Camelia, the Perl 6 bug

IRC log for #parrot, 2012-03-18

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
01:13 dalek parrot/remove_sub_flags: d69483e | Whiteknight++ | t/ (4 files):
01:13 dalek parrot/remove_sub_flags: Fix more failing test files. Delete a few tests that were explicitly relying on deprecated behavior
01:13 dalek parrot/remove_sub_flags: review: https://github.com/parrot/parrot/commit/d69483eed4
01:13 dalek parrot/remove_sub_flags: e3b5b75 | Whiteknight++ | t/src/embed/api.t:
01:13 dalek parrot/remove_sub_flags: Fix t/src/embed/api.t. Without setting the config hash, there are no search paths to load_language from. The test was previously passing for the wrong reason
01:13 dalek parrot/remove_sub_flags: review: https://github.com/parrot/parrot/commit/e3b5b759f6
01:13 dalek parrot/remove_sub_flags: bb9b863 | Whiteknight++ | / (4 files):
01:13 dalek parrot/remove_sub_flags: Delete t/pir/pbc_fallback.t. It is unnecessary. Update PDD19 to be more clear about :postcomp and :immediate behavior in modern Parrot. Update t/pmc/sub.t, removing invalid old tests and fixing all other broken tests
01:13 dalek parrot/remove_sub_flags: review: https://github.com/parrot/parrot/commit/bb9b8639e6
01:14 whiteknight it's amazing how I'm finding tests that apparently were passing, but didn't do the right thing
01:14 whiteknight poorly written tests are worse than no test at all
01:39 dukeleto ~~
01:54 alester joined #parrot
02:15 kid51 joined #parrot
02:15 aloha (parrot/parrot) Issues opened : 738 (Build failures under clang and g++) by jkeenan : https://github.com/parrot/parrot/issues/738
02:17 kid51 alester: Could you take a look at GI 738 and see if any of your recent work might have caused a problem with g++ builds?
02:17 alester Yes, and I'm about to push it.
02:17 kid51 Merci beaucoup
02:18 kid51 dukeleto: Can you post any error output from those clang-optimize build failures in GI 738?
02:18 kid51 (or paste)
02:19 * kid51 goes back to writing a talk where he channels alester vintage 2004 :-)
02:19 alester Heh, what about, testing?
02:20 kid51 Yes, I'll be giving a Perl Testing the Basics talk at DC-Baltimore Perl Workshop and at YAPC::NA
02:22 dalek parrot: 4735f5b | petdance++ | ChangeLog:
02:22 dalek parrot: fix typos
02:22 dalek parrot: review: https://github.com/parrot/parrot/commit/4735f5b7b7
02:22 dalek parrot: 138dc62 | petdance++ | / (2 files):
02:22 dalek parrot: Standard program code as unsigned char
02:22 dalek parrot: review: https://github.com/parrot/parrot/commit/138dc625e0
02:24 kid51 Hmm, 138dc62 didn't fix the problem.  Are there more fixes to come?
02:26 nopaste "kid51" at 192.168.1.3 pasted "138dc62 did not fix g++ build failure" (633 lines) at http://nopaste.snit.ch/123373
02:32 alester How does one build nap?
02:32 alester nqp?
02:32 alester also, I can't figure out how frontend/parrot2/prt0.pbc.c: gets created
02:35 kid51 Well, in my "normal" build log (gcc), it first shows up here:
02:35 kid51 597:./miniparrot -Iruntime/parrot/include -o frontend/parrot2/prt0.pbc frontend/parrot2/prt0.pir
02:35 kid51 598:./miniparrot frontend/parrot2/build.pir frontend/parrot2/prt0.pbc
02:35 kid51 599:frontend/parrot2/prt0.pbc.c
02:36 alester kthx
02:36 alester That's my last build bummer.
02:37 alester It's not including parrot/api.h
02:37 plobsing joined #parrot
02:37 alester which is now mandatory if the compiler is to get the function prototypes
02:37 nopaste "kid51" at 192.168.1.3 pasted "prt0 logged during an all gcc build" (7 lines) at http://nopaste.snit.ch/123374
02:42 alester ok, so then what builds frontend/parrot2/prt0.pir   if that's what builds the .c?
