Camelia, the Perl 6 bug

IRC log for #parrot, 2009-02-03

Parrot | source cross referenced

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:04 nopaste "kid51" at 70.85.31.226 pasted "Build failure at r36305" (431 lines) at http://nopaste.snit.ch/15487
00:08 chromatic Confirmed.
00:08 purl confirmed is, like, "i've been told by the person", speculative = got it from some other .git or the prev conversion, from github, and from pause
00:08 rg forget confirmed
00:08 purl rg: I forgot confirmed
00:08 kid51 NotFound:  ping
00:09 NotFound pong
00:09 AndyA joined #parrot
00:10 kid51 I believe one of your recent commits causes the build error I just pasted.
00:10 kid51 r36305, probably.
00:12 NotFound One moment...
00:12 chromatic include/../io/io_private.h should be $(IO_DIR)/io_private.h
00:14 kid51 src/pmc/parrotinterpreter.pmc
00:14 chromatic in the Makefile at least
00:14 chromatic ... but the PMC probably shouldn't use io_private.h symbols.
00:14 * kid51 starts to prepare dinner
00:15 NotFound Can "../../src/io/io_private.h" be a quick fix?
00:16 Whiteknight io_private.h is supposed to be private to the IO subsystem only
00:16 Whiteknight if you're using something from there outside src/io/, then it isn't private and you should put it in a differnt .h file
00:17 NotFound Whiteknight: I know, but this is a quick attempt to test the feature
00:17 Whiteknight ah, okay </unnecessary lecture>
00:19 dalek parrot: r36306 | NotFound++ | trunk/src/pmc/parrotinterpreter.pmc:
00:20 dalek parrot: attempt to quick fix the bug introduced in r36305
00:20 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36306/
00:24 kid51 NotFound:  If this is an experimental feature, why not develop it in a branch rather than trunk?
00:25 Whiteknight branches are for suckers
00:25 * Whiteknight kids, of course
00:26 NotFound kid51: is very short, and if I open a branch no one will test it ;)
00:26 kid51 *This* kid has used lots of branches
00:26 Whiteknight I want to make a branch to do some more CC stuff, but there are too many test failures right now
00:26 NotFound Sorry for the inconvenience, In my build systems it compiled fine.
00:26 Whiteknight I won't get clean regression testing info like this
00:27 kid51 Granted, all that t/native_pbc/ stuff should have been done in a branch as well.
00:28 Whiteknight definitely did need doin' though
00:28 skv joined #parrot
00:28 kid51 But I have spent many cycles dealing with those failures over past 3 days.  I could have been working on other problems.
00:28 kid51 I could have watched the Super Bowl!
00:29 kid51 I could have been a contender!
00:29 Whiteknight I'm a philadelphia fan, the football season ended for me two weeks ago
00:30 * jhorwitz concurs
00:30 kid51 NotFound:  As a big advocate of branches, I would have tested anything you developed in a branch and requested help with testing.
00:31 hercynium joined #parrot
00:34 jhorwitz NotFound: still fails for me
00:38 TiMBuS joined #parrot
00:39 Whiteknight it's working like a charm on my system. NotFound, what are you trying to add?
00:41 NotFound Ok, i'll comment it out until doing it on a better way.
00:44 dalek parrot: r36307 | NotFound++ | trunk/src/pmc/parrotinterpreter.pmc:
00:44 NotFound Whiteknight: a way to redirect the standard in, out and err handles
00:45 dalek parrot: commented out temporarily the code inserted in r36305
00:45 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36307/
00:51 kid51 Same error at 36307:  make: *** No rule to make target `include/../io/io_private.h', needed by `src/pmc/parrotinterpreter.o'.  Stop.
00:52 * jhorwitz too  :(
00:53 dalek parrot: r36308 | allison++ | branches/pdd28str_part2:
00:53 dalek parrot: [pdd28str] Renaming string conversion functions.
00:53 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36308/
01:00 NotFound A make dependence? I don't know from that comes from :?
01:10 dalek parrot: r36309 | NotFound++ | trunk/src/pmc/parrotinterpreter.pmc:
01:10 dalek parrot: commented out temporarily the code inserted in r36305 - harder approach, you may need a realclean
01:10 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36309/
01:11 chromatic Just a reconfigure.
01:17 dalek parrot: r36310 | allison++ | branches/pdd28str_part2:
01:17 dalek parrot: [pdd28str] Renaming memory-related string functions.
01:17 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36310/
01:22 s1n Coke: my rt id is s1n as well
01:41 TiMBuS joined #parrot
02:10 dalek parrot: r36311 | allison++ | branches/pdd28str_part2:
02:11 dalek parrot: [pdd28str] Renaming case-changing string functions.
02:11 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36311/
02:16 Limbic_Region joined #parrot
02:34 shragei joined #parrot
02:45 ask_ joined #parrot
03:19 jimk joined #parrot
03:21 jimk joined #parrot
03:23 kid51 joined #parrot
03:30 bacek_ joined #parrot
03:32 gravity joined #parrot
03:39 skv joined #parrot
04:10 Fayland joined #parrot
04:20 jimmy joined #parrot
04:23 dalek parrot: r36312 | allison++ | branches/pdd28str_part2:
04:23 dalek parrot: [pdd28str] Renaming character class searching string functions.
04:23 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36312/
04:41 dalek parrot: r36313 | allison++ | branches/pdd28str_part2:
04:41 dalek parrot: [pdd28str] Renaming string functions for changing charset and encoding.
04:41 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36313/
04:58 dalek parrot: r36314 | allison++ | branches/pdd28str_part2:
04:58 dalek parrot: [pdd28str] Renaming a collection of string creation/manipulation functions.
04:58 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36314/
05:12 dalek parrot: r36315 | allison++ | branches/pdd28str_part2:
05:12 dalek parrot: [pdd28str] Renaming 'Parrot_string_split' to 'Parrot_str_split'.
05:12 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36315/
05:17 skv joined #parrot
05:35 dalek parrot: r36316 | allison++ | branches/pdd28str_part2:
05:35 dalek parrot: [pdd28str] Renaming 'string_split', also to 'Parrot_str_split', collapsing the
05:35 dalek parrot: two string splitting functions into one.
05:35 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36316/
06:01 Andy joined #parrot
06:40 slavorgn joined #parrot
06:42 dalek parrot: r36317 | allison++ | trunk/docs/pdds/pdd28_strings.pod:
06:43 dalek parrot: [pdd] Some additional deprecations and name changes for string functions.
06:43 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36317/
07:07 masak joined #parrot
07:12 uniejo joined #parrot
07:40 rurban_ joined #parrot
07:55 iblechbot joined #parrot
07:58 mberends joined #parrot
08:00 masak Rakudo doesn't build over here.
08:02 masak http://gist.github.com/57395
08:06 masak must be a recent change in Parrot that broke it.
08:09 jimmy joined #parrot
08:09 jimmy good afternoon, masak.
08:09 masak 你好, jimmy!
08:10 masak 好九不见了.
08:11 jimmy I came back home to celebrate the Spring Festival.
08:11 masak yes, I remember you told me.
08:12 masak happy 牛 year!
08:13 jimmy wow...you're following the fashion!!
08:13 masak I guess that joke is old by now.
08:13 jimmy happy 牛 year.
08:13 jimmy no...
08:14 jimmy It's still the Spring Festival.
08:15 masak meeting &
09:00 janus joined #parrot
09:00 janus hi
09:00 purl salut, janus.
09:01 janus how to file tickets? trac won't let me because of missing privs :/
09:01 janus is rt still acceptable?
09:01 TiMBuS have you verified after signing up?
