2014-10-14T00:01:16Z Bicyclidine quit (Ping timeout: 260 seconds)
2014-10-14T00:34:33Z psilord joined #sbcl
2014-10-14T02:27:32Z scymtym_ quit (Ping timeout: 250 seconds)
2014-10-14T02:27:58Z slyrus quit (Ping timeout: 258 seconds)
2014-10-14T02:38:26Z christoph_debian quit (Ping timeout: 240 seconds)
2014-10-14T02:46:26Z specbot quit (Disconnected by services)
2014-10-14T02:46:30Z specbot joined #sbcl
2014-10-14T02:48:38Z flip214_ joined #sbcl
2014-10-14T02:49:03Z antoszka_ joined #sbcl
2014-10-14T02:49:34Z brucem_ joined #sbcl
2014-10-14T02:49:44Z milosn_ joined #sbcl
2014-10-14T02:50:11Z scymtym_ joined #sbcl
2014-10-14T02:51:05Z brucem_ quit (Changing host)
2014-10-14T02:51:05Z brucem_ joined #sbcl
2014-10-14T02:51:34Z brucem quit (Disconnected by services)
2014-10-14T02:51:37Z brucem_ is now known as brucem
2014-10-14T02:52:03Z christoph_debian joined #sbcl
2014-10-14T02:53:44Z Hache_ quit (Remote host closed the connection)
2014-10-14T02:54:00Z oleo__ quit (*.net *.split)
2014-10-14T02:54:00Z loke_ quit (*.net *.split)
2014-10-14T02:54:01Z milosn quit (*.net *.split)
2014-10-14T02:54:01Z scymtym quit (*.net *.split)
2014-10-14T02:54:01Z ams quit (*.net *.split)
2014-10-14T02:54:02Z flip214 quit (*.net *.split)
2014-10-14T02:54:02Z weissschloss quit (*.net *.split)
2014-10-14T02:54:02Z akkad quit (*.net *.split)
2014-10-14T02:54:03Z joshe quit (*.net *.split)
2014-10-14T02:54:03Z antoszka quit (*.net *.split)
2014-10-14T02:54:03Z |3b| quit (*.net *.split)
2014-10-14T02:54:37Z akkad joined #sbcl
2014-10-14T03:00:57Z oleo__ joined #sbcl
2014-10-14T03:06:17Z weissschloss joined #sbcl
2014-10-14T03:06:30Z loke_ joined #sbcl
2014-10-14T03:19:23Z LiamH quit (Quit: Leaving.)
2014-10-14T03:29:08Z nyef quit (Quit: G'night all.)
2014-10-14T03:35:13Z echo-area joined #sbcl
2014-10-14T05:01:58Z oleo__ quit (Quit: Verlassend)
2014-10-14T05:22:28Z pranavrc joined #sbcl
2014-10-14T05:22:28Z pranavrc quit (Changing host)
2014-10-14T05:22:28Z pranavrc joined #sbcl
2014-10-14T05:38:04Z edgar-rft quit (Quit: lifeform experiment abandoned because continuation corrupted)
2014-10-14T05:40:39Z jsnell_ quit (*.net *.split)
2014-10-14T05:40:39Z irsol quit (*.net *.split)
2014-10-14T05:46:08Z jsnell_ joined #sbcl
2014-10-14T05:46:08Z irsol joined #sbcl
2014-10-14T05:49:33Z karswell` quit (Ping timeout: 246 seconds)
2014-10-14T06:22:47Z prxq joined #sbcl
2014-10-14T06:27:25Z aldo joined #sbcl
2014-10-14T06:32:51Z slyrus joined #sbcl
2014-10-14T06:36:15Z aldo quit (Ping timeout: 272 seconds)