02:44 alester Aha, found it.
02:49 kid51 Whatever it is, it's not affecting gcc build.
03:09 dalek parrot: 6fac2cd | petdance++ | frontend/parrot2/build.pir:
03:09 dalek parrot: built programs have to include parrot/api.h
03:09 dalek parrot: review: https://github.com/parrot/parrot/commit/6fac2cdec2
03:16 aloha (parrot/parrot) Issues closed : 738 (Build failures under clang and g++) by jkeenan : https://github.com/parrot/parrot/issues/738
03:30 alester kid51: http://tt.taptinder.org/bu​ildstatus/parrot/master/4
03:45 wagle_ joined #parrot
04:25 preflex_ joined #parrot
05:06 aloha (parrot/parrot) Issues reopened : 729 (Memory leak with hop_buckets in ops) by petdance : https://github.com/parrot/parrot/issues/729
05:33 dalek parrot: ab45201 | petdance++ | src/ops/core_ops.c:
05:33 dalek parrot: rebootstrapped to get the fix from GitHub #729
05:33 dalek parrot: review: https://github.com/parrot/parrot/commit/ab4520154a
05:37 aloha (parrot/parrot) Issues closed : 729 (Memory leak with hop_buckets in ops) by petdance : https://github.com/parrot/parrot/issues/729
05:47 aloha (parrot/parrot) Issues opened : 739 (Fix prototype for Parrot_get_config_hash_bytes) by petdance : https://github.com/parrot/parrot/issues/739
06:02 sorear Why is aloha opped?
07:36 cotto seen fperrad
07:36 aloha fperrad was last seen in #parrot 21 hours 10 mins ago joining the channel.
08:19 mvuets joined #parrot
08:35 fperrad joined #parrot
08:40 bacek joined #parrot
08:51 mvuets| joined #parrot
09:42 bacek joined #parrot
09:53 jsut_ joined #parrot
10:30 preflex_ joined #parrot
10:33 bacek joined #parrot
10:46 dalek parrot: daad3e8 | fperrad++ | / (5 files):
10:46 dalek parrot: [mingw] fix build
10:46 dalek parrot:
10:46 dalek parrot: broken since https://github.com/parrot/parrot/commit/​d923903bdf2183c223cca1b2887055fcbac0e588
10:46 dalek parrot: where warning implicit-function-declaration becomes an error.
10:46 dalek parrot: so, now probe process.h which declares _getpid.
10:46 dalek parrot: review: https://github.com/parrot/parrot/commit/daad3e865e
10:50 bacek joined #parrot
10:58 mj41 joined #parrot
11:03 travis-ci joined #parrot
11:03 travis-ci [travis-ci] parrot/parrot#172 (master - daad3e8 : Francois Perrad): The build is still failing.
11:03 travis-ci [travis-ci] Change view : https://github.com/parrot/par​rot/compare/ab45201...daad3e8
11:03 travis-ci [travis-ci] Build details : http://travis-ci.org/parrot/parrot/builds/891576
11:03 travis-ci left #parrot
11:07 bacek joined #parrot
11:24 bacek joined #parrot
11:32 lateau joined #parrot
12:04 dalek rakudo/nom: 5264177 | moritz++ | t/spectest.data:
12:04 dalek rakudo/nom: run pre-post.t
12:04 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/5264177fe4
12:11 whiteknight joined #parrot
12:37 whiteknight good morning, #parrot
12:46 kid51 joined #parrot
12:48 kid51 msg dukeleto Did alester's fixes clear up clang build failures along with g++ failures (https://github.com/parrot/parrot/issues/738)?
12:48 aloha OK. I'll deliver the message.
12:49 kid51 Hmm, getting failure in checkdepend.t
12:56 nine http://www.infoworld.com/d/application-devel​opment/van-rossum-python-not-too-slow-188715
12:56 nine Apparently Van Rossum still thinks a GIL an appropriate solution
12:56 whiteknight nine: yeah, I saw that
12:56 whiteknight nine: that's because he probably hasn't seen the awesome system we're building :)
12:57 nine Can't wait to see it finished either ;)
12:57 whiteknight I still need to get those green threads working on windows first. I'm slacking there.
