|
Parrot 4.8.0 "Spix's Macaw" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC Set by moderator on 20 September 2012. |
|||
|
00:01
kid51_ joined
00:05
jimk joined
00:06
benabik joined
00:18
kid51_ joined,
benabik joined
00:47
rurban_mobile joined
00:51
benabik joined
|
|||
| rurban_mobile | kid51 again merged the threads branch into master, 3rd time. sigh | 01:00 | |
| I did it only once. | |||
| dalek | rrot: 57382f8 | rurban++ | t/dynpmc/select.t: Avoid TODO in t/dynpmc/select.t on known passes |
01:21 | |
|
01:24
benabik joined
|
|||
| benabik | I think kid51 pulled while threads was in master, so pulls afterwards create a merge between old master and new master. | 01:25 | |
| If we had just committed an "anti-threads" revert commit, this wouldn't happen. | |||
| (For reference, this is _exactly_ why rewriting published history is frowned upon.) | 01:27 | ||
| rurban_mobile | kid51: please do git fetch; git reset --hard origin/master; you are always merging threads in your commits. | 01:28 | |
| benabik | I think we need kid51 to do something like `git rebase --onto origin/master origin/threads <branch>` for all of his branches. | ||
| rurban_mobile | benabik: do you want to do this anti-thread revert commit? | 01:29 | |
| benabik | And we need to never rewrite master in the future. | ||
| Is threads currently merged in parrot/master? | |||
| rurban_mobile | nope, of course not. that's whay I have to cherry-pick from him and revert his commits | ||
| but rebase is worse. reset hard origin/master is easier | 01:30 | ||
|
01:30
travis-ci joined
|
|||
| travis-ci | [travis-ci] parrot/parrot#587 (master - 57382f8 : Reini Urban): The build was broken. | 01:30 | |
| [travis-ci] Change view : github.com/parrot/parrot/compare/2...382f873591 | |||
| [travis-ci] Build details : travis-ci.org/parrot/parrot/builds/2552308 | |||
|
01:30
travis-ci left
|
|||
| benabik | Reset hard _loses data_ | 01:30 | |
| rurban_mobile | yes, it would loose his wrong merge | ||
| benabik | And all the work done on his branches. | ||
| rurban_mobile | it's only for master | 01:31 | |
| nope, only master | |||
| benabik | rurban_mobile: Can you give me an example branch where this happened? | 01:33 | |
| I want to see if he's been creating new merges or not... | |||
| It might be that the branches just pre-date the removal. | |||
| rurban_mobile | its on two of his latest branches: 840/has_libffi (already deleted) and one other with cpu | 01:34 | |
| cpu_type_eliminate_gpp_warning it was | |||
| f2d576e0e3bd3991e16bf9b04586b82d3094b7f6 is such a bad merge | 01:35 | ||
| Coke | ugh, we rewrote published history? who did that? | ||
| benabik | Yeah, you can see the repeated "Merge branch 'master' of git@github.com:parrot/parrot | ||
| rurban_mobile | Coke: I reset one bad merge (threads) | 01:36 | |
| benabik | I was against it at the time, although I didn't voice that terribly strongly. | ||
| Coke | rurban_mobile: please don't ever do that again. | ||
| rurban_mobile | git revert -m 1 did not work out for me this time | ||
| Coke: Sure. | |||
| benabik | rurban_mobile: As I recall, it actually needed to be "-m 2" | ||
| rurban_mobile | Now I have to cherry pick from kid's branches, and revert all his commits | 01:37 | |
| Coke | I am wondering if we have already done it once, will doing it one more time hurt any more? | ||
| I'd ask on list, and at least get buyin from dukeleto on how to clean it up. | 01:38 | ||
| (in general, though, let's avoid rewriting) | |||
