Camelia, the Perl 6 bug

IRC log for #parrot, 2009-11-30

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:02 dalek partcl-nqp: 1e4db42 | pmichaud++ | src/PmTcl/Commands.pm:
00:02 dalek partcl-nqp: Simple version of after(); t/cmd_after.t passes.
00:02 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/1e4db4244494e453e25b148a5e9847c652323de4
00:20 japhb pmichaud, the STATUS files for NQP-rx seems very out of date.  Should it just be deleted entirely?
00:26 Whiteknight you know, if we had aggregate literal syntax in PIR, NQP could just translate to that
00:29 japhb Whiteknight, you get right on that.  ;-)
00:30 * japhb busily trying to grok NQP-rx's workings
00:33 japhb pmichaud, also, there are a fair few variables set at the top of NQP-rx's Makefile.in that appear unnecessary (such as the ones for C compiling and linking).  Can those be removed, or are they for a non-obvious and/or future purpose?
00:48 theory joined #parrot
00:54 pmichaud any unnecessary variables in the top of the makefile can be used.  Some of them are left over from when pge/nqp (no -rx) was being used to build nqp-rx
00:56 japhb "can be used"?
00:56 japhb Do you mean "removed"?
00:57 abqar joined #parrot
00:58 japhb pmichaud, also: I'm trying to follow the build process.  What is the purpose of the 3-stage bootstrap?  The only differences I can find in the PIR between stage1 and stage2 are subids and a seemingly trivial reordering of "!cursor_caparray" arguments.
01:00 japhb pmichaud, Is the 3-stage bootstrap another holdover of building from PGE/old NQP?
01:42 bacek joined #parrot
02:04 Zak joined #parrot
02:12 jan joined #parrot
02:14 zak_ joined #parrot
03:10 cognominal_ joined #parrot
03:15 * Coke
03:18 dukeleto 'ello
03:38 davidfetter joined #parrot
03:38 JimmyZ joined #parrot
04:12 Coke pmichaud: ping
04:13 nopaste "Coke" at 72.228.52.192 pasted "this complains that it can't find the lexical '$body'" (16 lines) at http://nopaste.snit.ch/18900
04:15 Coke (oddly, it complains after passing every test...)
04:19 Coke incoming.
04:19 purl duck!
04:20 dalek partcl-nqp: 0241721 | (Will Coleda)++ | :
04:20 dalek partcl-nqp: Merge branch 'master' of github.com:partcl/partcl-nqp
04:20 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/0241721b82957f799846716971b09a33d5bc2e4d
04:20 dalek partcl-nqp: 8dbbda6 | (Will Coleda)++ | src/PmTcl/Commands.pm:
04:20 dalek partcl-nqp: Add basic args handling to [regexp]
04:20 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/8dbbda604b3d95e9643df4e75de7d51f07c90a50
04:20 dalek partcl-nqp: d844c93 | (Will Coleda)++ |  (4 files):
04:20 dalek partcl-nqp: Move tcllist into a subdir.
04:20 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/d844c93707a78b2cd3d1e5eef2b0175967898bb3
04:20 dalek partcl-nqp: e98a757 | (Will Coleda)++ | :
04:20 dalek partcl-nqp: Merge branch 'master' of github.com:partcl/partcl-nqp
04:20 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/e98a757cc282539d930a6e7ddc9d68737bc87308
04:20 dalek partcl-nqp: 11e9695 | (Will Coleda)++ |  (3 files):
04:20 dalek partcl-nqp: Add a string HLL class that has a more accurate get_bool vtable.
