z Chream_2 joined #sbcl
2017-10-25T21:37:20Z attila_lendvai quit (Quit: Leaving.)
2017-10-25T21:39:46Z Chream_2 quit (Read error: Connection reset by peer)
2017-10-25T21:39:52Z Chream_ joined #sbcl
2017-10-25T21:40:24Z Chream_2 joined #sbcl
2017-10-25T21:40:25Z Chream_ quit (Read error: Connection reset by peer)
2017-10-25T21:43:35Z Chream_2 quit (Read error: Connection reset by peer)
2017-10-25T21:44:14Z Chream_ joined #sbcl
2017-10-25T22:17:47Z sjl_ joined #sbcl
2017-10-25T22:20:10Z sjl quit (Ping timeout: 264 seconds)
2017-10-25T22:28:06Z sjl_ quit (Ping timeout: 258 seconds)
2017-10-25T22:42:50Z sjl_ joined #sbcl
2017-10-25T23:14:44Z stassats joined #sbcl
2017-10-25T23:39:46Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2017-10-26T00:48:25Z MetaYan joined #sbcl
2017-10-26T01:20:48Z milanj quit (Quit: This computer has gone to sleep)
2017-10-26T01:21:35Z DGASAU quit (Ping timeout: 248 seconds)
2017-10-26T01:22:01Z DGASAU joined #sbcl
2017-10-26T01:30:24Z Chream_2 joined #sbcl
2017-10-26T01:30:41Z Chream_ quit (Read error: Connection reset by peer)
2017-10-26T01:30:53Z Chream_ joined #sbcl
2017-10-26T01:30:53Z Chream_2 quit (Read error: Connection reset by peer)
2017-10-26T01:36:54Z Chream_2 joined #sbcl
2017-10-26T01:37:05Z Chream_ quit (Ping timeout: 248 seconds)
2017-10-26T01:37:33Z Chream_2 quit (Read error: Connection reset by peer)
2017-10-26T01:39:14Z Chream_ joined #sbcl
2017-10-26T01:51:08Z milanj joined #sbcl
2017-10-26T01:54:09Z Chream_ quit (Ping timeout: 248 seconds)
2017-10-26T01:55:42Z stassats quit (Ping timeout: 260 seconds)
2017-10-26T03:09:47Z Bike quit (Quit: Lost terminal)
2017-10-26T04:25:39Z pfdietz joined #sbcl
2017-10-26T04:29:07Z pfdietz: The random tester has stopped finding new bugs.  It's just showing that old one (rarely).  I'm running it on two cores nonstop with two different size settings.
2017-10-26T05:07:16Z oleo quit (Quit: Leaving)
2017-10-26T05:09:41Z shka_ joined #sbcl
2017-10-26T05:13:47Z whoman quit (Remote host closed the connection)
2017-10-26T05:16:01Z whoman joined #sbcl
2017-10-26T06:08:05Z scymtym quit (Ping timeout: 240 seconds)
2017-10-26T07:11:27Z shka_ quit (Ping timeout: 240 seconds)
2017-10-26T07:22:55Z scymtym joined #sbcl
2017-10-26T07:25:13Z Colleen quit (Quit: Colleen)
2017-10-26T07:25:49Z Colleen joined #sbcl
2017-10-26T07:48:09Z scymtym quit (Ping timeout: 258 seconds)
2017-10-26T07:49:08Z edgar-rft quit (Quit: edgar-rft)
2017-10-26T08:09:48Z attila_lendvai joined #sbcl
2017-10-26T08:09:48Z attila_lendvai quit (Changing host)
2017-10-26T08:09:48Z attila_lendvai joined #sbcl
2017-10-26T08:39:23Z attila_lendvai quit (Quit: Leaving.)
2017-10-26T08:49:05Z scymtym joined #sbcl
2017-10-26T08:51:47Z attila_lendvai joined #sbcl
2017-10-26T08:51:47Z attila_lendvai quit (Changing host)
2017-10-26T08:51:47Z attila_lendvai joined #sbcl
2017-10-26T08:52:38Z attila_lendvai quit (Client Quit)
2017-10-26T09:23:42Z milanj quit (Quit: This computer has gone to sleep)
2017-10-26T09:58:59Z edgar-rft joined #sbcl
2017-10-26T10:26:29Z angavrilov joined #sbcl
2017-10-26T10:41:18Z m00natic joined #sbcl
2017-10-26T11:36:11Z Bike joined #sbcl
2017-10-26T12:00:17Z rpg joined #sbcl
2017-10-26T12:13:33Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2017-10-26T12:16:54Z scymtym quit (Remote host closed the connection)
2017-10-26T12:17:10Z scymtym joined #sbcl
2017-10-26T12:43:27Z stassats joined #sbcl
2017-10-26T12:43:38Z Bike quit (Ping timeout: 252 seconds)
2017-10-26T12:44:25Z stassats: pfdietz: hurray, no more bugs!
2017-10-26T12:53:27Z Jesin quit (Quit: Leaving)
2017-10-26T12:58:26Z scymtym quit (Ping timeout: 246 seconds)
2017-10-26T13:12:00Z Bike joined #sbcl
2017-10-26T13:15:57Z stassats quit (Ping timeout: 240 seconds)
2017-10-26T13:18:24Z stassats joined #sbcl
2017-10-26T13:20:25Z m00natic left #sbcl
2017-10-26T13:25:18Z scymtym joined #sbcl
2017-10-26T13:27:11Z stassats quit (Quit: ERC (IRC client for Emacs 25.1.2))
2017-10-26T13:36:26Z sjl joined #sbcl
2017-10-26T13:40:19Z karswell_ joined #sbcl
2017-10-26T14:01:35Z sjl_ quit (Ping timeout: 240 seconds)
2017-10-26T14:02:02Z Jesin joined #sbcl
2017-10-26T14:04:22Z Jesin quit (Remote host closed the connection)
2017-10-26T14:05:33Z scymtym quit (Ping timeout: 246 seconds)
2017-10-26T14:09:29Z Jesin joined #sbcl
2017-10-26T14:15:01Z Jesin quit (Quit: Leaving)
2017-10-26T14:19:51Z oleo joined #sbcl
2017-10-26T15:00:13Z pfdietz_ joined #sbcl
2017-10-26T15:05:39Z rpg joined #sbcl
2017-10-26T15:06:46Z rpg quit (Client Quit)
2017-10-26T15:33:27Z Chream joined #sbcl
2017-10-26T15:34:02Z rpg joined #sbcl
2017-10-26T15:37:20Z pfdietz_ quit (Ping timeout: 246 seconds)
2017-10-26T15:37:31Z milanj joined #sbcl
2017-10-26T15:42:10Z pfdietz_ joined #sbcl
2017-10-26T15:43:30Z pfdietz_ quit (Client Quit)
2017-10-26T16:23:45Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2017-10-26T16:26:33Z pfdietz_ joined #sbcl
2017-10-26T16:36:35Z rpg joined #sbcl
2017-10-26T16:44:34Z scymtym joined #sbcl
2017-10-26T16:47:50Z pfdietz_ quit (Ping timeout: 252 seconds)
2017-10-26T16:56:24Z milanj quit (Quit: This computer has gone to sleep)
2017-10-26T17:05:49Z shka_ joined #sbcl
2017-10-26T17:06:09Z milanj joined #sbcl
2017-10-26T17:12:12Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2017-10-26T17:14:55Z pfdietz_ joined #sbcl
2017-10-26T17:17:08Z rpg joined #sbcl
2017-10-26T17:33:13Z attila_lendvai joined #sbcl
2017-10-26T17:33:13Z attila_lendvai quit (Changing host)
2017-10-26T17:33:13Z attila_lendvai joined #sbcl
2017-10-26T17:57:20Z attila_lendvai quit (Quit: Leaving.)