| rurban_mobile | It was an undo of one last merge. A simple git pull would have fixed it. | 01:40 | |
| benabik | I think a big NEVER REWRITE MASTER should be our policy. Even if the "wrong" history only existed for minutes, it's a race condition: Who pulled before the new one was pushed? | ||
| rurban_mobile: No. Git-pull creates new merged. | |||
| *merges | |||
| ttbot | Parrot ebe1e7c6 MSWin32-x86-multi-thread make error tt.taptinder.org/cmdinfo/97567 | ||
| benabik | Which is exactly what's been happening to poor kid51. | ||
| rurban_mobile | Oh, I see | ||
| sorear | benabik: this wouldn't be a problem if we could just contact kid51 (?) | 01:41 | |
| Coke | and anyone else who happens to be working off a bad master. | ||
| benabik | Right. | ||
| I think kid51 is the only one affected. I'm going to post to parrot-dev a probable fix... | 01:42 | ||
| rurban_mobile: Do you remember what commit was the re-write? | |||
| Blub. Actually, I don't have time for this. | 01:43 | ||
| Short version `git rebase --onto origin/master <removed master commit> <branch>` should be done for all affected branches. | 01:44 | ||
| This degenerates to `git reset origin/master` for any branch (local master?) which doesn't have any local changes. | |||
| obviously I don't suggest doing this for any branches published to parrot/parrot. Just use cherry-pick or wait until post-threads to merge those. | 01:45 | ||
| `git branch --contains <removed master>` gives all affected commits (unless that commit is still in threads? | 01:46 | ||
|
01:49
l3l1p joined
|
|||
| benabik said "don't have time" because before writing that up for the list and kid51, I would like to be far more sure I'm right, including determining the commit and testing it. | 01:51 | ||
| But homework and crying kid both call. | |||
|
02:08
l3l1p joined
02:12
sivoais joined
02:19
l3l1p joined
|
|||
| ttbot | Parrot 57382f87 MSWin32-x86-multi-thread make error tt.taptinder.org/cmdinfo/97579 | 02:52 | |
|
03:35
rurban_mobile joined
|
|||
| dalek | rtcl-nqp/nqp2: d3368fc | coke++ | / (4 files): Move "invoke" to a helper class, refer to it as a WVal |
03:36 | |
| rtcl-nqp/nqp2: 7d076be | coke++ | src/Partcl/Actions.pm: switch ::op to use callmethod. |
03:52 | ||
| rurban_mobile | I think it should be git revert -m 1 a54a581d2392a885c70093349382b5da696ec4dd | 03:54 | |
| benabik | rurban_mobile: A real "revert merge" commit would involve merging threads, then reverting. It's probably easier to try to fix kid51 right now. | 04:14 | |
| MikeFair goes to bed as well. | 04:27 | ||
| rurban_mobile | msg kid51 you based your branches on a rewritten master, which still includes the wrong threads merge. big problem. please git reset --hard origin/master and rebase your branches onto the new master then please. | 04:36 | |
| aloha | OK. I'll deliver the message. | ||
|
05:55
l3l1p joined
07:08
brrt joined
07:33
Psyche^ joined
|
|||
| brrt | hi #parrot | 07:36 | |
|
07:53
lucian joined
11:39
mtk joined
11:52
rurban_mobile joined
11:57
Liz joined
11:58
woolfy joined
|
|||
| rurban | hi #parrot | 12:07 | |
| brrt | hi rurban | 12:12 | |
| rurban | ewontfix.com/2/ on our thread cancellation problem | 12:20 | |
| dalek | rrot: 64fd80f | rurban++ | t/pmc/ (4 files): [cage] Fix 4 asynch typos in tests |
12:24 | |
|
12:34
travis-ci joined
|
|||
| travis-ci | [travis-ci] parrot/parrot#588 (master - 64fd80f : Reini Urban): The build was fixed. | 12:34 | |
| [travis-ci] Change view : github.com/parrot/parrot/compare/5...fd80fb3e8a | |||
| [travis-ci] Build details : travis-ci.org/parrot/parrot/builds/2556109 | |||