04:20 dalek partcl-nqp: (reclaims a few individual tests)
04:20 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/11e9695cae910d81170e8fcc99417f297fbe4d72
04:20 dalek partcl-nqp: 9dc018b | (Will Coleda)++ | src/PmTcl/Commands.pm:
04:20 dalek partcl-nqp: add args handling to [for]
04:20 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/9dc018bb1eac41d00478d31372b2e52a84a649d7
04:20 dalek partcl-nqp: 15c4413 | (Will Coleda)++ | TODO:
04:20 dalek partcl-nqp: update task list
04:20 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/15c441393e4790e8714b8eab7e91c6644513fcb1
04:26 dalek partcl-nqp: 637f097 | (Will Coleda)++ | build/Makefile.in:
04:26 dalek partcl-nqp: pmichaud++ made this test pass
04:26 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/637f097128df854f4b9d3b23827606f3c3aea59b
04:38 vad joined #parrot
04:46 nopaste joined #parrot
04:56 patspam joined #parrot
05:18 nopaste joined #parrot
06:42 dalek tracwiki: v1 | chromatic++ | chromaticTasks
06:42 dalek tracwiki: initial creation
06:42 dalek tracwiki: https://trac.parrot.org/parrot/wiki/ch​romaticTasks?version=1&action=diff
07:08 uniejo joined #parrot
07:30 cotto joined #parrot
08:04 iblechbot joined #parrot
08:18 bacek joined #parrot
08:43 mikehh joined #parrot
09:00 fperrad joined #parrot
09:01 fperrad_ joined #parrot
09:16 riffraff joined #parrot
09:27 fperrad_ joined #parrot
09:43 mberends joined #parrot
09:46 payload joined #parrot
09:47 riffraff joined #parrot
09:57 abqar_ joined #parrot
10:05 fperrad_ joined #parrot
10:06 JimmyZ joined #parrot
10:09 mikehh joined #parrot
10:38 mikehh joined #parrot
10:49 mikehh joined #parrot
10:55 mikehh joined #parrot
11:04 dalek parrot: r42823 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
11:04 dalek parrot: [distutils] some refactors and doc
11:04 dalek parrot: review: https://trac.parrot.org/parrot/changeset/42823/
11:11 mikehh joined #parrot
11:17 mikehh joined #parrot
11:23 bacek joined #parrot
12:15 patspam joined #parrot
12:26 purl joined #parrot
12:29 mikehh joined #parrot
12:33 payload joined #parrot
12:41 mikehh joined #parrot
12:46 plobsing joined #parrot
12:47 mikehh joined #parrot
12:50 dalek parrot: r42824 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
12:50 dalek parrot: [distutils] handles options from command line
12:50 dalek parrot: review: https://trac.parrot.org/parrot/changeset/42824/
12:53 mikehh joined #parrot
12:57 mikehh joined #parrot
13:01 nopaste joined #parrot
13:03 zak_ joined #parrot
13:05 payload joined #parrot
13:06 mikehh joined #parrot
13:14 mikehh joined #parrot
13:19 fperrad joined #parrot
13:20 mikehh joined #parrot
13:21 JimmyZ joined #parrot
13:21 bluescreen joined #parrot
13:28 cognominal_ joined #parrot
13:29 zak_ joined #parrot
13:34 payload joined #parrot
13:38 clinton joined #parrot
13:38 mberends_ joined #parrot
13:38 mberends_ left #parrot
13:43 clinton joined #parrot
13:49 clinton joined #parrot
13:53 nopaste joined #parrot
13:57 zak_ joined #parrot
14:01 * Coke yawns
14:02 Coke hachi++
14:05 Coke <the critic>HACHI MACHI!</critic>
14:05 iblechbot joined #parrot
14:07 cognominal_ joined #parrot
14:17 zak_ joined #parrot
14:23 Coke msg fperrad I hestitate to apply that patch to partcl-nqp, as the PIR seems a little more convoluted to me. (probably just because it's different) - is there a goal here other than reducing the dependency on perl5?
14:23 purl Message for fperrad stored.
14:24 Coke (If anyone is interesting in hacking on partcl-nqp as a test for working with nqp-rx, btw, have at it. Happy to answer any questions. =-)
14:28 davidfetter joined #parrot
14:31 Coke how is ".namespace" written in NQP ?
14:31 patspam joined #parrot
14:38 nopaste joined #parrot
14:38 Coke hurm. "module { ... }" seems close.
14:38 Coke_ joined #parrot
14:40 pmichaud good morning, #parrot
14:42 Coke hio
14:42 particle howdy
14:47 davidfetter oh hai
14:49 ruoso joined #parrot
14:51 Coke pmichaud: moved string_trim into _tcl::string_trim
14:53 pmichaud Coke: excellent.