09:01 TiMBuS it sends you an email
09:01 janus oh, sigh... didn't now i have to
09:01 janus okies, will wait then, thanks for the pointer :)
09:02 janus s/now/k$&/
09:02 TiMBuS np. also apparently trac has a bug, dont change your user settings
09:02 TiMBuS it might be fixed by now but just in case
09:03 janus good to know though
09:03 moritz trac has a bug? unbelievable... *SCNR*
09:03 janus hehe
09:09 janus fyi, trying to get parrot ready on openbsd/sparc64, hppa could be the next target
09:12 janus oh, and only one failing test on amd64 left. i386 and macppc reports will hopefully follow soon (i want to get 0.9.0 into the next release if possible)
09:12 szabgab anyone with November foo here? http://sial.org/pbot/34854
09:14 * moritz summons masak
09:14 szabgab oh and btw I still don't know where is the latest rakudo ?
09:15 moritz I think it's on github, but so far there hasn't been an announcement
09:15 moritz but since the migration nothing has changed significantly in rakudo
09:15 szabgab so does that mean none of you are currently doing any rakudo related work ?
09:16 moritz (except that parrot broke rakudo in a few ways, it seems)
09:16 moritz well, I added a few test files
09:16 moritz but it currently feels a bit lonely in rakudo dev
09:18 szabgab actually I was trying to use HTML::Template but that blew up in my face
09:19 szabgab so I was hoping to run the November test suit to see if the bug is in my code or not
09:22 szabgab perl6 t/html-template/01.t also fails (in November)
09:32 szabgab cannot build parrot at 36319
09:33 dalek parrot: r36320 | cotto++ | trunk/src/jit/i386/jit_defs.c:
09:33 dalek parrot: [jit] fix a neglected reference to Parrot_str_new_from_cstring
09:33 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36320/
09:34 cotto szabgab, that should fix it.  I just noticed that myself.
09:34 szabgab http://sial.org/pbot/34855
09:34 * szabgab trying
09:34 cotto or at least make the stupid bird less broken
09:35 * cotto is not happy with Parrot atm
09:41 alvar joined #parrot
09:43 moritz t/dynpmc/gdbmhash.t and t/dynpmc/rational.t fail here
09:54 szabgab after rebuilding parrot and rakudo I get:
09:54 szabgab parrot: symbol lookup error: /home/gabor/work/parrot/runtim​e/parrot/dynext/perl6_ops.so: undefined symbol: const_string
09:55 szabgab rebuilding the whole thing again...
10:00 cotto This does not appear to be the most thorough of function name changes.
10:06 dalek parrot: r36321 | cotto++ | trunk/languages/pipp/src/pmc:
10:06 dalek parrot: [pipp] update PhpArray after the string function renaming
10:06 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36321/
10:06 szabgab yeah, rebuilding with realclean did not help either, I get the above whe trying to "make" November
10:14 dalek parrot: r36322 | cotto++ | trunk/languages/pipp/src/pmc/phpfloat.pmc:
10:14 dalek parrot: [pipp] one last change for Pipp PMCs
10:14 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36322/
10:32 tomyan joined #parrot
10:51 masak szabgab: I get similar failures when trying to build rakudo.
10:51 masak pmichaud: ping
10:53 szabgab masak: rakudo compiled for me and I can even run   perl6 -e'say 24'
10:54 szabgab but not much more I think
10:55 masak szabgab: doesn't work here.
10:55 szabgab its the weather then
10:56 * masak growls
10:57 masak szabgab: what do you mean by 'but not much more I think'? things work, but only for 'say 24'?
10:58 szabgab simple scripts work
10:59 masak szabgab: let me rephrase. do you hafve any reason to think that things that used to work don't?
10:59 masak nope, still getting the error after make realclean/make cycle.
11:00 masak http://gist.github.com/57467
11:00 szabgab but this code:   BEGIN {@*INC.push('/home/');}
11:00 szabgab already blows up with
11:01 szabgab parrot: symbol lookup error: /home/gabor/work/parrot/runtim​e/parrot/dynext/perl6_ops.so: undefined symbol: const_string
11:01 masak same error as I have.
11:08 masak szabgab++ # [perl #62992]
11:20 szabgab masak: actually after I sent that I though I should just add it as an additional test to the pugs repo
11:20 szabgab the question, where to ? it is not in the Spec
11:21 szabgab should there be a Bugs directory where we can add tests that are triggered by bugs in one of the implementations?
11:57 skv joined #parrot
12:02 kid51 joined #parrot
12:07 kj joined #parrot
12:17 dalek parrot: r36323 | fperrad++ | trunk/src/dynpmc:
12:17 dalek parrot: [pdd28str] fix build (see merge in r36319)
12:17 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36323/
12:25 integral joined #parrot
12:29 Zaba_ joined #parrot
12:29 janus meh, still no luck on sparc64 (miniparrot dumps core)
12:30 rg1 joined #parrot
12:38 janus src/cpu_dep.c:90: defined(__sparc) -> defined(__sparc__) && !defined(__sparc64__) seems to fix... at least the build continues now
12:39 janus could that be something os-specific?
12:40 janus granted, i'm poking around in an area where others will have more clue
12:55 skv joined #parrot
12:56 masak szabgab: t/spec/integration or something similar might be the place. but I'm not the one to ask, moritz is.
13:12 szabgab moritz: see, you are being asked ^^^^
13:14 * masak grows impatient and bisects
13:15 nopaste "kid51" at 71.247.42.79 pasted "New test failures for 2 t/dynpmc/ tests at r36322" (494 lines) at http://nopaste.snit.ch/15488
13:15 braceta joined #parrot
13:16 kid51 Those failures come from something that happened in the last 9 hours.
13:17 skv joined #parrot
13:18 kid51 Somewhere between r36310 and 36322.
13:18 kid51 Have to go to $job now.
13:19 masak kid51: thank you.
13:19 kid51 Based on Smolder reports, probably between 36310 and 36320.  (I think there was a big merge there somewhere.)
13:20 masak I think it's reasonable to test implementations like Rakudo when doing merges.
13:25 iblechbot joined #parrot
13:41 NotFound dynpmc failures were fixed in r36323
13:42 NotFound <dalek> parrot: [pdd28str] fix build (see merge in r36319)
13:43 dalek joined #parrot
13:48 moritz szabgab: integration/ is fine by now, although I generally try to put the regression tests into existing files
14:07 szabgab if anyone is planning to attend GPW, please note I am trying to organize a Perl 6 hackathon afterwards
14:07 szabgab http://conferences.yapceurope.o​rg/gpw2009/wiki?node=Hackathon
14:13 Whiteknight joined #parrot
14:16 mberends joined #parrot
14:17 gryphon joined #parrot
14:21 Whiteknight have the packfile issues been rectified?
14:25 Coke pmichaud, s1n: I added you to the group parrotbugadmin; Based on what you said robrt said, that should make you a bugadmin of both parrot and perl6 queues.
14:30 * masak is beginning to suspect that the merge rev 36319 did the breaking
14:34 Infinoid janus: If __sparc isn't defined, is that something openbsd declares differently?  If we switch to __sparc__, will the check still work on solaris?
14:35 Infinoid Looking at the code in question, I think it intends to detect the platform, not the OS, so we should probably use something shared by both OSes.
14:37 alinbsp joined #parrot
14:38 skv joined #parrot
14:40 Infinoid janus: After googling a bit, I fear this stuff might not be very consistent.  I suppose the alternative is to just check for both __sparc and __sparc__ and !__sparc64__, and cross the sparc64/solaris case when we come to it.
14:41 janus Infinoid: where should the code in question be used? on 32bit sparc only?
14:43 janus from what i know, when __sparc64__ is defined also __sparc__ is
14:44 Infinoid Not being at all familiar with the sparc platform, I couldn't say for sure.  This check dates back to Dan Sugalski back in 2002, so we might just ask him.
14:44 Coke rakudo-hackers; is languages/perl6 now obsolete?
14:46 Coke also, can I get commit emails for rakudo on github, or am I reduce to an rss feed?
14:47 Infinoid Coke: When I asked that last week, the answer was yes, perl6 is obsolete (and they'll clean it up at some point).
14:47 Infinoid No idea about github commit emails... but the rss does show up in here.