|
12:34
travis-ci left
12:56
PacoAir joined
13:16
rurban_mobile joined
13:25
bluescreen joined
|
|||
| ttbot | Parrot 64fd80fb MSWin32-x86-multi-thread make error tt.taptinder.org/cmdinfo/97742 | 13:44 | |
| Coke points at boingboing.net/2012/09/24/adobe-rel...e-cod.html for a nice new(?) font. | 13:45 | ||
| dalek | rrot/rurban/clang++-gh844: d75558b | rurban++ | / (4 files): [GH #844] add missing auto::alignof files Add files to MANIFEST Use auto::alignof as last auto step |
14:09 | |
| rrot/rurban/clang++-gh844: d582800 | rurban++ | src/dynpmc/gziphandle.pmc: [GH #844] Make gziphandle C++ safe Wrap #include <zlib.h> into PMC_HEADER_PREAMBLE to be able to use ATTR gzFile file. This is now C++ safe and clang++ compiles fine. |
14:36 | ||
| rurban | clang++ all PASS smolder.parrot.org/app/projects/rep...ails/31718 | 14:44 | |
|
14:48
benabik joined
14:52
dmalcolm joined
|
|||
| rurban | checking now threads with g++ | 14:54 | |
| dalek | Heuristic branch merge: pushed 17 commits to parrot/rurban/clang++-gh844 by rurban | 14:59 | |
| rurban | nine: pcc_unproxy_context had a worrying pointer bug. Parrot_Interp * const target_interp => Parrot_Interp const target_interp | 15:03 | |
| So Parrot_thread_create_proxy worked on a wrong ptr... | 15:04 | ||
| g++ and esp. clang++ really give us superior diagnostics. | 15:05 | ||
| dalek | rrot: fdd0ca3 | rurban++ | src/dynpmc/gziphandle.pmc: [GH #844] Make gziphandle C++ safe Wrap #include <zlib.h> into PMC_HEADER_PREAMBLE to be able to use ATTR gzFile file. This is now C++ safe and clang++ compiles fine. |
15:29 | |
| rurban | clang++ does not compile with this alone, it's just a cherry-pick from rurban/clang++-gh844 needed for threads also | ||
|
15:53
travis-ci joined
|
|||
| travis-ci | [travis-ci] parrot/parrot#589 (master - fdd0ca3 : Reini Urban): The build was broken. | 15:53 | |
| [travis-ci] Change view : github.com/parrot/parrot/compare/6...d0ca3cc18c | |||
| [travis-ci] Build details : travis-ci.org/parrot/parrot/builds/2558751 | |||
|
15:53
travis-ci left
16:00
brrt left
|
|||
| ttbot | Parrot fdd0ca3c MSWin32-x86-multi-thread make error tt.taptinder.org/cmdinfo/97778 | 16:11 | |
| allison | Coke: just compatibly licensed, so we don't trip people up | 16:31 | |
| Coke: like BSD/MIT is totally fine, GPL is trickier (better to make it a package dependency) | |||
| Coke: but mostly, I never did grok the Parrot love of sucking everything into the tarball | 16:32 | ||
| Coke: it's just not necessary, the world is full of good packaging systems today | |||
| Coke | depends on if it's optional or a requirement, I think. and if you need to patch it. | 17:08 | |
| I endorse rakudo bundling some stuff that is pretty much required and needs patching, e.g. | |||
| all a tradeoff. in general, though, yes, let's default to making it an external dep. | 17:09 | ||
| allison: you still with ubuntu? | |||
| allison | Coke: even if it's required, it's better to use the packaged version than install your own | 17:11 | |
| thousands of man hours go into maintaining packages | |||
| fixing security issues | |||
| fixing bugs | |||
| tracking the latest from the library maintainers | |||
| embedding our own version means we lose all that value | 17:12 | ||
| (and it's not like Parrot is so overstaffed with volunteers that we can really afford to miss out on free development work) | |||
| Coke: I'm still an ubuntu volunteer, but not working at Canonical anymore | 17:13 | ||
| embedding a patched version of an upstream library is an even greater maintenance burden | 17:14 | ||