2017-10-26T18:29:58Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2017-10-26T18:35:15Z rpg joined #sbcl
2017-10-26T18:57:21Z Chream quit (Ping timeout: 240 seconds)
2017-10-26T19:02:47Z Chream joined #sbcl
2017-10-26T19:30:14Z scymtym_ joined #sbcl
2017-10-26T19:32:06Z scymtym quit (Ping timeout: 246 seconds)
2017-10-26T19:35:20Z scymtym_ quit (Ping timeout: 246 seconds)
2017-10-26T19:52:21Z Chream quit (Ping timeout: 240 seconds)
2017-10-26T20:12:35Z pfdietz_ quit (Ping timeout: 240 seconds)
2017-10-26T20:14:05Z pfdietz_ joined #sbcl
2017-10-26T20:17:21Z pfdietz_ quit (Remote host closed the connection)
2017-10-26T20:18:49Z pfdietz_ joined #sbcl
2017-10-26T20:33:20Z pfdietz_ quit (Ping timeout: 252 seconds)
2017-10-26T20:35:57Z angavrilov quit (Remote host closed the connection)
2017-10-26T20:49:45Z pfdietz_ joined #sbcl
2017-10-26T21:00:47Z shka_ quit (Ping timeout: 260 seconds)
2017-10-26T21:04:08Z pfdietz_ quit (Ping timeout: 252 seconds)
2017-10-26T21:05:39Z pfdietz_ joined #sbcl
2017-10-26T21:08:53Z Colleen quit (Quit: Colleen)
2017-10-26T21:09:33Z Colleen joined #sbcl
2017-10-26T21:10:11Z Colleen quit (Client Quit)
2017-10-26T21:10:38Z Colleen joined #sbcl
2017-10-26T21:14:44Z pfdietz_ quit (Ping timeout: 246 seconds)
2017-10-26T21:16:37Z pfdietz_ joined #sbcl
2017-10-26T21:19:13Z Aaryn joined #sbcl
2017-10-26T21:21:12Z pfdietz_ quit (Ping timeout: 258 seconds)
2017-10-26T21:27:09Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2017-10-26T21:49:27Z scymtym joined #sbcl
2017-10-26T21:56:12Z Bike quit (Ping timeout: 260 seconds)
2017-10-26T22:07:01Z sjl quit (Ping timeout: 240 seconds)
2017-10-26T22:12:28Z bentaisan joined #sbcl
2017-10-26T22:29:01Z karswell_ quit (Ping timeout: 240 seconds)
2017-10-26T22:32:21Z Bike joined #sbcl
2017-10-26T22:45:52Z jrm2 joined #sbcl
2017-10-26T22:46:27Z jrm quit (Ping timeout: 240 seconds)
2017-10-26T22:46:37Z cromachina joined #sbcl
2017-10-26T22:46:37Z jrm2 is now known as jrm
2017-10-26T22:48:18Z Jesin joined #sbcl
2017-10-26T23:44:12Z milanj quit (Quit: This computer has gone to sleep)
2017-10-27T00:23:18Z milanj joined #sbcl
2017-10-27T00:30:17Z rpg joined #sbcl
2017-10-27T00:40:53Z slyrus joined #sbcl
2017-10-27T01:04:54Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2017-10-27T01:12:05Z milanj quit (Quit: This computer has gone to sleep)
2017-10-27T01:43:30Z Chream joined #sbcl
2017-10-27T01:57:27Z Chream quit (Ping timeout: 240 seconds)
2017-10-27T02:00:51Z rpg joined #sbcl
2017-10-27T02:13:34Z rpg quit (Quit: Textual IRC Client: www.textualapp.com)
2017-10-27T02:46:27Z sjl joined #sbcl
2017-10-27T02:53:42Z Chream joined #sbcl
2017-10-27T03:51:05Z Jesin quit (Quit: Leaving)
2017-10-27T03:53:05Z Chream quit (Ping timeout: 248 seconds)
2017-10-27T04:00:01Z Bike quit (Quit: Lost terminal)
2017-10-27T04:23:06Z pfdietz quit
2017-10-27T04:50:32Z brucem quit (Ping timeout: 252 seconds)
2017-10-27T04:51:44Z |3b|`` quit (Ping timeout: 255 seconds)
2017-10-27T05:01:58Z sjl quit (Ping timeout: 258 seconds)
2017-10-27T05:03:38Z brucem joined #sbcl
2017-10-27T05:24:39Z milanj joined #sbcl
2017-10-27T05:29:17Z oleo quit (Quit: Leaving)
2017-10-27T05:49:52Z shka_ joined #sbcl
2017-10-27T06:03:11Z jack_rabbit quit (Ping timeout: 248 seconds)
2017-10-27T06:16:01Z scymtym quit (Ping timeout: 248 seconds)
2017-10-27T06:47:35Z shka_ quit (Ping timeout: 240 seconds)
2017-10-27T07:29:11Z scymtym joined #sbcl
2017-10-27T07:29:27Z attila_lendvai joined #sbcl
2017-10-27T07:29:27Z attila_lendvai quit (Changing host)
2017-10-27T07:29:27Z attila_lendvai joined #sbcl
2017-10-27T07:37:03Z attila_lendvai quit (Ping timeout: 252 seconds)
2017-10-27T08:09:53Z attila_lendvai joined #sbcl
2017-10-27T08:09:53Z attila_lendvai quit (Changing host)
2017-10-27T08:09:53Z attila_lendvai joined #sbcl
2017-10-27T08:41:57Z milanj quit (Quit: This computer has gone to sleep)
2017-10-27T08:43:45Z |3b|`` joined #sbcl
2017-10-27T09:15:30Z attila_lendvai quit (Read error: Connection reset by peer)
2017-10-27T09:16:24Z attila_lendvai joined #sbcl
2017-10-27T09:16:24Z attila_lendvai quit (Changing host)
2017-10-27T09:16:24Z attila_lendvai joined #sbcl
2017-10-27T09:17:28Z DeadTrickster_ joined #sbcl
2017-10-27T09:19:57Z DeadTrickster quit (Ping timeout: 240 seconds)
2017-10-27T09:51:24Z angavrilov joined #sbcl
2017-10-27T10:08:49Z m00natic joined #sbcl
2017-10-27T11:43:19Z karswell_ joined #sbcl
2017-10-27T12:09:52Z whoman quit (Quit: Leaving)
2017-10-27T12:18:05Z attila_lendvai quit (Quit: Leaving.)