14:48 Infinoid (for rakudo, and eclectus, and lazy-k, and etc etc)
14:48 Coke (rss) the diffs aren't in the raw feed, which makes it less than useful to me. ah well.
14:49 Infinoid True.  But links to the diffs are.
14:49 Infinoid All the bot needs is the list of changed files, and those are in the feed.
14:49 Coke in the feeds? I don't see them.
14:49 Coke oh, you mean in the bot output here.
14:49 Coke that isn't really useful to me. =-)
14:50 Coke I mean, on an ad hoc basis it is.
14:50 Infinoid Fair enough. :)
15:06 dalek parrot: r36324 | coke++ | trunk/DEPRECATED.pod:
15:06 dalek parrot: All entries need a tracking ticket.
15:06 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36324/
15:16 masak yup, r36319 is the culprit.
15:17 masak allison: ping
15:18 elmex joined #parrot
15:19 rg I can't help you, but you guys did read allisons mail to parrot-dev earlier today, right?
15:20 Coke http://lists.parrot.org/pipermail/p​arrot-dev/2009-February/001257.html
15:20 shorten Coke's url is at http://xrl.us/beer3q
15:20 Coke beer3q!
15:20 Coke shorten++
15:21 rg that one, exactly.
15:21 Coke I think the problem was that there were still usages /in core parrot/
15:22 Coke hopefully those are ok.
15:22 Coke (now)
15:26 * Coke re-discovers that writing is fun, but research can suck.
15:33 Andy joined #parrot
15:40 * rg can confirm, the build is fine. only some packfile related test errors remain.
15:40 uniejo_ joined #parrot
15:40 rurban_ joined #parrot
15:59 Infinoid janus: Is there a ticket open for this stuff?  If not, I will create one.
16:00 Theory joined #parrot
16:00 janus no, i'm still testing a diff
16:02 Infinoid Ok.  I want to ask the list about whether the assembly instructions in there are even valid (I suspect not, because instruction widths usually change when you go to 64 bits, but I don't know anything about sparc).  I will create a ticket summarizing the current situation, just so I have something to point to.
16:04 masak rg, Coke: thank you for the email reference.
16:04 masak I'll try to run the script on Rakudo.
16:10 Infinoid Hmm.  Doesn't openbsd use stack-smash protection by default?
16:10 davidfetter joined #parrot
16:11 Infinoid If so, the crash may have nothing to do with being on sparc64... it might just be due to trying to jump to code in a static array.
16:18 rurban Whiteknight: packfile not yet settled, but tests should be skipped, I think.
16:20 rurban As far as I know, simply the underlying constant lookup for numvals changed so the number pbs should be recreated also.
16:23 janus Infinoid: stack-smash protection is enabled by default. but as the build continued with my patch i think that could be it.  let's wait what the tests tell us :)
16:23 pmichaud hello, all.
16:23 Whiteknight rurban: thanks for the heads-up
16:23 moritz hello pmichaud. You have been missed.
16:24 pmichaud Yes, things have been... interesting around here.
16:24 pmichaud I should be back fully in an hour or so.
16:24 moritz great
16:24 Coke hopefully that's interesting in a good way.
16:24 pmichaud more interesting in a very distracting way.
16:24 Coke still one of the better alternatives. =-)
16:25 pmichaud Yes.
16:25 * particle imagines projectile vomit
16:25 pmichaud yes, there was some of that.
16:26 particle keytool--
16:26 moritz pmichaud: could you give masak commit access to rakudo? he has an un-breaking patch in his queu :)
16:26 pmichaud I thought I already did that.
16:26 Infinoid janus: Ok.  That means we might need a different way of doing things for the openbsd/sparc case.
16:26 Infinoid janus: http://groups.google.com/group/parrot-de​v/browse_thread/thread/ac06c3ba23da6eb1 , https://trac.parrot.org/parrot/ticket/271
16:26 shorten Infinoid's url is at http://xrl.us/beesat
16:27 masak pmichaud: checking.
16:27 masak pmichaud: nope.
16:27 janus Infinoid: thanks
16:27 pmichaud masak:  name...?
16:27 purl name is fine
16:28 masak at least not in such a way that I got it :)
16:28 masak pmichaud: masak
16:28 pmichaud masak: added.
16:28 masak pmichaud: confirmed. thank you.
16:28 masak it's nice to be back. :)
16:29 pmichaud #parrotsketch in 120
16:30 masak oh, right! today is tuesday!
16:30 rurban Infinoid: __sparcv9 is a bit too narrrow probably. that time sparcv9 was the latest, now we have v10
16:30 pmichaud I saw allison's post to the mailing list (strings merge) -- are we just suspending the deprecation policy between now and release?  Or did I miss a note somewhere...?
16:30 moritz I'll probably miss ps today, but I have nothing interesting to report anyway
16:31 Infinoid I haven't committed anything interesting, either.
16:32 Infinoid If I have time I'll make #ps, but I won't really have anything to say other than "I love writing rss->irc plugins, victim here!".
16:36 dalek rakudo: 8e67a4f | (Carl Masak)++ | src/ (3 files):
16:36 dalek rakudo: ran allison++'s script against the src/ files
16:36 dalek rakudo: Things now work again with Parrot r36319 and later.
16:36 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/8​e67a4f925c4d9ab6ff1842c8fa55b7e702f046f
16:36 shorten dalek's url is at http://xrl.us/beesb2
16:37 pmichaud masak++
16:38 Infinoid ask_: What are our chances of convincing shorten to ignore dalek?  Or alternately, ignoring github commit URLs?
16:41 moritz $ git-pull
16:41 moritz Updating 21f374f..8e67a4f
16:41 moritz src/ops/perl6.ops: needs update
16:41 moritz what can I do against that?
16:41 masak moritz: is yours clean?
16:42 moritz masak: I did a git-clean -f
16:42 moritz should it be clean after that?
16:42 kj pmichaud: what is the most common cause of "Failed to parse source" message? Usually, it's something like "syntax error in line x"
16:42 dalek parrot: r36325 | whiteknight++ | trunk/src/gc/system.c:
16:42 dalek parrot: [GC] update the documentation in this file to reflect it's new location
16:42 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36325/
16:42 masak moritz: no, not necessarily.
16:42 masak moritz: what does 'git status' say?
16:42 moritz ah, it says it's modified
16:42 moritz so how do I remove these modifications?
16:42 masak right.
16:42 masak git checkout
16:42 purl git checkout is just a command on top of git
16:43 masak purl: shut up.
16:43 purl make me
16:43 * masak makes purl shut up
16:43 pmichaud kj:  it means that whatever came back from attempting the grammar match wasn't a successful match.
16:43 masak moritz: git checkout src/
16:43 moritz masak: works, thanks
16:44 masak np
16:44 masak git is wonderful, but tricky in the beginning.
16:44 masak what one used to think of as 'revert' is simply 'checkout' :)
16:44 kj pmichaud: I've been trying to use 2 different parsers in 1 program, but somehow the second invocation always results in Failed to parse source message.
16:45 pmichaud kj:  are they different parser names?
16:45 pmichaud i.e., different grammars?
16:46 kj pmichaud: at first yes. later I tried to just get things working, so I figured that invoking the same parser twice would suffice
16:46 particle how do i add ignore files for git?
16:46 kj but that didn't really help
16:46 moritz particle: in the repo, or locally?
16:46 pmichaud http://github.com/guides/ignore-for-git
16:46 particle repo. i need to ignore generated files like *.dll
16:46 kj pmichaud: the first invocation is to parse a file, the second to parse input from the commandline
16:46 particle git status lists them currently
16:46 pmichaud particle:  http://github.com/guides/ignore-for-git
16:47 moritz particle: echo '*.dll' >> .gitignore
16:49 dalek rakudo: 0c50158 | (Carl Masak)++ | .gitignore:
16:49 dalek rakudo: ignore .dll files
16:49 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/0​c50158cba82aef182e0a8d22a3b9a91b1fc941b
16:49 shorten dalek's url is at http://xrl.us/beesfb
16:49 masak oops, committed. :)
16:51 particle oh, come on!