14:55 dalek partcl-nqp: efe5134 | (Will Coleda)++ | TODO:
14:55 dalek partcl-nqp: Add to the pile
14:55 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/efe51341918b1f7b51710b7f4e523f188b8329ce
14:55 dalek partcl-nqp: c916eb0 | (Will Coleda)++ | src/PmTcl/Commands.pm:
14:55 dalek partcl-nqp: Move string_trim into our ``private'' namespace.
14:55 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/c916eb03ec763005876e65f72542cb829f008057
14:58 Coke is there an "int()" in nqp?
15:00 pmichaud Coke: closest is   pir::set__IP($something)
15:03 Coke pmichaud: close enough. =-)
15:03 * Coke works on his secret evil plan.
15:05 particle what's the difference between pir::set__IP(...) and int() ?
15:06 particle seems like it could be pretty easy to support native types in nqp...
15:08 Coke particle: I think only the amount of sugar.
15:09 pmichaud I do plan to support the native types in nqp
15:09 pmichaud but it's not "easy"
15:10 pmichaud because lexical symbols in Parrot can only be PMCs
15:10 he Aahhh...  I've found the reason t/pmc/threads_14.pir fails on NetBSD.
15:10 pmichaud which means PAST needs a whole new notion of "lexical" in order to be able to support symbol<->native type mapping
15:10 pmichaud and something like:
15:11 pmichaud my int $a;  if 1 { my $b := $a + 2; }
15:11 pmichaud cannot work at all
15:11 he Parrot in this test ends up calling dlclose() twice with the same handle, and NetBSD's ld.so_elf has a check for whether the given handle is "valid", which it isn't in the second case.
15:11 pmichaud (without a *lot* of analysis)
15:11 he I've not found out how to fix it, though...
15:12 Coke he++
15:13 pmichaud int() would be a function call to the "int" function.... and NQP doesn't want to provide a library of builtins beyond whatever Parrot provides
15:13 he Fodder for trac, perhaps?
15:16 nopaste joined #parrot
15:17 Coke pmichaud: if you can figure out the "$body" error in the TODO, that'll reclaim two test files.
15:17 Coke (looks like those 2 tests run to completion, and then dump the error message.)
15:19 dalek partcl-nqp: 96d0a16 | (Will Coleda)++ |  (2 files):
15:19 dalek partcl-nqp: Add mostly complete [time]
15:19 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/96d0a16ffc4a4043d23ee3c8851724e14bce13bd
15:19 Coke also, I'm guessing If we do anything serious with [format], it'll probably end up as a 3rd grammar. Would you agree?
15:21 Coke (partcl's version just dispatches to sprintf.)
15:22 treed_ joined #parrot
15:22 dngor_ joined #parrot
15:22 ingyfoo joined #parrot
15:22 Tene_ joined #parrot
15:23 GeJ_ joined #parrot
15:23 TimToady_ joined #parrot
15:23 confound_ joined #parrot
15:23 frodwith_ joined #parrot
15:23 estrabd_ joined #parrot
15:23 viklund_ joined #parrot
15:24 he OK, trac ticket #1340 filed at https://trac.parrot.org/parrot/ticket/1340
15:25 Coke he++
15:26 dalek TT #1340 created by heidnes++: Parrot may end up calling dlclose() twice with the same handle, ref. ...
15:26 Coke pmichaud: I also wonder if http://www.tcl.tk/man/tcl8.​5/TclCmd/namespace.htm#M34 might be a useful way to implement core commands like array/info/string/namespace
15:27 Coke (which are themselves ensembles) - in partcl, I just did a giant switch statement.