2017-10-27T12:18:26Z attila_lendvai joined #sbcl
2017-10-27T12:36:09Z attila_lendvai quit (Quit: Leaving.)
2017-10-27T12:51:13Z karswell_ quit (Ping timeout: 248 seconds)
2017-10-27T12:51:43Z karswell_ joined #sbcl
2017-10-27T13:09:01Z cromachina quit (Read error: Connection reset by peer)
2017-10-27T13:17:45Z oleo joined #sbcl
2017-10-27T13:23:49Z Bike joined #sbcl
2017-10-27T13:32:29Z scymtym quit (Remote host closed the connection)
2017-10-27T13:32:48Z scymtym joined #sbcl
2017-10-27T13:33:41Z sjl joined #sbcl
2017-10-27T13:44:14Z stassats joined #sbcl
2017-10-27T14:03:27Z stassats quit (Ping timeout: 240 seconds)
2017-10-27T14:05:45Z Jesin joined #sbcl
2017-10-27T14:08:29Z foom2 joined #sbcl
2017-10-27T14:11:32Z foom quit (Ping timeout: 252 seconds)
2017-10-27T14:13:28Z whoman joined #sbcl
2017-10-27T14:34:05Z scymtym quit (Ping timeout: 252 seconds)
2017-10-27T14:57:24Z Chream joined #sbcl
2017-10-27T15:13:57Z Chream quit (Remote host closed the connection)
2017-10-27T15:16:11Z Chream joined #sbcl
2017-10-27T15:22:53Z milanj joined #sbcl
2017-10-27T15:29:19Z milanj quit (Quit: This computer has gone to sleep)
2017-10-27T15:31:02Z pfdietz joined #sbcl
2017-10-27T15:35:37Z pfdietz quit (Ping timeout: 258 seconds)
2017-10-27T15:42:33Z pfdietz joined #sbcl
2017-10-27T15:49:26Z pfdietz quit (Ping timeout: 252 seconds)
2017-10-27T15:51:13Z pfdietz joined #sbcl
2017-10-27T15:59:29Z scymtym joined #sbcl
2017-10-27T16:00:26Z pfdietz quit (Ping timeout: 252 seconds)
2017-10-27T16:28:29Z dougk quit (Ping timeout: 252 seconds)
2017-10-27T16:38:31Z Jesin quit (Quit: Leaving)
2017-10-27T16:40:04Z dougk joined #sbcl
2017-10-27T16:42:28Z Chream quit (Ping timeout: 240 seconds)
2017-10-27T16:44:39Z Jesin joined #sbcl
2017-10-27T16:54:13Z m00natic quit (Remote host closed the connection)
2017-10-27T17:12:12Z milanj joined #sbcl
2017-10-27T17:29:57Z fiveop joined #sbcl
2017-10-27T18:07:17Z jack_rabbit joined #sbcl
2017-10-27T18:08:15Z stassats joined #sbcl
2017-10-27T18:11:53Z jack_rabbit quit (Client Quit)
2017-10-27T18:12:23Z whoman quit (Remote host closed the connection)
2017-10-27T18:13:41Z whoman joined #sbcl
2017-10-27T18:17:46Z shka_ joined #sbcl
2017-10-27T18:28:26Z stassats: i have an obvious fix for the latest pfdietz report, but now wondering, why is it happening in the first place — args with the NIL type reaching ir1-optimize-combination
2017-10-27T18:28:48Z stassats: delete-lvar-use is supposed to mark the destination block for deletion
2017-10-27T18:32:34Z stassats: of course, delete-lvar-use isn't called on this arg, something calls %delete-lvar-use directly
2017-10-27T18:51:54Z jack_rabbit joined #sbcl
2017-10-27T19:13:08Z karswell_ quit (Read error: Connection reset by peer)
2017-10-27T19:14:55Z angavrilov quit (Remote host closed the connection)
2017-10-27T19:30:37Z foom2 is now known as foom
2017-10-27T19:41:18Z stassats: zeroed in on a better fix, and it's a one character fix
2017-10-27T19:41:25Z stassats: %delete-lvar-use to delete-lvar-use
2017-10-27T19:48:32Z Jesin quit (Quit: Leaving)
2017-10-27T20:19:44Z pfdietz joined #sbcl
2017-10-27T20:20:20Z pfdietz: Was that also 14 years old?
2017-10-27T20:23:14Z stassats: pfdietz: yes...
2017-10-27T20:27:11Z pfdietz quit (Ping timeout: 252 seconds)
2017-10-27T20:28:35Z stassats: the commit even says "remove bug reported by Paul Dietz on sbcl-devel 2003-09-14." not sure if that's for that part
2017-10-27T20:28:53Z stassats: back in the day people tended to bunch several things into a single commit
2017-10-27T20:31:25Z fiveop quit
2017-10-27T20:44:14Z shka_ quit (Ping timeout: 252 seconds)
2017-10-27T21:17:53Z Bike quit (Ping timeout: 248 seconds)
2017-10-27T21:47:24Z Bike joined #sbcl
2017-10-27T21:48:31Z Bike quit (Client Quit)
2017-10-27T21:48:55Z Bike joined #sbcl
2017-10-27T22:01:20Z sjl quit (Quit: WeeChat 1.9)
2017-10-27T22:05:16Z sjl joined #sbcl
2017-10-27T22:27:40Z milanj quit (Quit: This computer has gone to sleep)
2017-10-27T22:39:12Z cromachina joined #sbcl
2017-10-27T23:02:16Z milanj joined #sbcl
2017-10-27T23:23:51Z milanj quit (Quit: This computer has gone to sleep)
2017-10-27T23:49:36Z milanj joined #sbcl
2017-10-28T00:07:30Z pfdietz joined #sbcl
2017-10-28T01:00:35Z milanj quit (Quit: This computer has gone to sleep)
2017-10-28T01:17:14Z milanj joined #sbcl
2017-10-28T01:18:31Z stassats quit (Ping timeout: 255 seconds)
2017-10-28T01:35:25Z stassats joined #sbcl
2017-10-28T02:07:05Z stassats quit (Ping timeout: 240 seconds)
2017-10-28T02:42:31Z milanj quit (Quit: This computer has gone to sleep)
2017-10-28T02:42:56Z milanj joined #sbcl
2017-10-28T03:29:17Z sjl quit (Quit: WeeChat 1.9.1)
2017-10-28T03:30:05Z sjl joined #sbcl
2017-10-28T06:07:55Z Bike quit (Quit: Lost terminal)
2017-10-28T06:35:05Z angavrilov joined #sbcl
2017-10-28T07:55:03Z shka_ joined #sbcl
2017-10-28T09:23:39Z edgar-rft quit (Quit: edgar-rft)
2017-10-28T11:38:42Z scymtym quit (Ping timeout: 246 seconds)
2017-10-28T11:42:38Z milanj quit (Quit: This computer has gone to sleep)
2017-10-28T11:50:50Z stassats joined #sbcl
2017-10-28T11:57:32Z scymtym joined #sbcl
2017-10-28T12:09:12Z scymtym_ joined #sbcl
2017-10-28T12:11:19Z stassats: why is stack-analyze for DX variables, but not for special variables? both are quite similar
2017-10-28T12:13:05Z scymtym quit (Ping timeout: 246 seconds)
2017-10-28T12:24:15Z stassats: maybe because the size is not known?