16:51 purl EVERYBODY TO THE LIMIT!
16:51 masak sorry :/
16:51 particle now that i have a conflict, how do i merge?
16:51 masak the change should just go in, I think.
16:51 Whiteknight particle: you don't. you're trapped in repository limbo forever
16:51 particle no, i did git commit .gitignore
16:52 masak PerlJam: yes, but it's the same change.
16:52 particle then git told me somebody else modified the file
16:52 masak oops, PerlJam, sorry.
16:52 particle it's not the same change
16:52 moritz git reset --hard HEAD^ # something along these lines? just guessing...
16:52 particle i added many different extensions
16:52 masak particle: when you do a git pull, you get a conflict, right?
16:52 particle CONFLICT (content): Merge conflict in .gitignore
16:52 particle Automatic merge failed; fix conflicts and then commit the result.
16:52 masak particle: right.
16:52 masak edit the file.
16:53 masak 'git add' it
16:53 masak and commit.
16:53 pmichaud ...and push?
16:54 masak aye.
16:54 particle c:\Users\particle\dev\parrot​\rakudo\languages\rakudo>git pull
16:54 particle error: Entry '.gitignore' would be overwritten by merge. Cannot merge.
16:54 particle fatal: merging of trees 50360ac47e81d838ae10b221aa5e66f3811d8ed9 and 51e45f1c9fb4ab0687dfc6181d3b8db2e6cd86f2 failed
16:54 particle that's after editing and adding
16:54 masak particle: but did you commit after adding?
16:54 particle ...and trying to commit
16:55 particle fatal: cannot do a partial commit during a merge.
16:55 masak trying to commit?
16:55 masak ah.
16:55 masak 'git status'
16:55 particle ok, nm
16:55 masak what's left out?
16:55 particle i tried git commit .gitignore
16:55 masak aha.
16:55 particle i guess i need git commit
16:55 masak just git commit should do it.
16:55 masak takes a bit of getting used to.
16:56 kj pmichaud: I think I got it now (double invocation of same parser). Will look into this tonight
16:56 hercynium joined #parrot
16:56 pmichaud kj:  okay.  Until I get everything worked out with rakudo+parrot repos I'm a bit distracted as far as looking at anything else.
16:56 particle the fact that git st doesn't work will make it even harder to retrain my muscles
16:56 * Infinoid uses "stg st"
16:57 kj pmichaud: I understand, np.
16:57 masak particle: add an alias in .gitconfig
16:57 masak I did.
16:57 particle so, how do i get these changes to github?
16:58 masak 'git push origin master'
16:58 masak particle: http://gist.github.com/57619
16:59 particle XML error: syntax error
16:59 particle error: Error: no DAV locking support on http://git.rakudoperl.org/rakudo/rakudo.git/
16:59 particle error: failed to push some refs to 'http://git.rakudoperl.org/rakudo/rakudo.git'
16:59 moritz particle: ah, you're using http://, not git:// - dunno if that works on github
16:59 masak it doesn't.
17:00 particle does git:// forward from rakudoperl.org?
17:00 particle or do i need to config for git://github...
17:00 pmichaud github doesn't handle cnames well.
17:00 pmichaud and I think push might require the git@github.com address.
17:01 masak aye.
17:02 particle sigh.
17:05 Infinoid Whiteknight: http://www.sics.se/~psm/sparcstack.html is an interesting read.  I suspect IA64 has something similar; that architecture is full of weird s**t.
17:08 dalek parrot: r36326 | allison++ | branches/pdd28str_part2:
17:08 dalek parrot: Removing second strings development branch from the repository
17:08 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36326/
17:09 jan joined #parrot
17:14 pmichaud I've now created rakudo-commits@groups.google.com for those who want email notifications of commits to github.
17:17 Coke pmichaud: those commit messages don't include the diffs, correct?
17:18 pmichaud Coke: correct.
17:18 pmichaud They include a link to the diffs, though.
17:18 * Coke unsubs
17:18 pmichaud right, I don't know that it'll be that useful w/o the diffs, but it at least gives me a checklist
17:18 pmichaud for easily reviewing the diffs.
17:22 cotto It's a pre-posting party over in #ps
17:23 pmichaud Coke:  any suggestions for creating per-language Makefiles when Parrot's 'reconfigure.pl' isn't available?
17:23 Whiteknight Infinoid, thanks! I'm always looking for documentation about weird s**t platforms
17:24 cotto pmichaud, ping
17:24 Infinoid Whiteknight: The IA64 case is explained neatly in RT #19418
17:24 geof joined #parrot
17:25 pmichaud cotto: pong
17:25 Infinoid (Took some digging to get that, heh.)
17:25 cotto Do you have any recommendations on TT #256 (bad PAST->PIR generation)?
17:26 cotto I've tried to fix it (since I caused it), but I'm stuck.  I posted what I know in the tt.
17:27 pmichaud cotto: this sounds suspiciously like another a bug I ran into.
17:27 pmichaud just a sec
17:28 pmichaud (searching)
17:30 pmichaud cotto: it works fine before r36165 ?
17:31 cotto pmichaud, yes
17:32 dalek allison@perl.org | Debian/Ubuntu chroot Environment Setup:
17:32 dalek link: http://www.perlfoundation.org/parrot/index.​cgi?debian_ubuntu_chroot_environment_setup
17:32 shorten dalek's url is at http://xrl.us/beesjm
17:40 pmichaud cotto: I don't know what might be the cause, but I'm very suspicious of
17:40 pmichaud SET_ATTR_resize_threshold(INTERP, SELF, 8);
17:41 pmichaud since we seem to be having trouble with every 8th value.
17:41 cotto I'll look into that.
17:48 cotto That looks less suspicious.  ISTR that PCC uses FixedIntegerArrays in that area.  Anyway, changing the magic 8 to 64 didn't change the bug.
17:49 pmichaud hmmmm
17:49 pmichaud ResizablePMCArray seems to be confusing 'elements' with Size
17:51 pmichaud maybe not.  checking.
17:51 pmichaud did r36165 get reverted?
17:54 ask- joined #parrot
17:55 rg svn browsing says no.
17:55 pmichaud then why does my trunk not match the svn diff?
17:55 pmichaud ohhhh, nm.
17:55 pmichaud I'm looking at the wrong source.
17:56 pmichaud sorry about that.
17:56 rg np
17:56 barney joined #parrot
17:56 cotto I didn't revert it because it was a relatively obscure bug that make test didn't appear to catch.
17:57 pmichaud np, makes sense.
18:00 tomyan left #parrot
18:00 dalek parrot: r36327 | rurban++ | trunk/config/gen/makefiles/parrot_embed_pl.in:
18:00 dalek parrot: One more for TT #262, $linkflags missing
18:00 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36327/
18:02 Coke pmichaud: I haven't jumped off that bridge yet, and am always building inside of a parrot build.
18:02 pmichaud Coke: so, no ideas yet?  ;-)
18:02 Coke nope.
18:03 pmichaud okay.
18:03 Coke someone needs to work on getting an installed parrot usable for building languages.
18:03 Coke this is a small component of that.
18:04 pmichaud well, I don't know that in the general case we can/should expect Parrot to be building individual languages' Makefiles.
18:04 rurban The only question is: where do we put our perl5 libs?
18:04 pmichaud that's not the _only_ question.  :-)
18:05 rurban I'd like a seperate CPAN release for that
18:05 Coke rurban; if they're required to build modules, I would say create Bundle::ParrotDeveloper or something.
18:05 rurban The rest is quite simple :)
18:05 pmichaud rurban:  I don't know that in the general case we should expect Parrot to be building individual languages' Makefiles.
18:06 rurban But as seperate CPAN release we need some automated package creation makefile target
18:06 Coke I think eventually the goal is to have everything be available as a PBC instead, so we can build languages without forcing languages developers to use our build chain.
18:06 pmichaud Coke:  well, I still have to be able to detect things like  .o vs .obj, .dll versus .so, etc.