2014-10-14T06:47:34Z Krystof: my recollection too is that punycode is simple
2014-10-14T06:47:36Z Krystof: doable by hand
2014-10-14T06:48:01Z Krystof: of course that's a recollection from a decade ago so we'll probably find it's been extended to do three other unrelated things
2014-10-14T07:18:32Z fridim_ joined #sbcl
2014-10-14T07:22:14Z DGASAU quit (Read error: Connection reset by peer)
2014-10-14T07:23:13Z DGASAU joined #sbcl
2014-10-14T07:40:11Z antoszka_ is now known as antoszka
2014-10-14T08:18:42Z DGASAU quit (Read error: Connection reset by peer)
2014-10-14T08:19:20Z |3b|`` joined #sbcl
2014-10-14T08:19:21Z DGASAU joined #sbcl
2014-10-14T08:20:55Z |3b|`` is now known as |3b|
2014-10-14T08:29:35Z ams joined #sbcl
2014-10-14T08:40:21Z stassats joined #sbcl
2014-10-14T09:40:06Z scymtym joined #sbcl
2014-10-14T10:45:12Z edgar-rft joined #sbcl
2014-10-14T11:03:37Z scymtym quit (Ping timeout: 272 seconds)
2014-10-14T11:09:12Z echo-area quit (Remote host closed the connection)
2014-10-14T11:50:30Z scymtym_: i think i have a mostly complete implementation of the IDNA encoding direction (toASCII ≅ (punycode o nameprep)) based andreas fuchs' idna system
2014-10-14T11:50:56Z scymtym_: Krystof may be happy to hear that nameprep makes heavy use of the new sb-unicode facilities
2014-10-14T12:10:40Z eudoxia joined #sbcl
2014-10-14T12:22:20Z stassats quit (Ping timeout: 260 seconds)
2014-10-14T12:39:54Z stassats joined #sbcl
2014-10-14T12:53:06Z cmack joined #sbcl
2014-10-14T12:57:53Z Hache_ joined #sbcl
2014-10-14T12:59:42Z Krystof: \o/
2014-10-14T12:59:49Z Krystof: krzysz00 will also be happy to hear that
2014-10-14T13:00:11Z Krystof: meanwhile, in reproducible build land
2014-10-14T13:00:16Z psilord quit (Quit: Leaving.)
2014-10-14T13:00:18Z Krystof: ; Unboxed data:
2014-10-14T13:00:18Z Krystof: ; DF8:       000000000000D043
2014-10-14T13:00:18Z Krystof: ; E00:       0000805E
2014-10-14T13:00:20Z Krystof: vs
2014-10-14T13:00:27Z Krystof: ; DF8:       000000FFFFFFBF41
2014-10-14T13:00:27Z Krystof: ; E00:       0000004E
2014-10-14T13:02:36Z Krystof: oh, hey, look
2014-10-14T13:02:37Z Krystof: ;;; FIXME: Is it right to have TYPEP (and TYPE-OF, elsewhere; and
2014-10-14T13:02:37Z Krystof: ;;; perhaps SPECIAL-OPERATOR-P and others) be FOLDABLE in the
2014-10-14T13:02:38Z Krystof: ;;; cross-compilation host?
2014-10-14T13:03:53Z Krystof: (load-time-value 1f9 t) ; I don't think that works as expected either
2014-10-14T13:04:21Z Krystof: argh, #.(float most-positive-fixnum 1f0)
2014-10-14T13:04:26Z stassats: that's mine, isn't it?