12:58 PacoAir joined #parrot
13:00 nine We'll get there. Meanwhile, I got my moretasks.pir to run 100 % stable. This test creates 200,000 tasks executed on 8 threads which wait in a loop containing sleep 0.1 for a shared starter variable to become 1. Then all tasks just print their number and everything starts again. It gets killed eventually for eating all memory though.
13:10 whiteknight that's awesome. Why does it eat all memory? GC should clean all those pmcs up
13:13 nine Hard to say. I'm totally lacking tools and experience for digging that up. Till now I've spent my time getting it to run stable. And there's more to do in this regard because what the test not yet does is using schedule_proxied to write results back to the main thread.
13:15 whiteknight okay, that's fair. We can track down memory leaks later
13:16 nine All in all, I'd say things start to look really shiny
13:16 whiteknight yes, yes they do
13:17 whiteknight nine: if you have an official fan club, I'd like to join and get my decoder ring
13:22 nine Well, we can talk about such things, when things move closer to master ;)
13:23 dalek parrot: 3bb3ac9 | jkeenan++ | config/gen/makefiles/root.in:
13:23 dalek parrot: Add dependencies revealed by running t/src/checkdepend.t as part of 'make test'.
13:23 dalek parrot: review: https://github.com/parrot/parrot/commit/3bb3ac9c1c
13:39 travis-ci joined #parrot
13:39 travis-ci [travis-ci] parrot/parrot#173 (master - 3bb3ac9 : jkeenan): The build is still failing.
13:39 travis-ci [travis-ci] Change view : https://github.com/parrot/par​rot/compare/daad3e8...3bb3ac9
13:39 travis-ci [travis-ci] Build details : http://travis-ci.org/parrot/parrot/builds/892443
13:39 travis-ci left #parrot
13:52 * kid51 is unimpressed with travis-ci
13:56 dalek parrot: b714b90 | jkeenan++ | config/gen/makefiles/root.in:
13:56 dalek parrot: Extend use of Makefile variable $(INC_DIR) to two unused cases.
13:56 dalek parrot: review: https://github.com/parrot/parrot/commit/b714b9044d
14:03 whiteknight joined #parrot
14:11 travis-ci joined #parrot
14:11 travis-ci [travis-ci] parrot/parrot#174 (master - b714b90 : jkeenan): The build is still failing.
14:11 travis-ci [travis-ci] Change view : https://github.com/parrot/par​rot/compare/3bb3ac9...b714b90
14:11 travis-ci [travis-ci] Build details : http://travis-ci.org/parrot/parrot/builds/892653
14:11 travis-ci left #parrot
14:42 aloha (parrot/parrot) Issues opened : 740 (Make compiler detection on Windows more robust) by gerdr : https://github.com/parrot/parrot/issues/740
14:46 aloha (parrot/parrot) Issues opened : 740 (Make compiler detection on Windows more robust) by gerdr : https://github.com/parrot/parrot/issues/740
14:51 aloha (parrot/parrot) Issues opened : 741 (Remove redundant function declarations) by petdance : https://github.com/parrot/parrot/issues/741
14:56 lateau joined #parrot
15:05 dalek parrot/kid51/makefile_variables: c66fc77 | jkeenan++ | config/gen/makefiles/root.in:
15:05 dalek parrot/kid51/makefile_variables: Introduce use of Makefile variable $FRP_DIR.
15:05 dalek parrot/kid51/makefile_variables: review: https://github.com/parrot/parrot/commit/c66fc77b51
15:05 dalek parrot/kid51/makefile_variables: 4f50140 | jkeenan++ | config/gen/makefiles/root.in:
15:05 dalek parrot/kid51/makefile_variables: Introduce use of Makefile variable $FR_DIR.