18:07 rurban We need the installed parrot, some perl libs, and make
18:07 Coke yes, but we could provide that with parrot only.
18:07 Coke (that is, replace the build bits in perl with build bits in parrot.)
18:07 rurban Coke, one pbc is usually not enough. We have docs, man pages, and some langs need more pbc's.
18:07 Coke and ship parrot and parrot-dev rpms.
18:08 Coke rurban: my intent is not "one pbc"
18:08 rurban build bits in perl => parrot: parrot-2.0 earliest
18:08 Coke "I think eventually the goal" <---
18:08 szabgab has anyone looked at the broken rakudo thing:   parrot: symbol lookup error: /home/gabor/work/parrot/runtim​e/parrot/dynext/perl6_ops.so: undefined symbol: const_string
18:08 rurban make realclean
18:08 purl rumour has it make realclean is sometimes need in parrot to work against missing dependencies
18:08 pmichaud szabgab: you probably need make realclean
18:09 pmichaud allison checked in a major change to Parrot
18:09 pmichaud you probably also need the latest version of Rakudo
18:09 szabgab so where is Rakudo now ?
18:09 pmichaud szabgab: in transition, at the moment, but github.com/rakudo/rakudo
18:10 szabgab its not languages/perl6 right ?
18:10 szabgab so how do I kick start it ? git clone http://github.com/rakudo/rakudo ?
18:11 pmichaud szabgab: yes, into languages/rakudo of the parrot tree.
18:11 pmichaud I'm working on improving that today.
18:11 Tene joined #parrot
18:11 szabgab do I have to put it there or can I put the source in another directory ?
18:12 szabgab next to my parrot checkout dir ?
18:12 rurban Coke: a install.pbc?
18:12 pmichaud right now it only knows how to build from languages/rakudo/
18:12 pmichaud I'm working on that.
18:12 dalek parrot: r36328 | fperrad++ | trunk/t:
18:12 dalek parrot: [t] glob arguments if any
18:12 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36328/
18:12 shragei left #parrot
18:12 rurban Coke: with Parrot::Compiler, configure helpers and such
18:13 szabgab hmm I get warning: remote HEAD refers to nonexistent ref, unable to checkout.
18:14 dalek parrot: r36329 | fperrad++ | trunk/languages/lua/config/makefiles/root.in:
18:16 pmichaud oh, looks like I might be able to get the information I need from parrot_config
18:16 dalek parrot: [Lua] add a target 'codetest'
18:16 szabgab the correct command is     git clone http://github.com/rakudo/rakudo.git    (with the .git extension)
18:16 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36329/
18:17 pmichaud cotto: I don't have any clues on #256 at the moment.  The 'emit' method of CodeString isn't doing anything special here, so something is causing the ResizableStringArray to get the wrong arguments.
18:18 pmichaud my suspicion is that somewhere the 'size' value is getting confused in RSA
18:18 dalek parrot: r36330 | fperrad++ | trunk/languages/lua:
18:18 pmichaud or that there's an off-by-one error when growing the array
18:19 cotto pmichaud, ok.  I'll keep digging, poking people or putting it off.
18:19 szabgab after checking it out how do I build it ?    the usual make languages in the parrot directory  or do I need something else ?
18:19 pmichaud assuming that you have an up-to-date parrot, you should be able to do
18:19 pmichaud cd languages/rakudo
18:19 allison joined #parrot
18:19 pmichaud perl Configure.pl;
18:19 pmichaud make
18:20 * szabgab still cloning
18:21 dalek parrot: [Lua] fix line length
18:21 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36330/
18:22 szabgab cloning failed, lovely
18:22 szabgab error: Unable to find 185a5e35186aeb569a9f11019f141b7517f68d1d under http://github.com/rakudo/rakudo.git
18:24 masak joined #parrot
18:26 * szabgab rm -rf rakudo; clone again worked
18:27 szabgab pmichaud++ rakudo workz again! yippee
18:29 pmichaud #ps in 1
18:30 masak pmichaud: is the perl Configure.pl step mentioned in the README, or anywhere else?
18:30 pmichaud masak: not yet, but all of that will likely change very soon.
18:30 masak oki, good.
18:31 pmichaud I'm trying to decide how someone gets from rakudo to parrot
18:32 pmichaud i.e., do we (1) include a copy of parrot in the rakudo checkout, (2) include a script/target that checks out the (known working) version of parrot from svn.parrot.org, (3) tell people to obtain/build parrot before building rakudo, or ...?
18:32 masak good question.
18:32 purl Yeah, it is. I'm stumped.
18:33 barney Keep it simple and stick with 3)
18:33 dalek parrot: r36331 | fperrad++ | trunk/src/string/api.c:
18:33 chromatic joined #parrot
18:33 dalek parrot: [codingstd] fix c_parens
18:33 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36331/
18:34 pmichaud barney: that's not quite so simple -- we have to tell them *which* revision to check out.
18:34 pmichaud or perhaps we keep a 'rakudo' tag in the svn.parrot.org repository.
18:34 pmichaud so then it's just  https://svn.parrot.org/parrot/tags/rakudo
18:34 pmichaud that might work.
18:34 barney For Pipp it's simple. Use the same version as for Rakudo
18:34 cotto barney++
18:35 pmichaud probably should call it 'rakparrot' or something
18:35 cognominal joined #parrot
18:36 barney or working_parrot
18:36 pmichaud well, I'd like it to be clear that it's the version known to be working with rakudo.
18:38 pmichaud if other languages want to piggyback on the tag, that's great, but I won't be able to guarantee it for other langs :-)
18:38 dalek parrot: r36332 | whiteknight++ | trunk/src/gc/system.c:
18:38 allison pmichaud: is it a substantial advantage to target something other than the monthly release?
18:38 dalek parrot: [GC] adding some notes about related Trac tickets where things in the system.c file need to be cleaned up.
18:38 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36332/
18:38 Infinoid I'd aim for the releases and use a script to pull them from CPAN.
18:38 pmichaud allison: for Rakudo, yes.
18:38 pmichaud Both Jonathan and I frequently have to make changes to Parrot for MMD-type stuff and PCT-type stuff.
18:39 pmichaud waiting 4 weeks for people to use them in Rakudo isn't going to work.
18:41 pmichaud at least, I don't think it will work for the next 3-4 months.
18:41 pmichaud By the time we get to the summer, perhaps.
18:42 Whiteknight what is Druid?
18:42 cotto Druid?
18:42 Infinoid purl, druid?
18:42 purl well, druid is a totally nerfed class
18:42 masak https://github.com/masak/druid/
18:42 masak purl: Druid is https://github.com/masak/druid/
18:42 purl ...but druid is a totally nerfed class...
18:43 masak purl: no, Druid is https://github.com/masak/druid/
18:43 purl okay, masak.
18:43 allison pmichaud: I ask, because we planned the whole release schedule to give language developers a sane point to synchronize with. seems a bit like reinventing the wheel if we need another layer of sane points to synchronize with.
18:43 Coke allison: if monthly releases correspond to sane points, that would work.
18:44 pmichaud allison: I'd agree if Parrot was able to keep up with Rakudo (more)
18:44 pmichaud allison: but even at PDS I noted that Rakudo probably would not be able to use Parrot's monthly release schedule.
18:44 chromatic Given how Rakudo pushes PCT, I'm not sure that's possible.
18:44 chromatic allison, your SVN merge program uses the wrong branch name in commit messages.
18:45 pmichaud As long as Rakudo is "out in front" on PCT and various other Parrot features, I don't think we can afford to have week-or-longer delays in development waiting for a Parrot release to catch up.
18:46 allison chromatic: "my SVN merge program" would be *me*, though I probably should automate it
18:46 pmichaud I'm not asking Parrot to provide additional sane points -- I'm simply saying that Rakudo will keep its own notion of sane points as a superset of Parrot's list.
18:46 chromatic You merged the MMD branch back to trunk twice, according to the commit messages :)
18:46 Coke I still think that languages should have a version that targets the major releases, even if it means features are missing from that release.