2017-10-28T12:25:12Z stassats: but does it have to be known? just save and restore RSP
2017-10-28T12:32:32Z stassats: or is it due to interleaving of unknown value return and dx
2017-10-28T12:37:38Z milanj joined #sbcl
2017-10-28T12:40:29Z Bike joined #sbcl
2017-10-28T12:46:00Z stassats: nip-values before calling return-multiple sounds wasteful
2017-10-28T12:58:39Z stassats quit (Ping timeout: 248 seconds)
2017-10-28T13:50:39Z bentaisan quit (Ping timeout: 246 seconds)
2017-10-28T15:11:57Z myrkraverk quit (Ping timeout: 240 seconds)
2017-10-28T15:12:50Z myrkraverk joined #sbcl
2017-10-28T15:13:25Z stassats joined #sbcl
2017-10-28T15:37:55Z milanj quit (Quit: This computer has gone to sleep)
2017-10-28T16:01:45Z edgar-rft joined #sbcl
2017-10-28T16:05:57Z tonton quit (Ping timeout: 240 seconds)
2017-10-28T16:07:42Z tonton joined #sbcl
2017-10-28T16:20:09Z bkst joined #sbcl
2017-10-28T16:23:58Z stassats quit (Ping timeout: 252 seconds)
2017-10-28T16:38:08Z milanj joined #sbcl
2017-10-28T16:38:14Z milanj quit (Remote host closed the connection)
2017-10-28T16:51:17Z scymtym_ quit (Ping timeout: 255 seconds)
2017-10-28T17:03:48Z Chream joined #sbcl
2017-10-28T17:04:47Z Jesin joined #sbcl
2017-10-28T17:05:06Z Jesin quit (Remote host closed the connection)
2017-10-28T17:18:01Z Chream quit (Ping timeout: 240 seconds)
2017-10-28T17:41:35Z scymtym joined #sbcl
2017-10-28T17:54:30Z Jesin joined #sbcl
2017-10-28T18:49:11Z leah2 quit (Ping timeout: 255 seconds)
2017-10-28T19:03:45Z leah2 joined #sbcl
2017-10-28T19:25:57Z angular_mike___ quit
2017-10-28T19:27:26Z angular_mike___ joined #sbcl
2017-10-28T19:40:46Z stassats joined #sbcl
2017-10-28T20:31:29Z shka_ quit (Ping timeout: 248 seconds)
2017-10-28T20:36:29Z Xof joined #sbcl
2017-10-28T20:40:45Z Jesin quit (Quit: Leaving)
2017-10-28T20:40:57Z stassats: even when stack-analyze works properly it produces some strange code
2017-10-28T20:41:13Z stassats: but i'm still not close to understanding stack-analyze
2017-10-28T20:42:53Z stassats: what if we had a separate stack for values, a lot of shuffling around could be avoided
2017-10-28T20:43:14Z stassats: and stack-analyze falls off
2017-10-28T20:51:56Z stassats: it wouldn't really be a stack, just a location for values/arguments
2017-10-28T21:03:48Z stassats: an experiment would be useful
2017-10-28T21:05:47Z Chream joined #sbcl
2017-10-28T21:06:31Z Chream quit (Remote host closed the connection)
2017-10-28T21:06:47Z Chream joined #sbcl
2017-10-28T21:12:57Z shka_ joined #sbcl
2017-10-28T21:17:20Z Chream quit (Remote host closed the connection)
2017-10-28T21:17:36Z Chream joined #sbcl
2017-10-28T21:25:53Z shka_ quit (Ping timeout: 248 seconds)
2017-10-28T21:33:19Z sjl quit (Ping timeout: 248 seconds)
2017-10-28T23:35:01Z Chream quit (Ping timeout: 255 seconds)
2017-10-29T00:29:15Z stassats: i'm liking more and more the idea of having a dedicated return stack
2017-10-29T01:55:53Z sjl joined #sbcl
2017-10-29T02:03:27Z stassats quit (Ping timeout: 260 seconds)
2017-10-29T02:19:23Z stassats joined #sbcl
2017-10-29T02:23:59Z stassats quit (Ping timeout: 248 seconds)
2017-10-29T04:10:51Z Bike quit (Quit: Lost terminal)
2017-10-29T04:30:54Z sjl quit (Ping timeout: 258 seconds)
2017-10-29T05:48:05Z White_Flame quit (Ping timeout: 240 seconds)
2017-10-29T05:58:32Z White_Flame joined #sbcl
2017-10-29T07:09:17Z whoman quit (Read error: Connection reset by peer)
2017-10-29T08:16:42Z shka_ joined #sbcl
2017-10-29T10:19:20Z scymtym quit (Ping timeout: 255 seconds)
2017-10-29T10:45:54Z scymtym joined #sbcl
2017-10-29T10:46:55Z easieste joined #sbcl
2017-10-29T10:50:42Z easieste quit (Client Quit)
2017-10-29T11:55:16Z Bike joined #sbcl
2017-10-29T11:56:48Z leah2 quit (Ping timeout: 246 seconds)
2017-10-29T12:00:18Z leah2 joined #sbcl
2017-10-29T13:35:29Z josemanuel joined #sbcl
2017-10-29T14:08:58Z Chream joined #sbcl
2017-10-29T14:17:21Z minion quit (Remote host closed the connection)
2017-10-29T14:17:54Z Shinmera quit (Remote host closed the connection)
2017-10-29T14:18:12Z minion joined #sbcl
2017-10-29T14:18:17Z Shinmera joined #sbcl
2017-10-29T14:20:32Z abbe quit (Ping timeout: 240 seconds)
2017-10-29T14:23:03Z abbe joined #sbcl
2017-10-29T15:21:21Z sjl joined #sbcl
2017-10-29T16:36:34Z attila_lendvai joined #sbcl
2017-10-29T16:36:34Z attila_lendvai quit (Changing host)
2017-10-29T16:36:34Z attila_lendvai joined #sbcl
2017-10-29T17:00:38Z cromachina_ joined #sbcl
2017-10-29T17:00:47Z Chream quit (Read error: Connection reset by peer)