2014-10-14T13:04:38Z Krystof: yes
2014-10-14T13:04:48Z stassats: without ltv it forces to cons it at runtime
2014-10-14T13:04:58Z Krystof: sure, we need ltv
2014-10-14T13:05:10Z Krystof: I think we need (load-time-value (make-single-float <bits>) t)
2014-10-14T13:06:01Z stassats: i'm thinking to stop bothering with consing and just make a C wrapper which calculates things correctly
2014-10-14T13:06:45Z stassats: so, sleep_single_float(), sleep_double_float(), sleep_ratio()
2014-10-14T13:07:00Z Krystof: totally happy with that
2014-10-14T13:07:16Z stassats: because this monstrosity still conses somewhere
2014-10-14T13:07:22Z Krystof: it is utterly horrible
2014-10-14T13:07:22Z stassats: on some targets, that is
2014-10-14T13:07:29Z Krystof: and I think has about 4 separate host-xc leaks
2014-10-14T13:08:25Z stassats: yeah, (typep 1000000000 'fixnum) doesn't sound great either
2014-10-14T13:09:32Z Krystof: ok, I will let you work on a reimplementation of this
2014-10-14T13:09:35Z Krystof: thank you
2014-10-14T13:24:46Z oleo joined #sbcl
2014-10-14T13:35:36Z Krystof: remember this number: 8579
2014-10-14T13:59:09Z DGASAU quit (Remote host closed the connection)
2014-10-14T14:01:34Z DGASAU joined #sbcl
2014-10-14T14:03:44Z pranavrc quit
2014-10-14T14:07:32Z pkhuong: Krystof: do you have architectural tricks you'd use to ensure bitwise reproducibility, if you were to do things from scratch?
2014-10-14T14:08:29Z pkhuong: ways to make it easier to binary search on files/forms and then on compiler phases, for example.
2014-10-14T14:09:38Z stassats: building each commit and seeing where it differs is a simple way
2014-10-14T14:10:39Z Krystof: pkhuong: emitting enough metainformation about fops that made writing an emacs tool for analysing xc fasls would probably be the first thing
2014-10-14T14:10:56Z Krystof: normally, once I've identified the symptom and the toplevel form it corresponds to, the problem is relatively obvious
2014-10-14T14:11:29Z stassats: but this failures are non-obvious, wee need an implementation with completely crooked constants and behaviours
2014-10-14T14:11:47Z stassats: or rather, some failures are not exposed by ccl or clisp
2014-10-14T14:12:13Z Krystof: (though, the last one I've just fixed is a counterexample: it took me for ever to think of the transform for concatenate)
2014-10-14T14:12:15Z stassats: grumpy cl, really 1024-long arrays, really only 40 arguments, 16-bit fixnums
2014-10-14T14:12:20Z stassats: or was it 50
2014-10-14T14:12:32Z Krystof: I'm not worried about fixnums any more
2014-10-14T14:13:00Z Krystof: we build essentially bitwise-identically under hosts with n different fixnum ranges
2014-10-14T14:13:20Z psilord joined #sbcl
2014-10-14T14:13:26Z Krystof: stable set operations, hash table iterators, and similar would be helpful
2014-10-14T14:13:46Z stassats: and re sleep differences, i don't think i'll be able to work on it soon, a) not much time, b) it needs do something different on windows, and i don't have it handy
2014-10-14T14:14:05Z Krystof: ok, I will try to construct a "correct"ish lisp versoin
2014-10-14T14:14:22Z stassats: will only get home in the november, so, not much sbclish things until then
2014-10-14T14:15:06Z scymtym_: does it make sense to automate core comparison for clisp and ccl hosts or are there remaining known differences?
2014-10-14T14:15:14Z Krystof: there are remaining differences
2014-10-14T14:15:16Z Krystof: down to 2 files
2014-10-14T14:15:21Z scymtym_: i see, thanks
2014-10-14T14:15:37Z Krystof: then there might be lurking gotchas in genesis
2014-10-14T14:20:25Z krzysz00 quit (Ping timeout: 260 seconds)
2014-10-14T14:38:21Z pkhuong: I can add stable set/hash ops.