18:46 pmichaud Coke: yes, we'll have regular releases for Rakudo.
18:46 pmichaud Coke:  and those will be based on the Parrot monthly releases.
18:46 allison Coke: though, the actual tag on the commit message is correct, says 'pdd28str'.
18:46 * Coke dodges.
18:47 allison pmichaud: fair enough, I just wish we had a way to make it simpler. Something to think about.
18:47 pmichaud most people playing with Perl 6 still will tend to want to use the head version of Rakudo, which means that we'll need something more up-to-date than Parrot's latest monthly release.
18:48 allison Coke: actually, at this point I think the commit messages say I've merged in pdd27mmd about 5 times (I keep making the same cut-n-paste mistake)
18:48 Coke allison: you mean "chromatic"
18:48 pmichaud (and people developing Perl 6 definitely need the head version of Rakudo, but I don't necessarily want them to have to have Parrot trunk.)
18:48 pmichaud (because Parrot trunk might not work with Rakudo head)
18:49 Coke using a tag on parrot is probably the simplest way to do it. I had just been saying "use revision foo", but that's clumsy.
18:49 allison Coke: yes, paying more attention to #parrotsketch than #parrot
18:49 pmichaud right, I think I'm going for the tag approach, yes.
18:49 ask- one thing git has that's really nice is that the corresponding feature to "svn externals" makes you tie the external not just to another repository, but to a particular commit in the other repository.
18:50 Coke I do that at work: have a tag called 'stage' that I regularly delete and re-tag.
18:51 Infinoid Whiteknight: Please note (if you didn't already notice), TT #271 is related to the #273 you just created.  (It raises some questions for one specific platform affected by those cases.)
18:52 ask- coke: that doesn't work - when you go back in history, your "stage tag" doesn't go back in history.
18:52 Whiteknight Infinoid: Yeah, I need to link the two or something
18:52 Whiteknight #271 reports a breakage, #273 is a call to cleanup the whole damn file
18:52 ask- (doesn't work = at best a clumsy workaround).
18:52 Coke ask: I don't need history. I need "what is the current build associated with stage."
18:53 Coke (and it does go back, in that I can say "give me what was in stage as of last wednesday". (Or at least I think it does.)
18:53 Coke this was as opposed to having a "tag" that I had to merge in changes from. it's always just a snapshot of something "good"
18:54 Coke *change to
18:54 Coke (clumsy workaround) What would you recommend?
18:55 pmichaud I've used 'latest' tags in various svn repos to indicate "latest stable" and it works just fine.
19:10 ask- coke: with svn I was doing it the same way (with one branch I'd follow from the parent repository).    It's just awful though.  It breaks when you want to go back 10 commits. Or when you branch a release branch.  Or when you want to use an old tag.  etc etc.   At first the thing with having to do an explicit commit to update the submodule ("external") in a git repository was a hassle, but it solves all those issues
19:13 * ask- has to run to the office.
19:16 Coke (with svn) yes, but we're stuck with svn.
19:16 Coke (at the very least, I am stuck with it at my work.)
19:27 Whiteknight allison: did you see that email I sent out about "self"?
19:28 allison Whiteknight: just saw it this morning, haven't had a chance to reply yet
19:28 rurban pmichaud: a typical package layout is here http://code.google.com/p/cygwin-rurban/source/br​owse/trunk/release/parrot/CYGWIN-PATCHES/README
19:28 Whiteknight okay, I'm relieved just to know somebody received it
19:28 shorten rurban's url is at http://xrl.us/beesz3
19:29 rurban debian and redhat is similar
19:30 pmichaud rurban: at the moment I'm more interested in what happens when people type 'make install'
19:30 dalek tracwiki: v108 | coke++ | ParrotRoadmap
19:30 pmichaud for example, if someone is trying to build the latest version of rakudo.
19:30 dalek tracwiki: https://trac.parrot.org/parrot/​wiki/ParrotRoadmap?version=108
19:30 shorten dalek's url is at http://xrl.us/bees2d
19:30 rurban right now some /usr/runtime files also get installed.
19:31 rurban but a basic make install is everything together which was build with make.
19:32 pmichaud where do the .pbc files end up?
19:32 rurban /usr/lib/parrot/library/
19:32 rurban If they get built at all. for some there are only .pir's
19:32 pmichaud that didn't happen when I did 'make install' this morning.
19:33 pmichaud why aren't they being built, then?
19:33 rurban missing make targets
19:33 pmichaud I don't understand.
19:33 rurban such as   /usr/lib/parrot/library/PGE/Perl6Grammar.pir e.g. :)
19:33 pmichaud they're built when I do a normal 'make'
19:34 rurban Ah, I see. So they are just missing in the MANIFEST
19:34 allison pmichaud: but not installed
19:34 pmichaud allison: why not installed?
19:34 pmichaud shouldn't we install precompiled versions of the libraries?
19:34 pmichaud or tools, or whatever we choose to call them?
19:34 rurban So we can add a simple logic to the installer script to always copy .pbc's on library .pir's
19:35 confound_ joined #parrot
19:36 rurban But I see that with my MANIFEST.generated patch all those pbc's are in
19:36 rurban here it is: runtime/parrot/library/PGE/P6Grammar.pbc          [main]
19:37 pmichaud That's P6Grammar.pbc
19:37 pmichaud that one doesn't exist.
19:37 pmichaud It's Perl6Grammar.pbc
19:37 rurban oops
19:38 rurban no it's missing
19:38 rurban That deserves a new ticket
19:39 pmichaud I issued a ticket earlier for it.
19:39 pmichaud TT #272
19:40 rurban I want to extend it to all generated pbc's.
19:40 pmichaud feel free.
19:40 pmichaud I just know I ran into that issue when trying to build Rakudo.  :-)
19:41 rurban Your fix is just a single line to MANIFEST.generated
19:43 dalek parrot: r36333 | rurban++ | trunk/MANIFEST.generated:
19:43 dalek parrot: Fix TT #272, Perl6Grammar.pbc not in installed parrot
19:43 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36333/
19:45 chromatic What did const_string become in the great STRING API renaming?
19:46 pmichaud Parrot_str_new, I think.
19:47 chromatic Looks like Parrot_str_new_constant.
19:47 pmichaud Parrot_str_new_constant.
19:47 chromatic I couldn't find the commit for it for some reason.
19:47 andrew_ joined #parrot
19:48 allison chromatic: const_string  Parrot_str_new_constant, yes
19:48 allison chromatic: they're listed in the pdd, look for "(was const_string)"
19:49 chromatic Ah, I had to touch the dynpmc *.pmc files.
19:57 Coke can we not list the old names for things in the PDD?
19:58 Coke It's helpful for a transition, but long term, it's superfluous
19:58 Coke (and for a transition, a wiki page is probably better.)
19:59 rurban For older languages it would be very helpful
19:59 allison Coke: I needed them before the transition because I was deciding what to name them (that is, it was part of the design)
19:59 rurban (kea-cl e.g.)
19:59 allison Coke: but, we can remove them to a wiki page (or just remove them entirely) after
20:00 Coke TT #221; did I miss a commit?
20:00 Coke I didn't see, e.g., "socket" get removed.
20:01 Coke hurm.
20:02 Coke Ok, rephrase. I don't see an entry in PBC_COMPAT that corresponds to removing the socket-related opcodes, or a list on the ticket that says what was removed.
20:02 dalek parrot: r36334 | allison++ | trunk/DEPRECATED.pod:
20:02 dalek parrot: [cage] Deprecation and removal of socket opcodes completed, removing entry from
20:02 dalek parrot: DEPRECATED.pod.
20:02 purl deprecated.pod is pretty big, looking now :(
20:02 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36334/
20:02 ask_ joined #parrot
20:02 * turbov21 hmms: I like --test in the Configure.pl file, what a cool way to see if you're system can make Parrot.
20:03 Coke purl, no, DEPRECATED.pod is <reply>
20:03 purl okay, Coke.