2017-10-29T17:01:05Z sjl quit (Ping timeout: 258 seconds)
2017-10-29T17:02:56Z cromachina quit (Ping timeout: 252 seconds)
2017-10-29T17:03:59Z sjl joined #sbcl
2017-10-29T17:09:22Z Jesin joined #sbcl
2017-10-29T17:18:27Z scymtym quit (Ping timeout: 246 seconds)
2017-10-29T18:06:50Z sjl quit (Ping timeout: 246 seconds)
2017-10-29T18:47:20Z sjl joined #sbcl
2017-10-29T19:10:32Z adlai quit (Ping timeout: 252 seconds)
2017-10-29T19:16:52Z adlai joined #sbcl
2017-10-29T19:19:04Z Chream joined #sbcl
2017-10-29T19:38:02Z Bicyclidine joined #sbcl
2017-10-29T19:38:05Z Bicyclidine quit (Client Quit)
2017-10-29T20:10:05Z angavrilov quit (Remote host closed the connection)
2017-10-29T20:29:33Z scymtym joined #sbcl
2017-10-29T20:34:05Z Chream quit (Remote host closed the connection)
2017-10-29T22:12:12Z josemanuel quit (Quit: leaving)
2017-10-29T22:35:45Z shka_ quit (Ping timeout: 248 seconds)
2017-10-29T22:37:15Z stassats joined #sbcl
2017-10-29T23:29:25Z Bicyclidine joined #sbcl
2017-10-29T23:35:01Z stassats quit (Ping timeout: 240 seconds)
2017-10-29T23:50:01Z aeth quit (Ping timeout: 240 seconds)
2017-10-29T23:54:51Z aeth joined #sbcl
2017-10-30T00:07:47Z Jesin quit (Quit: Leaving)
2017-10-30T00:14:17Z Jesin joined #sbcl
2017-10-30T01:38:01Z attila_lendvai quit (Quit: Leaving.)
2017-10-30T01:45:28Z edgar-rft quit (Ping timeout: 240 seconds)
2017-10-30T02:33:33Z Bicyclidine quit (Ping timeout: 246 seconds)
2017-10-30T03:43:49Z pfdietz: Yay, sourceforge is down again.
2017-10-30T03:45:56Z edgar-rft joined #sbcl
2017-10-30T03:53:54Z White_Flame quit (Ping timeout: 258 seconds)
2017-10-30T03:59:03Z White_Flame joined #sbcl
2017-10-30T04:01:25Z shka_ joined #sbcl
2017-10-30T04:32:28Z sjl quit (Ping timeout: 255 seconds)
2017-10-30T04:41:03Z jack_rabbit quit (Ping timeout: 248 seconds)
2017-10-30T04:46:41Z White_Flame quit (Ping timeout: 240 seconds)
2017-10-30T04:49:36Z White_Flame joined #sbcl
2017-10-30T04:52:17Z Colleen quit (Ping timeout: 248 seconds)
2017-10-30T04:54:26Z jack_rabbit joined #sbcl
2017-10-30T06:12:54Z angavrilov joined #sbcl
2017-10-30T06:39:52Z oleo quit (Quit: Leaving)
2017-10-30T06:43:26Z scymtym quit (Ping timeout: 255 seconds)
2017-10-30T07:59:19Z shka_ quit (Quit: Konversation terminated!)
2017-10-30T08:01:32Z mood quit (Ping timeout: 258 seconds)
2017-10-30T08:02:24Z mood joined #sbcl
2017-10-30T08:07:31Z shka joined #sbcl
2017-10-30T08:09:56Z scymtym joined #sbcl
2017-10-30T08:14:33Z salva quit (Remote host closed the connection)
2017-10-30T08:21:08Z salva joined #sbcl
2017-10-30T09:38:28Z scymtym quit (Ping timeout: 240 seconds)
2017-10-30T10:55:22Z scymtym joined #sbcl
2017-10-30T10:59:42Z m00natic joined #sbcl
2017-10-30T11:11:49Z Bicyclidine joined #sbcl
2017-10-30T11:52:02Z Bicyclidine quit (Ping timeout: 255 seconds)
2017-10-30T12:14:30Z attila_lendvai joined #sbcl
2017-10-30T12:33:16Z Colleen joined #sbcl
2017-10-30T12:43:32Z Jesin quit (Quit: Leaving)
2017-10-30T13:14:22Z stassats joined #sbcl
2017-10-30T13:17:47Z stassats quit (Read error: Connection reset by peer)
2017-10-30T13:17:53Z stassats` joined #sbcl
2017-10-30T13:28:55Z cromachina_ quit (Read error: Connection reset by peer)
2017-10-30T13:29:18Z sjl joined #sbcl
2017-10-30T13:54:24Z Jesin joined #sbcl
2017-10-30T13:57:40Z scymtym: stassats`: Fare clarified that the apparent problem with ASDF 3.3.0 actually only occurs when combining UIOP 3.3.0 with an ASDF /older/ than ASDF 3.3.0. there don't seem to be other severe known issues. are you ok with going through with the upgrade as planned?
2017-10-30T13:58:07Z stassats`: i still don't like the silly warnings
2017-10-30T13:58:19Z stassats`: but i don't know what to do about that
2017-10-30T13:59:03Z Shinmera: scymtym: You just reminded me to report my 3.3 issue to the mailing list
2017-10-30T13:59:06Z Shinmera goes off to do that
2017-10-30T14:00:40Z scymtym: Shinmera: whom does it affect? is it severe enough to warrant postponing the upgrade?
2017-10-30T14:02:01Z Shinmera: scymtym: It affects me and is really annoying.
2017-10-30T14:02:17Z Shinmera: It basically breaks some of my applications.
2017-10-30T14:02:22Z stassats`: that's how i describe ASDF as a whole
2017-10-30T14:03:14Z scymtym: Shinmera: Fare mentioned a bug fix 3.3.0.1 and suggested upgrading to that version instead. do you know whether that fix might be for your issue?
2017-10-30T14:03:50Z Shinmera: scymtym: I haven't verified yet. I'm working on it now.