15:05 dalek parrot/kid51/makefile_variables: review: https://github.com/parrot/parrot/commit/4f50140ec8
15:45 Psyche^ joined #parrot
15:46 dalek parrot/remove_sub_flags: d6a5a9b | Whiteknight++ | / (9 files):
15:46 dalek parrot/remove_sub_flags: Add languages/parrot/parrot.pir to the build, since we can't auto-load .pir files anymore. Fix several test files, removing obsolete tests. Remove some debugging stuff from src/packfile/api.c. Fix PCT to load P6object.pbc as necessary. Only one test still failing.
15:46 dalek parrot/remove_sub_flags: review: https://github.com/parrot/parrot/commit/d6a5a9b784
15:51 catalin joined #parrot
15:56 catalin joined #parrot
15:59 nine Isn't Parrot_block_GC_mark(interp); supposed to prevent the GC from running on interp?
16:00 whiteknight yes
16:00 whiteknight it increases a counter. If the counter is non-zero GC should not run
16:03 nine Interesting... according to gdb self->gc_mark_block_level is 1, but this is inside gc_gms_mark_and_sweep which should not be possible
16:07 benabik G'morning, #parrot
16:13 benabik Huh.  We're starting to get those introduction letters to -dev.  I guess I should do that whole "introduce yourself to the community" thing, huh?  ;-)
16:14 kid51 benabik: +1
16:19 lucian joined #parrot
16:21 moritz even better if you have a project idea
16:23 benabik moritz: Given that I added three or four projects to the ideas page, I think I do.  :-D
16:31 mj41 joined #parrot
16:35 dalek parrot/kid51/makefile_variables: 818a583 | jkeenan++ | config/gen/makefiles/root.in:
16:35 dalek parrot/kid51/makefile_variables: Remove two superfluous instances of Makefile variable $(MINIPARROT).
16:35 dalek parrot/kid51/makefile_variables: review: https://github.com/parrot/parrot/commit/818a583a4a
16:45 dalek rakudo/nom: d1505ae | moritz++ | src/ (2 files):
16:45 dalek rakudo/nom: Typed exceptions from PRE/POST failures
16:45 dalek rakudo/nom:
16:45 dalek rakudo/nom: jnthn++ for the help
16:45 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/d1505aee7c
16:59 dalek parrot/kid51/makefile_variables: 6e8c693 | jkeenan++ | config/gen/makefiles/root.in:
16:59 dalek parrot/kid51/makefile_variables: Introduce Makefile variable $(FRPTWO_DIR) for 'frontend/parrot2'.
16:59 dalek parrot/kid51/makefile_variables: review: https://github.com/parrot/parrot/commit/6e8c693840
17:11 nine whiteknight: like I feared: I think we're taking read accesses too lightly
17:12 nine whiteknight: even read accesses can lead to allocation of new PMCs and thus to GC runs
17:13 benabik How does read access allocate PMCs?
17:13 dalek parrot/kid51/makefile_variables: 97162f8 | jkeenan++ | config/gen/makefiles/root.in:
17:13 dalek parrot/kid51/makefile_variables: Introduce Makefile variable $(RUN_INC_DIR) for runtime/parrot/include.
17:13 dalek parrot/kid51/makefile_variables: review: https://github.com/parrot/parrot/commit/97162f8b79
17:30 dukeleto msg kid51 what makes you unimpressed with travis ci?
17:30 aloha OK. I'll deliver the message.
17:31 dalek parrot/kid51/makefile_variables: cec7cc8 | jkeenan++ | config/gen/makefiles/root.in:
17:31 dalek parrot/kid51/makefile_variables: Use Makefile variable $(LIBRARY_DIR) more consistently.
17:31 dalek parrot/kid51/makefile_variables: review: https://github.com/parrot/parrot/commit/cec7cc87af
17:31 dalek parrot/kid51/makefile_variables: 21b8758 | jkeenan++ | config/gen/makefiles/root.in:
17:31 dalek parrot/kid51/makefile_variables: Cannot use a Makefile variable in an include statement.  Fix.
17:31 dalek parrot/kid51/makefile_variables:
17:31 dalek parrot/kid51/makefile_variables: (Or so it appears.)