20:03 allison Coke: they're not in src/ops/io.ops anymore, whatever the history was
20:03 ask- joined #parrot
20:04 rurban turbov21: you'll also need --verbose if some build step fails
20:04 Whiteknight DEPRECATED.pod?
20:04 Whiteknight purl DEPRECATED.pod?
20:04 purl bugger all, i dunno, whiteknight
20:05 Coke Whiteknight: that's a way to tell purl to shut up about something.
20:05 turbov21 rurban: thanks.  if this build fails again (after Configure.pl seemed to pass), i'll be sure to try that.
20:05 Coke and not listen to people when they say it again.
20:05 masak Coke: really?
20:05 purl no, not really
20:05 Whiteknight well tell purl to shut up again, for me
20:05 Coke DEPRECATED.pod is interesting.
20:05 Coke DEPRECATED.pod?
20:06 rurban perl Configure.pl --debugging=1 --optimize=-g3 gives me -DNDEBUG
20:07 turbov21 What does "[PGE.pbc] Segmentation fault" mean?
20:07 Whiteknight turbov21: it means Parrot segfaulted while trying to build PGE.pbc
20:07 particle it's one bit of undigested crayon in an otherwise perfect turd :(
20:08 Whiteknight which in turn means that there's some bug somewhere doing something wrong, some how
20:08 turbov21 Whiteknight, that narrows it down.
20:08 turbov21 At least I know it's not the cat.
20:08 Whiteknight turbov21: Sorry, but it's a very unhelpful (but all-too-common) error message. What revision are you building?
20:09 Whiteknight and what platform?
20:09 rurban try different runcore's: ./parrot -f compilers/pge/PGE.pir    or -t
20:09 turbov21 The latest trunk build (i just ran svn a few minutes ago) on Ubuntu, 32-bit AMD
20:10 turbov21 (apologies for being vague, i was hoping that might have been a more specific error.)
20:11 dalek parrot: r36335 | allison++ | trunk/docs/gettingstarted.pod:
20:12 dalek parrot: [cage] Updating getting started guide for currently supported ways of getting
20:12 dalek parrot: Parrot, and new URLs.
20:12 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36335/
20:12 Whiteknight turbov21: run a "make realclean && perl Configure.pl && make"
20:12 particle turbov21: try ./parrot examples/pir/mandel.pir # to see if parrot built
20:13 Whiteknight particle: parrot would have built if it was trying to compile PGE.pbc
20:13 particle well, i should say "to see where the segfault might be"
20:14 Whiteknight ah, okay
20:14 turbov21 I'm waiting for the latest "make realclean && perl Configure.PL --without-icu --test --verbose && make" cycle to finish
20:14 particle ok
20:15 turbov21 "./parrot -f compilers/pge/PGE.pir" doesn't do anything, it's like a print "\n"
20:16 turbov21 "./parrot -t compilers/pge/PGE.pir" dumps what looks like PIL code to my screen
20:17 rurban good. -f is for the fastcore which you don#t have. -t is for tracing to see where it crashes.
20:17 turbov21 "./parrot examples/pir/mandel.pir #" shows an ascii fractal
20:18 rurban without --verbose and --test it's much faster
20:18 turbov21 is there anyway to dump -t into a file? there's too much to scroll back to the top
20:19 Coke turbov21: use shell redirect.
20:19 turbov21 like: > file
20:19 Coke ./parrot -t foo.pir 2> trace.out
20:19 Coke or > if it's stdout. I think it's stderr, though.
20:19 Coke PGE.pbc segfault has a ticket (or two!) already.
20:19 Whiteknight ./parrot -t foo.pir > trace.out 2&>1
20:20 Coke I even did an svn-bisect to find the offending commit.
20:20 rurban and the winner is....
20:21 Whiteknight ...I hope he doesn't say "whiteknight did it!"
20:21 turbov21 According to changelog, this is 0.9.0
20:21 chromatic That's it, you're on segfault detail.
20:22 Coke it was fperrad with an innocuous change.
20:22 rurban which rev?
20:22 purl which rev are you at?
20:22 Coke https://trac.parrot.org/parrot/ticket/261
20:22 Coke "Looks like the segfault started happening in r36176."
20:23 Coke https://trac.parrot.org/parrot/changeset/36176
20:23 turbov21 how would i check the rev?
20:24 Coke "svn info ."
20:24 turbov21 Last Changed Rev: 36332
20:24 * turbov21 notes that svn command
20:24 rurban r36176 looks completely innocent
20:25 Coke sure does. but that's where the bisect led me.
20:25 Coke I even double checked after one more time, and 36175 works, and 36176 doesn't.
20:25 rurban You also have a file .parrot_current_rev
20:25 Coke I wouldn't rely on that.
20:25 rurban maybe it falsely detected an exec runcore?
20:26 Coke since you have to run Configure.pl to update it.
20:26 rurban can you compare to resulting config_lib.pasm's if there's a major difference
20:28 turbov21 Here's what the trace.out looks like: http://kerovia.net/trace.out.txt
20:29 rurban looks good
20:29 purl O_O
20:29 Coke I don't think that's from your segfault, is it?
20:29 Coke (note that -t1 can impact the running of a given program to change the result.)
20:30 rurban grep execcapable config_lib.pasm => 0 or 1
20:30 rurban grep jitcapable config_lib.pasm # is probably 1
20:33 rurban amd 32-bit right?
20:35 Coke rurban: good call: https://trac.parrot.org/pa​rrot/ticket/261#comment:3
20:35 Coke it's whatever feather is. =-)
20:35 dalek parrot: r36336 | allison++ | trunk:
20:35 dalek parrot: [strings] Renaming 'Parrot_str_equal' to 'Parrot_str_not_equal' so the name
20:36 dalek parrot: matches the actual logic of the function.
20:36 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36336/
20:36 Coke rurban: has_exec_protect vanished.
20:36 rurban :)
20:37 Coke ah.
20:37 rurban But the segfault in comment 3 runs a slow_core. There's a false NCI call
20:37 Coke because he updated auto/jit.pm to look for files named _c but didn't update the existing files to match.
20:38 Coke oh, wait, I'm looking in the previous version here.
20:38 Coke (the diff is still good.)
20:38 rurban Looks like /usr/lib/libicudata.so.34 is the culprit
20:39 turbov21 but i'm cinfiguring with --without-icu
20:42 Coke it is possible our two segfaults are not related.
20:42 turbov21 ah.
20:42 rurban turbov21: your failure is probably just the jit, or exec core, which is falsely enabled
20:42 rurban the failure in  https://trac.parrot.org/pa​rrot/ticket/261#comment:3 is interesting because there it runs the slow_core
20:43 rurban And I also predict that's AMD specific only...
20:43 Coke how can I tell if I'm on an amd box?
20:46 rurban uname -a
20:46 purl Infobot 0.43.3 alpha (oznoid+#perl)
20:46 Coke Linux feather 2.6.18-6-xen-686 #1 SMP Sun Feb 10 22:43:13 UTC 2008 i686 GNU/Linux
20:46 rurban xen
20:46 rurban cat /proc/cpuinfo
20:46 elmex joined #parrot
20:47 Coke vendor_id       : GenuineIntel
20:47 rurban model name is enough normally
20:47 Coke model name      : Intel(R) Core(TM)2 Duo CPU     E6550  @ 2.33GHz
20:47 rurban cat /proc/cpuinfo| grep model
20:47 rurban Same as mine :)
20:47 Coke so, there goes that theory.
20:47 dalek tracwiki: v5 | allison++ | ChrootSetup
20:47 rurban shit
20:47 dalek tracwiki: https://trac.parrot.org/parro​t/wiki/ChrootSetup?version=5
20:48 alinbsp joined #parrot
20:48 rurban On cygwin it does not crash, and on my various linux vmware images also not.
20:48 turbov21 rurban, is there a way to properly enable exec core?
