00:04:53 -!- oleo [~oleo@xdsl-78-35-175-179.netcologne.de] has quit [Ping timeout: 244 seconds] 00:05:24 oleo [~oleo@xdsl-84-44-209-36.netcologne.de] has joined #sbcl 00:09:34 White_Flame: fwiw, I thought that was funny 00:11:35 -!- foom [jknight@nat/google/x-youbdbbrrmpocscc] has quit [Ping timeout: 264 seconds] 00:16:11 i would like to commit http://paste.lisp.org/display/141520. any objections? 00:17:38 i object! 00:18:30 i no longer object, since now _you_ have to merge the NEWS file 00:23:48 haha, did you just push something? 00:23:55 yep 00:24:22 enjoy your 3MB larger core files 00:25:36 #'struct-slot is 10 bytes larger than (lambda (x) (struct-slot x)), because #'struct-slot performs the type checks in the XEP 00:26:32 a way to dissuade it from doing that would be EXPLICIT-CHECK in defknown, but it doesn't use defknown, but proclaim 00:26:38 and defknown isn't suited for this 00:33:11 foom [jknight@nat/google/x-sxhlneuzsfkxlogv] has joined #sbcl 00:38:21 *nyef* sighs. 00:38:29 I'm going to have to merge NEWS again, aren't I? 00:39:56 On the other hand, I still have to figure out what's going on with a lot more of IR1 before I can really do anything with my current pending patches. 00:43:33 -!- DalekBaldwin [~user@ip68-231-209-84.oc.oc.cox.net] has quit [Read error: Connection reset by peer] 00:43:53 DalekBaldwin [~user@ip68-231-209-84.oc.oc.cox.net] has joined #sbcl 00:44:34 *stassats* envisions a trick to save some space: when checking the type and signaling the error, use # for the error, instead of 'X, since the layout is already stored in the boxed constants for type-checking, that will save space 00:45:26 Plausible, yes. 00:45:29 here comes -64KB 00:46:12 Only 2998KB to go! 00:46:54 i also need to add some code to make the error to be displayed the same way 00:48:32 now, if it saved 640KB y'all talking about, i would've committed it 00:48:56 but 64 is not worth the complication 00:48:59 :) 00:49:50 i can dust off my VOP machinery changes, they are worth 800KB of savings and potentially more 00:49:52 -!- slyrus [~chatzilla@dsl092-019-253.sfo1.dsl.speakeasy.net] has quit [Ping timeout: 252 seconds] 00:51:08 Correctness before speed or space. 00:52:07 On the other hand, I'm not happy with the way my correctness fix (for what seems to be a fairly rare condition) breaks tail-call optimization, so... 00:54:26 is there a clever algorithm for checking types when it's subclassed? 00:54:36 there must be something more clever than iterative over a list 00:55:00 though it will be probably too clever for sbcl 00:56:33 -!- segv- [~mb@95-91-242-0-dynip.superkabel.de] has quit [Ping timeout: 244 seconds] 01:04:20 apparently, the whole typechecks thing is now not needed 01:04:30 typecheckfuns 01:04:59 or at least some part of it related to structures 01:06:04 i'm currently running tests for http://paste.lisp.org/display/141521 and would like to commit it if nobody objects 01:06:40 you commit, anyone wanting to object will do that later 01:07:07 stassats: ok, thanks 01:07:23 lp 1264902 01:07:23 https://bugs.launchpad.net/bugs/1264902 01:08:43 Ah, no wonder it's unfamiliar, there's no confirmation and no priority assigned. 01:09:16 12-30, everyone was partying hard 01:09:22 to notice 01:09:43 how do confirmation and priority assignment usually happen? 01:10:33 reproduce => think up of tags => set to Triage 01:11:39 or see that it's bollocks, set to Invalid, write up a message trying to belittle the reporter 01:13:30 On the other hand, I'm thinking that eight of the fifteen "high" importance bugs aren't really very high importance. 01:13:36 ok, deleting src/code/typecheckfuns.lisp saves 32KB and "WTF is this?" 01:14:49 stassats: thanks for the explanation 01:17:25 *stassats* has a feeling that his defstruct decimation may end horribly 01:18:34 but my whole premise seems to be sound 01:19:26 a) first class functions to find/remove are common, b) (setf (documentation #'closure) "doc") is broken 01:21:57 ... I thought that nikodemus did something about the latter? 