15:27 leto joined #parrot
15:28 Util joined #parrot
15:28 confound joined #parrot
15:28 ruoso joined #parrot
15:28 cognominal_ joined #parrot
15:28 slavorg joined #parrot
15:29 workbench joined #parrot
15:33 Infinoid joined #parrot
15:33 dcolish joined #parrot
15:33 ascent joined #parrot
15:33 dukeleto_ joined #parrot
15:33 pmichaud_ joined #parrot
15:33 davidfetter joined #parrot
15:33 clinton joined #parrot
15:33 bluescreen joined #parrot
15:33 riffraff joined #parrot
15:33 mberends joined #parrot
15:33 cotto joined #parrot
15:33 bacek_at_work joined #parrot
15:33 Hunger joined #parrot
15:33 s1n joined #parrot
15:33 redbrain joined #parrot
15:33 xenoterracide joined #parrot
15:33 kthakore joined #parrot
15:33 KatrinaTheLamia joined #parrot
15:33 szabgab joined #parrot
15:33 gaz joined #parrot
15:33 zostay joined #parrot
15:33 particle joined #parrot
15:35 jan joined #parrot
15:35 he_ joined #parrot
15:35 cxreg joined #parrot
15:39 slavorg joined #parrot
15:41 japhb joined #parrot
15:41 Khisanth joined #parrot
15:42 vad joined #parrot
15:43 payload joined #parrot
15:43 sri joined #parrot
15:44 fperrad joined #parrot
15:45 Whiteknight joined #parrot
15:47 jjore joined #parrot
15:49 PacoLinux joined #parrot
15:53 Coke msg pmichaud huh! http://sourceforge.net/tracker/?func=detail&amp​;atid=110894&amp;aid=2901433&amp;group_id=10894 has been resolved.
15:53 purl Message for pmichaud stored.
15:55 Coke msg pmichaud diff of Tcl.n docs at: http://tcl.cvs.sourceforge.net/viewvc/tcl/tcl/d​oc/Tcl.n?r1=1.18&amp;r2=1.18.2.1&amp;sortby=rev
15:55 purl Message for pmichaud stored.
15:56 Psyche^ joined #parrot
15:57 Coke pmichaud++ # getting doc updates back into tcl mainline.
15:58 he Further info on TT#1340: this appears to be clone()'s fault: it clones a library, but the two instances share the dlhandle value.
16:01 Coke PerlJam: hey, partcl misses you!
16:02 PerlJam Coke: it looks lke partcl has been doing just fine without me :)
16:03 * Coke wonders how to split up Commands.pm
16:03 PerlJam is it becoming unwieldy?
16:03 * PerlJam looks
16:03 Coke it's going to be once we add things like array/string/namespace
16:04 Coke so I think we want some individual files, and perhaps a bucket for the rest.
16:04 PerlJam sounds like naturalish divisions to me
16:04 PerlJam (put all array commands together, all string commands together, etc)
16:06 PerlJam I've noticed that some commands have an extensive subcommand structure.  Those should probably get their own files.  (I don't know if that neatly follows the divisions I just mentioned or not though)
16:06 lucian joined #parrot
16:09 * Coke is thinking Commands/{string,namespace}.pm - pick a name for the leftovers. =-)
16:11 particle does the organization of the tcl tests offer any suggestions for how the commands should be divided?
16:12 Coke I don't particularly care to go for a test file <-> implementation file 1:1
16:12 * Coke picks "main"
16:16 dalek partcl-nqp: 816ba49 | (Will Coleda)++ |  (4 files):
16:16 dalek partcl-nqp: Move Commands.pm -> commands/main.pm in preparation for splitting.
16:16 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/816ba49bdeb45fbd47187ef73208602163fff891
16:24 Whiteknight how do I do parallel testing again?
16:24 Whiteknight I thought it was TEST_JOBS=5, but that doesnt seem to be right
16:24 Coke make TEST_JOBS=3 test
16:24 Coke presuming your harness supports it.
16:25 Coke is this parrot?
16:25 Whiteknight yes, parrot
16:25 Whiteknight that incantation doesn't seem to do it. Are there perl dependencies for that?
16:25 Coke yes.
16:26 Coke look at line 103 of t/harness  or so
16:27 Coke (you need TAP::Harness)
16:28 particle you need test::harness 3.x
16:28 dalek partcl-nqp: 79d9ee8 | (Will Coleda)++ |  (4 files):
16:28 dalek partcl-nqp: break out [string] in anticipation of implementing the rest of the ensemble command.