2017-10-30T14:07:36Z Shinmera: scymtym: It's fixed in 3.3.0.1
2017-10-30T14:07:43Z stassats joined #sbcl
2017-10-30T14:10:31Z scymtym: Shinmera: thanks
2017-10-30T14:12:37Z Jesin quit (Quit: Leaving)
2017-10-30T14:13:17Z stassats` quit (Quit: ERC (IRC client for Emacs 25.1.2))
2017-10-30T14:15:03Z scymtym: i'm pushing the ASDF upgrade then. we can revert in case anything breaks badly
2017-10-30T14:22:14Z Bicyclidine joined #sbcl
2017-10-30T14:23:48Z scymtym notices that we are still in freeze
2017-10-30T14:31:02Z Jesin joined #sbcl
2017-10-30T15:33:01Z oleo joined #sbcl
2017-10-30T16:13:53Z slyrus quit (Ping timeout: 248 seconds)
2017-10-30T16:28:38Z attila_lendvai quit (Remote host closed the connection)
2017-10-30T16:55:20Z Jesin quit (Quit: Leaving)
2017-10-30T17:17:32Z Xof: grmbl
2017-10-30T17:17:35Z Xof: grmbl grmbl
2017-10-30T17:17:39Z Xof: tar: sbcl-1.4.1-x86-64-linux/contrib/vanilla-module.mk: Cannot stat: No such file or directory
2017-10-30T17:17:58Z Xof: lucky I have lots of other things to do at work so I don't have any chance of going home anyway
2017-10-30T17:18:09Z Jesin joined #sbcl
2017-10-30T17:41:40Z pfdietz_ joined #sbcl
2017-10-30T17:46:17Z pfdietz_ quit (Ping timeout: 252 seconds)
2017-10-30T17:49:56Z pfdietz_ joined #sbcl
2017-10-30T17:50:10Z scymtym: Xof: thank you for doing the release work
2017-10-30T17:50:34Z pfdietz_: And I hope you are feeling better.
2017-10-30T17:59:29Z pfdietz_ quit (Ping timeout: 252 seconds)
2017-10-30T17:59:47Z pfdietz_ joined #sbcl
2017-10-30T18:04:08Z pfdietz_ quit (Ping timeout: 258 seconds)
2017-10-30T18:05:05Z m00natic quit (Remote host closed the connection)
2017-10-30T18:33:18Z dougk: stassats: is there any reason the primitive object for symbol has ":set-trans %set-symbol-global-value" on the value slot, but we also explicitly define a vop that translates %set-symbol-global-value on every backend?
2017-10-30T18:33:56Z dougk: i would have guessed that the primitive object's set-trans prevails, but it does not.
2017-10-30T18:45:14Z jdz quit (Ping timeout: 255 seconds)
2017-10-30T18:48:51Z jdz joined #sbcl
2017-10-30T19:21:26Z stassats: ugh, with-alien leaks out my new special form, i wonder if i can cheat and make *alien-stack-pointer* be treated as NSP
2017-10-30T19:22:06Z stassats: dougk: no immediate idea
2017-10-30T19:43:04Z stassats: ugh, why is my vop not being translated, i've tried all result-type combinations
2017-10-30T19:43:33Z stassats: i guess i'll go ahead with treating (let ((*alien-stack-pointer* *alien-stack-pointer*))) as a different special form on non-x86oids
2017-10-30T19:47:07Z angavrilov quit (Remote host closed the connection)
2017-10-30T19:48:10Z stassats: and i still have no idea what a "funny function" is
2017-10-30T19:49:50Z dougk: in your "(let ((*alien-stack-pointer" you mean any special rebound to itself, not just alien-stack-pointer, right?
2017-10-30T19:50:14Z stassats: dougk: doing nothing related to special variables
2017-10-30T19:50:52Z stassats: and special casing to *alien-stack-pointer*
2017-10-30T19:51:08Z stassats: basically, exactly (*alien-stack-pointer* *alien-stack-pointer*) will actually save and restore NSP
2017-10-30T19:57:53Z Jesin quit (Quit: Leaving)
2017-10-30T19:58:16Z stassats: i can just add %primitive and restoring-nsp to the walker, but then some third party code may not like it
2017-10-30T20:38:33Z scymtym quit (Ping timeout: 246 seconds)
2017-10-30T21:17:29Z shka quit (Ping timeout: 246 seconds)
2017-10-30T21:19:20Z dougk: stassats: ftb in contribs on ppc after "Fix stack exhaustion in with-alien on non-x86oids." - caught ERROR: (during macroexpansion of (SB-PCL::%DEFMETHOD-EXPANDER SOCKET-RECEIVE ...)) unexpected special form SB-SYS:%PRIMITIVE
2017-10-30T21:19:26Z scymtym joined #sbcl
2017-10-30T21:39:14Z Jesin joined #sbcl
2017-10-30T21:43:21Z Bicyclidine quit (Ping timeout: 240 seconds)
2017-10-30T21:44:21Z stassats: dougk: that's why the (let ((*alien-stack-pointer* *alien-stack-pointer*))) thing
2017-10-30T21:45:03Z stassats: except now i have a sb-sprof failure
2017-10-30T21:52:13Z sjl quit (Ping timeout: 258 seconds)
2017-10-30T21:54:27Z dougk: oh, i wasn't following closely enough. thanks
2017-10-30T22:10:24Z stassats: sb-sprof may not actually be related to *alien-stack-pointer*
2017-10-30T22:10:39Z stassats: it gets CFP from a context but it's 0
2017-10-30T22:10:52Z stassats: it's probably in a foreign function, but why didn't that happen before?
2017-10-30T22:14:34Z stassats: looks like it fails even after reverting the nsp-touching changes
2017-10-30T22:15:54Z stassats: it interrupted pthread_sigmask
2017-10-30T22:16:28Z stylewarning: ; 167:       4C8BCA           MOV R9, RDX
2017-10-30T22:16:28Z stylewarning: ; 16A:       498BD1           MOV RDX, R9
2017-10-30T22:16:28Z stylewarning: ; 16D:       4C8BCA           MOV R9, RDX
2017-10-30T22:16:32Z stylewarning: :/
2017-10-30T22:16:42Z stassats: meh, it's free
2017-10-30T22:16:53Z stassats: a move here, a move there
2017-10-30T22:17:13Z stylewarning: stassats: is it free if it gets executed an exponential number of times
2017-10-30T22:17:24Z milanj joined #sbcl
2017-10-30T22:17:29Z stylewarning: like 137438953472 / 2 times
2017-10-30T22:17:41Z stassats: well, some icache usage, but other than that, probably free
2017-10-30T22:17:47Z stassats: or close to it
2017-10-30T22:18:25Z Shinmera: microarchitecture is going to eliminate the moves most likely anyway.
2017-10-30T22:18:29Z Shinmera: register renaming and all that
2017-10-30T22:18:47Z stylewarning: what does anything mean anymore
2017-10-30T22:19:04Z Bicyclidine joined #sbcl
2017-10-30T22:19:16Z Shinmera: Nobody really understands modern CPUs
2017-10-30T22:21:11Z sjl joined #sbcl
2017-10-30T22:24:45Z stassats: i'm declaring sb-sprof broken on -c-stack-is-control-stack
2017-10-30T22:25:04Z stassats: whether it worked before or something's changed, it clearly can't just grab CFP and assume it's from lisp
2017-10-30T22:26:09Z stassats: just how on earth did it work before?