01:22:18 yes, about (setf (documentation 'closure) "doc") 01:23:38 Ahh. The difference between the symbol and the function? 01:24:15 right, the only way to attach it to a closure would be either through a hashtable or through an additional slot 01:25:09 #'hash-table-test => # is now #, which is nice 01:28:40 and i deleted about 500 lines, which makes things simpler 01:28:59 though !DEFSTRUCT-WITH-ALTERNATE-METACLASS still scares me 01:29:39 fridim_ [~fridim@65.93.77.146] has joined #sbcl 01:33:47 You know, we could add a boxed slot to the SIMPLE-FUN primobj. 01:37:13 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 240 seconds] 01:41:53 i accidentally committed trailing-whitespace-cleanup in NEWS. sorry for the large diff 02:02:59 -!- ASau [~user@46.115.48.210] has quit [Ping timeout: 240 seconds] 17:18:04 ccl-logbot [~ccl-logbo@setf.clozure.com] has joined #sbcl 17:18:04 17:18:04 -!- names: ccl-logbot slyrus senj fiveop davazp jiacobucci1 specbot minion kludge` stassats` fridim__ xani prxq oleo |3b| Bike segv- Blkt foom reb White_Flame ams bege fikusz christoph_deb ubii angavrilov edgar-rft milosn fe[nl]ix psilord jdz @Krystof xymox Posterdati ivan`` luis nicdev scymtym redline6561 izirku asedeno kanru antoszka yauz jsnell daimrod pkhuong mood faheem hzp joshe ferada Munksgaard brucem jaimef flip214 pchrist 17:21:51 attila_lendvai [~attila_le@5.251.216.208] has joined #sbcl 17:21:52 -!- attila_lendvai [~attila_le@5.251.216.208] has quit [Changing host] 17:21:52 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 17:35:43 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 17:40:14 -!- jdz [~jdz@212.36.34.246] has quit [Quit: Leaving] 18:15:12 drmeister [~drmeister@155.247.96.196] has joined #sbcl 18:19:44 -!- drmeister [~drmeister@155.247.96.196] has quit [Remote host closed the connection] 18:22:35 drmeister [~drmeister@155.247.96.196] has joined #sbcl 18:24:09 -!- senj [~senj@unaffiliated/senj] has quit [Quit: Sleep Now] 18:25:46 drmeiste_ [~drmeister@155.247.96.196] has joined #sbcl 18:25:49 -!- drmeister [~drmeister@155.247.96.196] has quit [Read error: Connection reset by peer] 18:38:44 leuler [~user@p548FA040.dip0.t-ipconnect.de] has joined #sbcl 18:48:51 i cannot look closer now, but there may be a regression: http://cl-test-grid.appspot.com/blob?key=19tnnpinqb 18:49:44 another one: http://cl-test-grid.appspot.com/blob?key=1i4459ffbs 18:50:18 ok 18:50:22 there seem to be more 18:50:29 i'll have a closer look later 18:50:55 i'll start with cl-store 18:52:46 cl-store is doing something funky with structures 19:01:25 fnv is broken without my help, i only uncover it 19:01:52 it defines a copy function manually, while not specifier :copier nil 19:02:47 scymtym: what else do you have? 19:10:13 pullrequested fnv 19:10:51 ccl full-warned about this before, but asdf pays attention to the third return of C-F only on SBCL 19:22:45 and opened a cl-store issue, the way it uses internals is not great 19:23:09 more reasons to change internals! don't let people get too comfortable 19:25:20 -!- drmeiste_ [~drmeister@155.247.96.196] has quit [Ping timeout: 244 seconds] 19:33:36 ppc fails somewhere at disassemble, i smell bootstrap issues 19:33:36 Sorry, I couldn't find anything for fails somewhere at disassemble, i smell bootstrap issues. 19:35:20 there have been changes recently in that area 19:35:50 yep 19:46:55 i don't really get why it failed 19:49:15 senj [~senj@unaffiliated/senj] has joined #sbcl 19:49:34 it looks like it should work but it doesn't 19:50:19 scymtym_ [~user@ip-78-94-192-97.unitymediagroup.de] has joined #sbcl 19:51:02 stassats`: sorry, i had to ride home first 19:52:16 there is also this: http://cl-test-grid.appspot.com/blob?key=1ak9w81jlh 19:53:16 kotovy [~kotovy@91.197.131.143] has joined #sbcl 19:53:20 these seem to be the root causes for the new failures 20:03:49 -!