17:31 dalek parrot/kid51/makefile_variables: review: https://github.com/parrot/parrot/commit/21b87585bd
17:32 dukeleto optimized clang is still broken on master: http://travis-ci.org/#!/parrot/parrot/jobs/892654
17:32 * dukeleto has gotten rid of the travis timeout problem, for now
17:36 * dukeleto builds an optimized clang parrot
17:40 dukeleto the optimized clang failure are all GMP/bigint related, seemingly
17:40 dukeleto I can reproduce the issues on my laptop
17:44 dukeleto these are the optimized clang failures on my machine (which is slightly different than on travis) http://smolder.parrot.org/app/​projects/report_details/26690
17:52 dukeleto VMware wants to put their virtual machines in your browser: http://blog.chipx86.com/2012/03/13/w​sx-virtual-machines-in-your-browser/
17:53 benabik I heard you like browsers, so we put a computer running a browser in your browser.
17:57 dukeleto what could go wrong?
18:04 dalek rakudo/nom: 924712d | moritz++ | src/ (2 files):
18:04 dalek rakudo/nom: move PRE/POST error message logic to the error class; trim condition text
18:04 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/924712d262
18:19 not_gerd joined #parrot
18:20 nine benabik: I'm actually not sure about that part. I think I remember seeing some stack traces that looked like. Also I remember a comment saying that parameter passing can cause GC
18:22 benabik nine: Oh, yes.  I think PCC results in allocation.  Oy.
18:23 nine Which would be a bit catastrophic if it happens while accessing from a different thread. But simply Parrot_block_GC_marking is not the solution either, since the thread whose data we're accessing might just now run the GC
18:38 aloha (parrot/parrot) Issues opened : 742 (Parrot_run_callback uses void* to pass arbitrary data) by gerdr : https://github.com/parrot/parrot/issues/742
18:44 janus joined #parrot
18:48 not_gerd any objections to dropping support for Windows versions < Windows 2000?
18:49 Coke what does it buy us?
18:50 not_gerd fixing win64 g++ build without having to hack the configuration system
18:51 not_gerd actually, minimum requirement already *is* windows 2000 according to socket.c - it just seems the configuration system is broken...
18:53 bacek joined #parrot
19:06 not_gerd found the issues - strawberry perl uses gmake instead of dmake now
19:28 aloha (parrot/parrot) Issues closed : 740 (Make compiler detection on Windows more robust) by gerdr : https://github.com/parrot/parrot/issues/740
19:33 aloha (parrot/parrot) Issues opened : 743 (Fix g++ issues with Strawberry Perl on Win64) by gerdr : https://github.com/parrot/parrot/issues/743
19:41 benabik joined #parrot
19:44 benabik joined #parrot
20:04 mj41 joined #parrot
20:17 perlite_ joined #parrot
20:31 not_gerd left #parrot
20:44 donaldh joined #parrot
20:46 preflex_ joined #parrot
21:06 dalek rakudo/nom: 8f4a063 | jonathan++ | src/ (2 files):
21:06 dalek rakudo/nom: POST block gets $_ set to the result.
21:06 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/8f4a063196
21:08 plobsing joined #parrot
21:15 dalek rakudo/nom: 1968b89 | jonathan++ | src/ops/perl6.ops:
21:15 dalek rakudo/nom: If one POST fails, don't run any more. The spectest for this still fails as it expects POST to run in declaration order, not reverse order.
21:15 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/1968b89b59
22:27 preflex_ joined #parrot
23:06 bacek_at_work joined #parrot
23:15 bacek_at_work ~~
23:16 aloha joined #parrot
23:17 bacek_at_work aloha, humans
23:25 humans aloha, bacek_at_work
23:57 bacek_at_work seen alester
23:57 aloha alester was last seen in #parrot 20 hours 26 mins ago saying "kid51: http://tt.taptinder.org/bu​ildstatus/parrot/master/4".
23:57 bacek_at_work msg alester we should probably install headerizer & co with parrot to help building something like nqp with new set of compiler flags.
23:57 aloha OK. I'll deliver the message.

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

Parrot | source cross referenced