2017-10-30T22:29:01Z aeth quit (Ping timeout: 240 seconds)
2017-10-30T22:31:16Z aeth joined #sbcl
2017-10-30T22:42:56Z stassats: let's see if sb-sprof suddenly starts to lose on ppc
2017-10-30T22:43:12Z stassats: it may be that the new sb-sprof.asd is catching more errors
2017-10-30T22:44:30Z stassats: building vanilla 1.4.1 to test that
2017-10-30T22:45:29Z stassats: the fix to sb-sprof should be trivial, just weird how it was never noticed before
2017-10-30T22:51:01Z stassats: not failing on ppc, but pthread_sigmask is showing up in the profile
2017-10-30T22:51:36Z White_Flame quit (Read error: Connection reset by peer)
2017-10-30T22:52:00Z stassats: the CFP register might just be still alive
2017-10-30T22:53:50Z White_Flame joined #sbcl
2017-10-30T22:55:56Z Bicyclidine quit (Ping timeout: 252 seconds)
2017-10-30T23:14:32Z whoman joined #sbcl
2017-10-30T23:19:33Z stylewarning: How should ; 76:       660F1F840000000000 NOP be interpreted?
2017-10-30T23:21:04Z dougk: where's that come from - an x86-64 disassembly? doesn't seem to be a valid NOP encoding
2017-10-30T23:25:17Z stassats: stylewarning: as a NOP?..
2017-10-30T23:26:49Z stassats: dougk: it's a valid 9-byte nop
2017-10-30T23:27:24Z dougk: true. is it one that we would emit?
2017-10-30T23:27:27Z dougk: i guess
2017-10-30T23:27:57Z dougk: indeed. emit-long-nop will do that
2017-10-30T23:28:14Z stylewarning: why is it so kooky
2017-10-30T23:28:35Z stassats: stylewarning: what do you mean?
2017-10-30T23:28:41Z dougk: we use the suggested encoding of a NOP of each length
2017-10-30T23:28:47Z stylewarning: I see, ok.
2017-10-30T23:28:48Z dougk: it's 1 instruction in every case.
2017-10-30T23:31:52Z stassats: fixed sb-sprof by not touching the caller frames of foreign functions
2017-10-30T23:32:28Z stassats: could handle that, just like backtraces do, but then it could handle more than two frames, like on x86oids
2017-10-30T23:57:34Z White_Flame quit (Ping timeout: 258 seconds)
2017-10-31T00:02:48Z White_Flame joined #sbcl
2017-10-31T00:36:21Z eschatologist quit (Ping timeout: 240 seconds)
2017-10-31T00:37:06Z stassats: the IR1 is annoying me again
2017-10-31T00:37:32Z stassats: specifically that you can still end up processing a deleted block
2017-10-31T00:37:42Z pfdietz: New random tester bug?
2017-10-31T00:37:49Z rpg joined #sbcl
2017-10-31T00:38:17Z stassats: pfdietz: it was actually caused by %delete-lvar-use -> delete-lvar-use -> more deleted blocks -> more checks for deletion
2017-10-31T00:38:35Z stassats: all these deletion checks are silly, the IR should always be normalized
2017-10-31T00:38:39Z stassats: maybe that's slow, i don't know
2017-10-31T00:38:58Z pfdietz: I figured it was caused by a recent change, since it started showing up at fairly high frequency.
2017-10-31T00:40:12Z stassats: i have no experience with other compilers, what do they do? do they maintain the IR in a good state?
2017-10-31T00:40:32Z pfdietz: I don't know.
2017-10-31T00:41:23Z stassats: someday i'll know what to do with it, for now, one more node-to-be-deleted-p
2017-10-31T00:41:42Z attila_lendvai joined #sbcl
2017-10-31T00:43:51Z stassats: picking up a big book on compilers is so daunting
2017-10-31T00:46:28Z pkhuong: stassats: IME, people tend to either always normalise (the type people especially), or normalise after small passes.
2017-10-31T00:47:23Z stassats: i don't think deleted code can pose a performance problem, just code clutter problem
2017-10-31T00:48:07Z stassats: and correctness
2017-10-31T00:48:20Z stassats: still no idea how to solve https://bugs.launchpad.net/sbcl/+bug/1661911
2017-10-31T00:48:29Z pfdietz: It would be annoying to get compiler warnings from dead code.
2017-10-31T00:49:11Z stassats: pfdietz: that's delayed until late enough
2017-10-31T00:49:44Z stassats: usually
2017-10-31T00:50:55Z stassats: ; caught WARNING: ;   Proclaiming SB!C::NODE-TO-BE-DELETED-P to be INLINE
2017-10-31T00:50:56Z stassats: hello
2017-10-31T01:19:58Z milanj quit (Quit: This computer has gone to sleep)
2017-10-31T02:04:32Z rpg quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
2017-10-31T02:42:13Z stassats quit (Ping timeout: 255 seconds)
2017-10-31T03:02:18Z rpg joined #sbcl
2017-10-31T03:08:35Z rpg quit (Quit: Textual IRC Client: www.textualapp.com)
2017-10-31T03:35:28Z Bike quit (Quit: Lost terminal)
2017-10-31T03:39:30Z attila_lendvai quit (Quit: Leaving.)