2014-10-14T14:39:03Z Krystof: at this point I'm not so worried
2014-10-14T14:39:06Z Krystof: I'm nearly done
2014-10-14T14:45:06Z attila_lendvai joined #sbcl
2014-10-14T14:45:06Z attila_lendvai quit (Changing host)
2014-10-14T14:45:06Z attila_lendvai joined #sbcl
2014-10-14T14:48:55Z Krystof: csr21@aleph-null:/tmp/tmp.RTBcQIdqYE$ diff -qr {sbcl,ccl32}/obj | wc -l
2014-10-14T14:48:55Z Krystof: 1
2014-10-14T14:49:58Z Krystof: so
2014-10-14T14:49:59Z Krystof: close
2014-10-14T14:57:55Z kanru` joined #sbcl
2014-10-14T14:59:39Z Krystof: though, the answer is "3 files differ" between sbcl and ccl64, which is quite interesting
2014-10-14T15:02:46Z nyef joined #sbcl
2014-10-14T15:03:25Z eudoxia quit (Read error: Connection reset by peer)
2014-10-14T15:04:25Z eudoxia joined #sbcl
2014-10-14T16:05:39Z krzysz00 joined #sbcl
2014-10-14T16:12:25Z scymtym_ quit (Read error: Connection reset by peer)
2014-10-14T16:12:57Z scymtym_ joined #sbcl
2014-10-14T16:17:37Z Krystof: argh, this is driving me crazy
2014-10-14T16:19:05Z nyef: Oh?
2014-10-14T16:19:33Z Krystof: oh
2014-10-14T16:19:35Z Krystof: oh
2014-10-14T16:19:37Z Krystof: oh wait
2014-10-14T16:19:47Z Krystof: hm
2014-10-14T16:20:25Z nyef: Entertaining, but uninformative.
2014-10-14T16:21:17Z Krystof: this fasl has a bunch of ^C5$@$@^CE^CA1^E^A sequences in it
2014-10-14T16:21:38Z Krystof: loosely, that translates as an alien-pointer-type
2014-10-14T16:22:02Z Krystof: the fasl with sbcl as a host occasionally reuses one
2014-10-14T16:22:12Z Krystof: the fasl with ccl32 as a host behaves identically
2014-10-14T16:22:52Z nyef: So... constant coalescence?
2014-10-14T16:22:53Z Krystof: the fasl with ccl64 as a host behaves almost identially, but there's a difference 100kB or so into the fasl
2014-10-14T16:23:12Z nyef: Hrm.
2014-10-14T16:24:52Z eudoxia quit (Quit: Lost terminal)
2014-10-14T16:25:05Z Krystof: I wonder if it's a not-quite-portable cache
2014-10-14T16:25:56Z nyef: With a bit of luck, it'll turn out to be a CCL bug.
2014-10-14T16:26:25Z stassats: minion: advice on compiler?
2014-10-14T16:26:25Z minion: #12000: Looking for a compiler bug is the second-to-last resort. The last resort is blaming bad RAM. It's never the correct hypothesis.
2014-10-14T16:27:31Z nyef: I love that one, if only because, as compiler maintainers, a great number of our bugs are, by definition, compiler bugs.
2014-10-14T16:28:03Z stassats: even without being a maintainer i regulary manage to find compiler bugs
2014-10-14T16:28:28Z stassats: i guess that speaks about the quality of CL compilers...
2014-10-14T16:29:16Z nyef: I still think that it'd be interesting to build a verified CL compiler.
2014-10-14T16:39:26Z loke_ quit (Read error: Connection reset by peer)
2014-10-14T16:43:33Z pkhuong: stassats: I feel like once you start working on a compiler, you get better at tickling compiler bugs
2014-10-14T16:43:59Z pkhuong: My GCC ICE (and outright coredumps) rate is higher than everyone else I work with.
2014-10-14T16:44:17Z stassats: well, maybe you become better at recognizing that it's a compiler bug and not just your code
2014-10-14T16:54:14Z Krystof: :hash-function #'sxhash
2014-10-14T16:54:16Z Krystof: hm hm hm
2014-10-14T16:59:54Z Hache_: what?
2014-10-14T17:00:02Z Hache_: can a custom hash function be used?
2014-10-14T17:00:15Z stassats: of course!
2014-10-14T17:00:35Z Hache_: but
2014-10-14T17:00:36Z Hache_: but
2014-10-14T17:00:40Z Hache_: is it sbcl specific?