20:48 rurban solaris-64int also not
20:48 rurban try and preach
20:48 Coke allison: if you wrap code blocks in {{{ }}} they will render as if in <pre>
20:49 allison Coke: ah, thanks
20:49 rurban we believe that the changed tests changed the exec or jit capability detection
20:49 rurban jit is usually harmless, exec is quite hard to get right.
20:50 Coke rurban: my problem seems to be that exec is /no longer/ detected.
20:50 turbov21 I'm assuming since Parrot is a VM, you can't turn off jit
20:51 Coke (hurm. probably more like <code>, but you get the idea.
20:51 dalek tracwiki: v6 | allison++ | ChrootSetup
20:51 rurban sure you can turn it off, see ./parrot -hg
20:51 rurban sure you can turn it off, see ./parrot -h
20:51 dalek tracwiki: https://trac.parrot.org/parro​t/wiki/ChrootSetup?version=6
20:52 rurban you have the following runcore options: -S -j -f -C -b -g
20:54 Coke (not all of which may work on your particular built parrot)
20:58 dalek tracwiki: v7 | allison++ | ChrootSetup
20:58 dalek tracwiki: https://trac.parrot.org/parro​t/wiki/ChrootSetup?version=7
21:00 bacek joined #parrot
21:03 rob joined #parrot
21:05 dalek parrot: r36337 | rurban++ | trunk/t/native_pbc:
21:05 dalek parrot: TT #254: Rebuild 32-bit and 64-bit intel numbers, as the previous ones
21:05 dalek parrot:          get the index access wrong.
21:05 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36337/
21:09 moritz src/jit/i386/core.jit: In function ‘Parrot_eq_s_s_ic_exec’:
21:09 moritz src/jit/i386/core.jit:1149: error: ‘Parrot_str_equal’ undeclared (first use in this function)
21:09 moritz ENOBUILD
21:14 rg my guess would be allison overlooked something in r36336
21:14 dalek tracwiki: v8 | allison++ | ChrootSetup
21:15 dalek tracwiki: https://trac.parrot.org/parro​t/wiki/ChrootSetup?version=8
21:17 rg seems like just that one occurrence in that file.
21:19 rob what the best way to handle constants in PAST, I was thinking of creating var nodes with a const flag in the symbol table.. it seems kinda ugly though
21:19 particle PAST::Val
21:20 particle many constants are automatically made into PAST::Val nodes anyway, for example an int constant will become a PAST::Val node
21:21 rob but how do I associate a past val with an identifier?
21:22 dalek parrot: r36338 | allison++ | trunk:
21:22 dalek parrot: [strings] Make 'Parrot_str_equal' the primary, and have 'Parrot_str_not_equal' call it.
21:22 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36338/
21:23 rg allison: can you check moritz' report and (probably) rename the occurence in src/jit/i386/core.jit before we get really confusing results?
21:25 jonathan pmichaud: ping
21:27 allison rg: okay
21:27 particle rob: you want to put the Val inside a Var
21:28 jonathan or maybe if they are true constants and just simple values, maintain a compile-time table of them.
21:28 particle generally via an Op
21:28 rob ok
21:34 dalek tracwiki: v9 | allison++ | ChrootSetup
21:34 dalek tracwiki: https://trac.parrot.org/parro​t/wiki/ChrootSetup?version=9
21:42 dalek tracwiki: v46 | rg++ | Languages
21:42 dalek tracwiki: https://trac.parrot.org/parr​ot/wiki/Languages?version=46
21:43 Coke_afk ugh.
21:43 Coke ww
21:44 dalek parrot: r36339 | allison++ | trunk/src/pmc/task.pmc:
21:44 dalek parrot: [strings] Use the new sane-logic 'Parrot_str_equal' in the Task PMC (as an example).
21:44 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36339/
21:47 dalek tracwiki: v10 | allison++ | ChrootSetup
21:47 Util Trying to commit fix for TT#256; first commit since svn.parrot.org repo change.
21:47 Util Commit fails with (403). Using Trac ID "Util" and current Trac password.
21:47 Util Does something need to be poked to restore my commit bit?
21:47 dalek tracwiki: https://trac.parrot.org/parro​t/wiki/ChrootSetup?version=10
21:50 kj joined #parrot
21:50 Coke Util: your account doesn't have 'developer' access.
21:50 Coke I would add you, but as a sanity check I tried to find your CLA but don't see it.
21:50 kj g'evening all
21:51 Util CLA?
21:51 purl i guess CLA is Contributor License Agreement or http://www.perlfoundation.org/​contributor_license_agreement or http://www.parrot.org/foundation/legal
21:51 Util I signed one at YAPC some year past - Robert Bruce Gray III
21:51 kj I'd like to override the standard syntax error handling sub, that comes with PGE-generated grammars. Anyone an idea how to do that?
21:52 Coke kj: something like this: http://code.google.com/p/partcl/sourc​e/browse/trunk/runtime/tcllib.pir#281
21:52 shorten Coke's url is at http://xrl.us/beetm7
21:53 Coke util, ah, your nick was missing.
21:53 Coke fixed on the CLA list...
21:54 Util Many thanks!
21:54 Infinoid Anyone else getting Packfile-related failures in t/pmc/eval.t?
21:54 rg yes, me too
21:54 kj Coke: aah. it took a while before I understood; it's the bold-printed line no, right?
21:55 Coke Util: try now.
21:55 Coke kj: yes.
21:55 kj Coke: thanks!
21:56 Coke I override the ws there.
21:56 rg infinoid: and all the t/pmc/packfile*.t
21:56 Coke hopefully you can use the same trick to override some other rule.
21:56 Infinoid rg: Funny, those pass here.
21:56 Util Coke: Works now. Thanks again!
21:56 dalek parrot: r36340 | Util++ | trunk:
21:56 dalek parrot: Fix TT #256, out-of-sync pointer in unshift_string() in PMC ResizableStringArray
21:56 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36340/
21:56 Coke Util++
21:56 kj yes instead of 'syntax error blablabla' I'd like to print "I'm sorry but I don't know what you mean" :-)
21:57 moritz are you writing an IRC bot? ;-)
21:57 kj moritz: no a text adventure engine
21:57 kj at least, that's the plan
21:57 moritz I think that's nearly the same ;-)
21:57 kj haha yep
22:00 Whiteknight joined #parrot
22:06 nopaste "infinoid" at 96.238.213.50 pasted "TT #271 test patch" (46 lines) at http://nopaste.snit.ch/15493
22:06 Infinoid janus: Does http://nopaste.snit.ch/15493 build for you on openbsd/sparc64?
22:07 dalek parrot: r36341 | allison++ | trunk/src/jit/i386/core.jit:
22:07 dalek parrot: [strings] Caught one stray reference to old version of 'Parrot_str_equal' and
22:07 dalek parrot: updated for backward compatibility.
22:07 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36341/
22:21 braceta left #parrot
22:35 bacek joined #parrot
22:53 hercynium joined #parrot
22:59 Whiteknight joined #parrot
23:01 dalek tracwiki: v11 | allison++ | ChrootSetup
23:01 dalek tracwiki: https://trac.parrot.org/parro​t/wiki/ChrootSetup?version=11
23:28 dalek parrot: r36342 | allison++ | trunk/ports/debian:
23:29 dalek parrot: [cage] Adding generated debian control file explicitly, so it can be
23:29 dalek parrot: used in Ubuntu packaging unchanged.
23:29 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36342/
23:41 rurban_ joined #parrot
23:42 bacek_ joined #parrot
23:46 s1n Coke_afk: thanks, i'll test it out later tonight
23:51 dalek parrot: r36343 | jkeenan++ | branches/closure_for_closure:
23:51 dalek parrot: Creating closure_for_closure in https://svn.parrot.org/parrot/branches
23:51 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36343/
23:51 dalek parrot: r36344 | jkeenan++ | tags/closure_for_closure-36342:
23:51 dalek parrot: Tagging trunk at r36342 so that the closure_for_closure can later be synched to it.
23:51 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36344/

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

Parrot | source cross referenced