| subclassing or wrapping with changes is better | 17:15 | ||
| basically, anything in the parrot tarball becomes the responsibility of the parrot developers | 17:16 | ||
| anything external is just about free | |||
| minor integration/testing with each new release of the particular distro/os | 17:17 | ||
| Coke | fair enough. | 17:18 | |
| "what about windows?" ;) | |||
| (no answer required). | 17:20 | ||
| allison | windows is it's own universe, make strawberry parrot :) | ||
| Active Perl and Strawberry Perl pretty much are the "perl-centric" packaging systems for Windows | 17:21 | ||
| but, that doesn't mean the Perl5 core itself should include all the cruft that windows needs | |||
|
17:23
contingencyplan joined
17:58
davidfetter joined
18:00
benabik joined
18:08
benabik joined
|
|||
| dalek | rrot/threads: 2244666 | dukeleto++ | src/call/context.c: Make g++ happy on the threads branch |
18:16 | |
| rurban | dukeleto: I think your fix is wrong, and the compiler is right. I'm stepping through the debugger right now. | 18:19 | |
| dalek | rrot/threads: 789cea7 | rurban++ | src/call/context.c: Revert "Make g++ happy on the threads branch" This reverts commit 22446667e8dbd29c9342f9402f8766f3dc295437. Wrong fix, see 02eece1 instead. |
18:23 | |
| rrot/threads: 8765cd4 | rurban++ | src/call/context.c: Fix target_interp in pcc_unproxy_context detected by g++ Parrot_Interp is a ptr to the interp struct already. This code section was never triggered in any testcase. I am trying to find a testcase for this. |
|||
|
18:27
travis-ci joined
|
|||
| travis-ci | [travis-ci] parrot/parrot#590 (threads - 2244666 : Jonathan "Duke" Leto): The build was fixed. | 18:27 | |
| [travis-ci] Change view : github.com/parrot/parrot/compare/0...446667e8db | |||
| [travis-ci] Build details : travis-ci.org/parrot/parrot/builds/2560773 | |||
|
18:27
travis-ci left
|
|||
| dalek | rrot/rurban/pmc2c-unused-gh836: ac0b135 | rurban++ | lib/Parrot/Pmc2c/Method.pm: [GH #836] Catch also Scheduler.destroy and empty Undef bodies |
18:45 | |
| rrot: 60cfaca | rurban++ | lib/Parrot/Pmc2c/Method.pm: [GH #836] Improve parsing pmc2c UNUSED arguments UNUSED(arg) amcros are now detected in pmc bodies and change the argument declaration to SHIM(). Empty bodies with return 1; are now detected and the two args are SHIMed. UNUSED(interp) cause a new warning to be replaced by UNUSED(INTERP). |
|||
| rrot: 6afb068 | rurban++ | config/auto/warnings.pm: warnings: Improve clang support clang requires warnings overrides for -Wno-parentheses-equality and adds -Wno-unused-result to src/ops/core_ops.c. (used in 29582199aef9ea9462 [GH #828] Optimize print_n/i and say_n/i) |
|||
| rrot: a95f757 | rurban++ | config/gen/makefiles/root.in: Add ccwarn::src/ops/core_ops.c SUFFIX OVERRIDE for using (void)Parrot_io_write_b Added with 29582199aef9ea94622b4202aee129521bc |
|||
| rrot: ef4b571 | rurban++ | src/ (3 files): Add more UNUSED() declarations to key.pmc and core.ops |
|||
| rrot: 4045afa | rurban++ | / (30 files): merge master into rurban/pmc2c-unused-gh836 Fix conflict in config/auto/warnings.pm |
|||
| rrot: 1d37b98 | rurban++ | config/auto/warnings.pm: fixup wrong config/auto/warnings.pm merge |
|||
| rrot: 562f9ab | rurban++ | lib/Parrot/Pmc2c/Method.pm: [GH #836] Allow UNUSED(word); the semicolon |
|||
| rrot: f4a63d5 | rurban++ | src/pmc/default.pmc: [GH #836] default thaw: SELF.init uses INTERP |
|||
| rrot: 7c350cf | rurban++ | config/gen/makefiles/root.in: pmc2c: use the current libs not the installed pmc2c favored installed Parrot::Pmc2c libs over the current ones, which made debugging Parrot::Pmc2c methods hard. |