- fridim__ [~fridim@65.93.77.146] has quit [Read error: Operation timed out] 20:04:03 stassats`: unused variables? 20:04:09 that's a commonish bootstrap failure mode 20:14:48 -!- edgar-rft [~GOD@HSI-KBW-109-193-013-113.hsi7.kabel-badenwuerttemberg.de] has quit [Quit: conversation terminated into permanent darkness] 20:19:17 no, undefined function 20:19:29 in the built image 20:19:40 define-instruction-format is supposed to define that function 20:21:14 scymtym_: defstruct previously defined its functions with (setf fdefinition), now it expands into defun 20:21:23 that's why all the duplicate warnings 20:22:24 stassats`: thanks, i still couldn't have a closer look 20:22:36 did already file all the necessary bugs? 20:22:46 cl-store and fnv 20:23:05 paiprolog, not sure where to file 20:26:11 https://github.com/quek/paiprolog 20:26:14 apparently 20:27:16 so, it redefines a structure with the same name, but with a different default value and constructor 20:27:20 i have no idea why 20:29:56 ok, reported 20:30:22 yet another instance of stricter warnings breaking bad code, good 20:43:00 back to ppc, x86 has the same kind of changes and yet works fine 20:44:04 oh, eval-when :compile-toplevel :execute 20:45:41 i'll defer to Douglas, i don't really want to think about the disassembler 20:47:19 -!- davazp [~user@177.Red-83-63-237.staticIP.rima-tde.net] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 20:55:49 stassats`: how come removing proclaim saves space? Don't redundant typechecks get removed before codegen anyhow? 20:56:51 foom: proclaim information itself takes space, and proclaim causes the type check to be performed inside XEP, not in the body, which is because of jumps and other things takes 10 bytes more space 20:58:05 Okay, and normally the proclaim information would be valuable, since callers can depend on it, but in this case, any caller that would use that already gets the inline transform applied. 20:58:15 right 20:58:24 Got it. 21:03:09 and there are about 2000 struct accessors in sbcl 21:05:56 some duplication could be resolved if structures were classes with a different metaclass, because it still defines a class for a structure 21:07:58 -!- kotovy [~kotovy@91.197.131.143] has quit [Ping timeout: 240 seconds] 21:21:24 -!- angavrilov [~angavrilo@217.71.227.190] has quit [Remote host closed the connection] 21:27:15 -!- fiveop [~fiveop@p5DDC79F5.dip0.t-ipconnect.de] has quit [] 21:39:21 pnpuff [~harmonic@unaffiliated/pnpuff] has joined #sbcl 21:40:06 -!- pnpuff [~harmonic@unaffiliated/pnpuff] has left #sbcl 21:44:18 -!- bege [~bege@S0106001d7e5132b0.ed.shawcable.net] has quit [Quit: leaving] 21:54:39 milanj [~milanj@cable-178-148-8-79.dynamic.sbb.rs] has joined #sbcl 22:01:00 ASau [~user@p54AFF665.dip0.t-ipconnect.de] has joined #sbcl 22:05:33 -!- leuler [~user@p548FA040.dip0.t-ipconnect.de] has quit [Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)] 22:17:37 sdemarre [~serge@91.180.106.182] has joined #sbcl 22:22:01 -!- segv- [~mb@95-91-211-99-dynip.superkabel.de] has quit [Ping timeout: 265 seconds] 22:35:38 -!- sdemarre [~serge@91.180.106.182] has quit [Ping timeout: 240 seconds] 23:06:16 -!- milanj [~milanj@cable-178-148-8-79.dynamic.sbb.rs] has quit [Quit: Leaving] 23:09:02 -!- christoph_deb [~christoph@ppp-46-244-238-183.dynamic.mnet-online.de] has quit [Ping timeout: 244 seconds] 23:17:02 -!- xani [~user@178.183.148.36.dsl.dynamic.t-mobile.pl] has quit [Remote host closed the connection] 23:19:04 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 23:45:38 cosmology [~cosmology@188.190.125.28] has joined #sbcl 23:48:47 -!- senj [~senj@unaffiliated/senj] has quit [Quit: Sleep Now] 23:59:56 senj [~senj@unaffiliated/senj] has joined #sbcl