2014-10-14T17:00:46Z stassats: yes
2014-10-14T17:00:52Z Hache_: oh
2014-10-14T17:00:53Z stassats: this is #sbcl
2014-10-14T17:00:58Z Hache_: yep
2014-10-14T17:01:35Z fridim_ quit (Ping timeout: 258 seconds)
2014-10-14T17:02:09Z Hache_: cool
2014-10-14T17:10:02Z stassats quit (Quit: ERC Version 5.3 (IRC client for Emacs))
2014-10-14T17:16:21Z krzysz00 quit (Ping timeout: 260 seconds)
2014-10-14T17:28:26Z karswell` joined #sbcl
2014-10-14T17:32:13Z kanru` quit (Ping timeout: 255 seconds)
2014-10-14T18:05:33Z Hache_ quit (Remote host closed the connection)
2014-10-14T18:25:16Z akkad quit (Excess Flood)
2014-10-14T18:25:52Z slyrus quit (Ping timeout: 260 seconds)
2014-10-14T18:25:55Z akkad joined #sbcl
2014-10-14T18:45:09Z slyrus joined #sbcl
2014-10-14T18:46:55Z fitzsim joined #sbcl
2014-10-14T18:59:09Z krzysz00 joined #sbcl
2014-10-14T19:25:58Z scymtym joined #sbcl
2014-10-14T19:43:53Z krzysz00: scymtym: Hove you pushed your IDNA stuff yet?
2014-10-14T20:15:01Z scymtym: krzysz00: not yet. i was going to make a list paste for review later today
2014-10-14T20:18:48Z krzysz00 quit (Ping timeout: 255 seconds)
2014-10-14T20:25:18Z slyrus quit (Ping timeout: 272 seconds)
2014-10-14T20:26:40Z krzysz00 joined #sbcl
2014-10-14T20:35:10Z krzysz00 quit (Ping timeout: 265 seconds)
2014-10-14T20:38:21Z slyrus joined #sbcl
2014-10-14T20:57:11Z krzysz00 joined #sbcl
2014-10-14T21:02:13Z gingerale quit (Ping timeout: 260 seconds)
2014-10-14T21:09:13Z slyrus quit (Ping timeout: 258 seconds)
2014-10-14T21:23:00Z slyrus joined #sbcl
2014-10-14T21:24:27Z scymtym: sb-idna work in progress is at http://paste.lisp.org/display/144040; punycode-encode is based on the system idna by andreas fuchs
2014-10-14T21:30:40Z prxq quit (Ping timeout: 260 seconds)
2014-10-14T21:35:00Z karswell` quit (Ping timeout: 250 seconds)
2014-10-14T21:48:35Z reb quit (Remote host closed the connection)
2014-10-14T21:54:15Z psilord quit (Quit: Leaving.)
2014-10-14T21:59:07Z krzysz00 quit (Ping timeout: 272 seconds)
2014-10-14T22:08:02Z edgar-rft quit (Quit: lifeform experiment abandoned into paranoid damage)
2014-10-14T22:13:04Z krzysz00 joined #sbcl
2014-10-14T22:16:46Z reb joined #sbcl
2014-10-14T22:21:51Z reb quit (Remote host closed the connection)
2014-10-14T22:26:21Z krzysz00 quit (Ping timeout: 246 seconds)
2014-10-14T22:31:49Z brown joined #sbcl
2014-10-14T22:32:05Z brown is now known as reb``
2014-10-14T22:52:00Z oleo is now known as Guest27251
2014-10-14T22:52:03Z krzysz00 joined #sbcl
2014-10-14T22:52:54Z oleo__ joined #sbcl
2014-10-14T22:53:30Z psilord joined #sbcl
2014-10-14T22:55:01Z Guest27251 quit (Ping timeout: 258 seconds)
2014-10-14T23:02:41Z krzysz00 quit (Ping timeout: 260 seconds)
2014-10-14T23:21:24Z attila_lendvai quit (Quit: Leaving.)