|||
| rrot: 847cb9b | rurban++ | / (13 files): [GH #836] Cleaned wrong UNUSED(arg) macros in all pmc's Due to an improved pmc2c compiler, which SHIMs all UNUSED args, a lot of wrong and/or badly placed UNUSED macros have been fixed. |
|||
| rrot: 7913e25 | rurban++ | / (2 files): [codingstd] GH #836 fix hard tabs |
|||
| rrot: 0525faa | rurban++ | / (16 files): Merge branch 'rurban/pmc2c-unused-gh836' Fixed conflict: \tconfig/auto/warnings.pm fixes already in master |
|||
| rrot: ac0b135 | rurban++ | lib/Parrot/Pmc2c/Method.pm: [GH #836] Catch also Scheduler.destroy and empty Undef bodies |
|||
| rrot: ce2c408 | rurban++ | lib/Parrot/Pmc2c/Method.pm: Merge branch 'rurban/pmc2c-unused-gh836' |
|||
| ttbot | Parrot ac0b1355 i386-linux-thread-multi make error tt.taptinder.org/cmdinfo/97812 | 18:49 | |
|
19:05
travis-ci joined
|
|||
| travis-ci | [travis-ci] parrot/parrot#593 (master - ce2c408 : Reini Urban): The build was fixed. | 19:05 | |
| [travis-ci] Change view : github.com/parrot/parrot/compare/0...2c408d3627 | |||
| [travis-ci] Build details : travis-ci.org/parrot/parrot/builds/2561119 | |||
|
19:05
travis-ci left
|
|||
| rurban | #ps in 5 min | 19:25 | |
| dalek | Heuristic branch merge: pushed 62 commits to parrot/gh838_cmdline_cleanup by rurban | 19:29 | |
| rrot: 19183d0 | dukeleto++ | / (2 files): Remove mentions of -d/-O/-v from --help output and docs/running.pod #838 |
|||
| rrot: bf60df5 | dukeleto++ | / (3 files): Remove more references to options that no longer exist #838 |
|||
| rrot: eaa1466 | rurban++ | / (41 files): Merge remote-tracking branch 'origin/master' into gh838_cmdline_cleanup |
19:30 | ||
| rrot: 4879083 | rurban++ | / (3 files): [GH #838] bootstrap-prt0, ChangeLog Remind to call bootstrap-prt0 in Makefile when prt0.winxed is changed. Called bootstrap-prt0 Added ChangeLog line, and whitespace fixes |
|||
| ttbot | Parrot 0525faa8 MSWin32-x86-multi-thread make error tt.taptinder.org/cmdinfo/97925 | 19:33 | |
| dalek | rrot: a5bb6bf | rurban++ | / (2 files): [codingstd] hard tabs |
19:54 | |
| rrot: 14fa643 | rurban++ | docs/binaries/parrot.pod: [codingstd] podcheck parrot.pod Fixed various non-conformant pod: empty =item, =over for bodies |
|||
| rurban | Oh! travis smoked the threads branch actually. Nice. Exactly as I wanted. | 19:55 | |
| travis-ci.org/#!/parrot/parrot/builds/2560855 | |||
| make cover actually does something useful. src/list.c and src/platform/generic/error.c src/extend_vtable.c src/disassemble.c have 0% | 20:03 | ||
| ttbot | Parrot 4879083e MSWin32-x86-multi-thread make error tt.taptinder.org/cmdinfo/97986 | 20:14 | |
| rurban | Question: Why is ImageIOThaw.set_pointer(void*) and not set_pointer(PackFile_ConstTable *value)? | 20:16 | |
|
20:18
tuxit joined
|
|||
| sorear | erm, is the latter even an option? | 20:24 | |
| isn't it forced to void* by the VTABLE spec for set_pointer? | 20:25 | ||
| rurban | set_pointer is always void*? This would make sense, yes | 20:27 | |
| sorear | I think so, but don't trust me. | 20:31 | |
| benabik | I think so. | 20:32 | |
| set_pointer is not a particularly safe operation. | |||
| ttbot | Parrot 14fa643c MSWin32-x86-multi-thread make error tt.taptinder.org/cmdinfo/97998 | 20:55 | |
|
21:20
perlite joined
22:21
whiteknight joined
|
|||
| whiteknight | good evening, #parrot | 22:24 | |
|
22:47
benabik joined
23:07
jevin joined
23:18
rurban_mobile joined
23:41
mdupont joined
23:52
rurban_mobile joined
|
|||