2017-10-31T03:50:32Z jack_rabbit quit (Quit: Leaving)
2017-10-31T03:59:17Z jack_rabbit joined #sbcl
2017-10-31T04:01:18Z shka joined #sbcl
2017-10-31T04:22:06Z adlai quit (Read error: Connection reset by peer)
2017-10-31T04:22:39Z adlai joined #sbcl
2017-10-31T04:36:06Z jack_rabbit quit (Remote host closed the connection)
2017-10-31T04:38:01Z jack_rabbit joined #sbcl
2017-10-31T04:46:06Z jrm quit (Read error: Connection reset by peer)
2017-10-31T04:46:27Z jrm joined #sbcl
2017-10-31T04:52:10Z jack_rabbit quit (Remote host closed the connection)
2017-10-31T04:55:15Z jack_rabbit joined #sbcl
2017-10-31T05:20:03Z milanj joined #sbcl
2017-10-31T05:41:22Z milanj quit (Quit: This computer has gone to sleep)
2017-10-31T05:55:13Z milanj joined #sbcl
2017-10-31T06:04:25Z milanj quit (Quit: Leaving)
2017-10-31T07:08:37Z sjl_ joined #sbcl
2017-10-31T07:11:13Z angavrilov joined #sbcl
2017-10-31T07:13:05Z sjl_ quit (Ping timeout: 248 seconds)
2017-10-31T07:20:35Z eschatologist joined #sbcl
2017-10-31T07:56:17Z Zisper joined #sbcl
2017-10-31T08:27:17Z Xof: argh, building the release stopped at updating the web page
2017-10-31T08:31:41Z shka quit (Ping timeout: 240 seconds)
2017-10-31T08:58:10Z sjl_ joined #sbcl
2017-10-31T09:02:27Z sjl_ quit (Ping timeout: 240 seconds)
2017-10-31T09:21:08Z bsmr joined #sbcl
2017-10-31T09:21:16Z bsmr left #sbcl
2017-10-31T09:27:41Z bsmr joined #sbcl
2017-10-31T09:39:12Z loke joined #sbcl
2017-10-31T09:39:15Z loke: Hello
2017-10-31T09:49:29Z minion quit (Remote host closed the connection)
2017-10-31T09:49:42Z minion joined #sbcl
2017-10-31T09:59:30Z specbot quit (Disconnected by services)
2017-10-31T09:59:34Z specbot joined #sbcl
2017-10-31T10:00:02Z liead joined #sbcl
2017-10-31T10:01:22Z nimiux_ joined #sbcl
2017-10-31T10:01:39Z adlai quit (Read error: Connection reset by peer)
2017-10-31T10:01:39Z minion quit (Read error: Connection reset by peer)
2017-10-31T10:01:39Z Xof quit (Ping timeout: 248 seconds)
2017-10-31T10:01:40Z nimiux quit (Ping timeout: 248 seconds)
2017-10-31T10:01:40Z jack_rabbit quit (Ping timeout: 248 seconds)
2017-10-31T10:01:41Z kini quit (Ping timeout: 248 seconds)
2017-10-31T10:01:52Z kini joined #sbcl
2017-10-31T10:02:01Z jack_rabbit joined #sbcl
2017-10-31T10:02:29Z minion joined #sbcl
2017-10-31T10:03:58Z brucem quit (Ping timeout: 240 seconds)
2017-10-31T10:07:08Z brucem joined #sbcl
2017-10-31T10:09:14Z fiveop joined #sbcl
2017-10-31T10:52:17Z attila_lendvai joined #sbcl
2017-10-31T11:08:00Z m00natic joined #sbcl
2017-10-31T11:10:34Z aeth quit (Read error: Connection reset by peer)
2017-10-31T11:12:05Z aeth joined #sbcl
2017-10-31T11:17:49Z whoman quit (Quit: Leaving)
2017-10-31T11:19:18Z Bike joined #sbcl
2017-10-31T12:20:34Z bsmr quit (Quit: Leaving)
2017-10-31T12:24:46Z bsmr joined #sbcl
2017-10-31T12:28:32Z Bike quit (Ping timeout: 246 seconds)
2017-10-31T12:37:56Z sjl_ joined #sbcl
2017-10-31T12:51:39Z Bike joined #sbcl
2017-10-31T12:52:46Z rgrau joined #sbcl
2017-10-31T13:08:26Z stassats joined #sbcl
2017-10-31T13:18:27Z bsmr quit (Quit: Leaving)
2017-10-31T13:18:27Z White_Flame quit (Ping timeout: 240 seconds)
2017-10-31T13:19:25Z sjl quit (Ping timeout: 255 seconds)
2017-10-31T13:20:12Z White_Flame joined #sbcl
2017-10-31T13:44:08Z jdz quit (Ping timeout: 240 seconds)
2017-10-31T13:47:19Z jdz joined #sbcl
2017-10-31T13:48:19Z sjl_ is now known as sjl
2017-10-31T14:25:27Z pfdietz left #sbcl
2017-10-31T14:45:06Z swflint quit (Ping timeout: 252 seconds)
2017-10-31T14:47:18Z swflint joined #sbcl
2017-10-31T15:22:01Z eschatologist quit (Ping timeout: 240 seconds)
2017-10-31T16:18:58Z Zisper quit (Ping timeout: 255 seconds)
2017-10-31T17:47:09Z eschatologist joined #sbcl
2017-10-31T17:52:21Z Jesin quit (Quit: Leaving)
2017-10-31T17:54:38Z m00natic quit (Remote host closed the connection)
2017-10-31T17:55:52Z Jesin joined #sbcl
2017-10-31T17:57:19Z stassats quit (Ping timeout: 248 seconds)
2017-10-31T17:59:30Z shka joined #sbcl
2017-10-31T18:12:02Z nimiux_ is now known as nimiux
2017-10-31T20:04:27Z rgrau quit (Ping timeout: 240 seconds)
2017-10-31T20:08:00Z fiveop quit (Quit: Lost terminal)
2017-10-31T20:23:48Z stassats joined #sbcl
2017-10-31T20:25:20Z stassats: i'm now able to fold (if x (print 10) (print 10)) to (print 10), previously only (if x 10 10) was folded
2017-10-31T20:26:19Z stassats: more points would be for (if x (progn (print 10) 1) (progn (print 10) 2)) => (progn (print 10) (if x 1 2))
2017-10-31T20:26:54Z stassats: and somewhere therein lies the path towards common sub-expression elimination
2017-10-31T20:28:56Z stassats: (if x (print 10) (print 20)) => (print (if x 10 20)) should be easy
2017-10-31T20:33:25Z stassats: but with more clever code deletion i need to stop code deletion notes
2017-10-31T20:40:33Z Posterdati joined #sbcl
2017-10-31T20:44:19Z angavrilov quit (Remote host closed the connection)
2017-10-31T20:45:26Z DeadTrickster_ quit (Remote host closed the connection)
2017-10-31T20:46:55Z shka quit (Ping timeout: 248 seconds)
2017-10-31T21:04:57Z sjl_ joined #sbcl
2017-10-31T21:05:30Z sjl___ joined #sbcl
2017-10-31T21:07:29Z sjl quit (Ping timeout: 252 seconds)
2017-10-31T21:09:37Z sjl_ quit (Ping timeout: 260 seconds)
2017-10-31T21:29:36Z sjl___ is now known as sjl
2017-10-31T22:03:11Z Bike quit (Ping timeout: 248 seconds)
2017-10-31T22:03:12Z ym quit (Quit: Leaving)
2017-10-31T22:04:47Z ym joined #sbcl
2017-10-31T22:10:58Z stassats: the cblocks-equivalent-p change would be perfect for the random tester
2017-10-31T22:11:09Z stassats: though i think i finally got it right
2017-10-31T22:20:32Z stassats: forking SBCL, how quaint
2017-10-31T22:20:40Z attila_lendvai quit (Quit: Leaving.)
2017-10-31T22:21:34Z stassats: i can't reply without being dismissive and sarcastic, i guess i won't reply at all
2017-10-31T22:47:57Z leah2 quit (Ping timeout: 246 seconds)
2017-10-31T22:55:27Z leah2 joined #sbcl
2017-10-31T23:02:57Z Bike joined #sbcl
2017-10-31T23:15:07Z cromachina joined #sbcl
2017-10-31T23:17:35Z whoman joined #sbcl