16:28 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/79d9ee81f813b2d290f9a329052b3d89c263ac67
16:28 Whiteknight I had to do a complete reinstall of my system, and now I don't have any of the things or settings that I used to hae
16:29 szbalint yeah, I run into the test::harness weirdness too
16:29 Coke `cpan TAP::Harness` should fixit.
16:29 szbalint earlier versions don't support parallel tests
16:32 * Coke needs to delete svn-bisect
16:36 Coke er, svnbinarysearch, I mean.
16:41 * Coke sees this in the .t for svn-bisect: BEGIN { $tests += 7; };
16:41 Coke (sneaky++)
16:41 Infinoid Coke: I stole that syntax from jq++.  It's more manageable to keep the counts with the tests
16:43 Coke Infinoid: any chance I can convince *you* to write svn-bisect run so I don't have to? =-)
16:44 Infinoid I can take a look at how easy it would be to do, at the very least.  I'll get back to you.
16:44 Coke Infinoid++
16:45 Coke I look forward to killing my only CPAN module when you do. =-)
16:47 Infinoid Coke: Your module looks for differences in stdout, or differences in exit code?
16:48 Infinoid "git bisect run" looks only at the exit code, it assumes an exit code of 0 means "before the change" and non-zero means "after"
16:49 Infinoid I've intentionally been staying away from saying "before" means "non-buggy" and "after" means "buggy", because you might be looking for things other than bugs.  But this interface kind of forces the issue
16:49 Coke Infinoid: exit code.
16:49 purl hmmm... exit code is not reliable.  It returns 0 if the file is invalid.
16:49 Coke this doesn't force the issue, you just have to be careful when documenting.
16:50 Infinoid Yeah.  It won't stop me from implementing it anyway
16:50 Coke 0 == "before" behavior, 1-124,126-127 == "after", 125==skip. methinks.
16:50 Coke Infinoid++ # can't stop the signal.
16:52 Coke If you like, I can whip up a sample script that shows the exit behavior. (say, using 125 when the build fails.)
16:52 Infinoid nah, it's pretty clear how the git version works
16:53 Coke whee!
16:54 Infinoid The tricky bit is how I handle metadata inside of svn-bisect while the run method reinvokes other methods... nothing to do with the user interface
16:56 Infinoid I do still want to implement "undo" one of these days.  But it would be nice if it could undo the whole run, not just the last action it took
16:59 nopaste joined #parrot
17:05 Infinoid parrot++ # having another thousand revisions every time I do an update
17:10 Coke Infinoid: you want a google wave invite?
17:10 Infinoid Coke: I'm already on it, thanks.  And I have invites too, if any parrot committers want one
17:11 Coke ditto here.
17:11 * Coke ponders sending an email to the list.
17:11 * Coke does so.
17:11 theory joined #parrot
17:11 moritz so does wave live up to the hype?
17:12 Infinoid Wave does what they say it does.  Whether it's useful or not is really up to the users.
17:13 Infinoid The with:public posts are even worse than usenet.  I haven't had the opportunity to do anything serious with it yet.
17:13 Coke Without people who want to collaborate on something that I want to work on, I have found it very noisy.
17:17 cotto_w0rk joined #parrot
17:18 nopaste joined #parrot
17:23 dalek parrot: r42825 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
17:23 dalek parrot: [distutils] add a step 'manifest'
17:23 dalek parrot: review: https://trac.parrot.org/parrot/changeset/42825/
17:27 iblechbot joined #parrot
17:32 Coke does p6 (nqp) require {}'s in "if <cond> { <block> }" ?
17:47 DrForr joined #parrot
17:47 moritz yes
17:48 Coke moritz++ # danke.
17:57 kurahaupo joined #parrot
18:00 payload joined #parrot
18:03 mikehh joined #parrot
18:05 chromatic joined #parrot
18:05 theory seen dukeleto
18:05 purl dukeleto was last seen on #parrot 14 hours, 46 minutes and 46 seconds ago, saying: 'ello
18:05 theory seen dukeleto_
18:05 purl dukeleto_ was last seen on #pdx.pm 32 days, 2 hours, 11 minutes and 22 seconds ago, saying: 'ello  [Oct 29 15:52:59 2009]
18:05 Coke I see a pattern.
18:06 Coke seen dukeleto__
18:06 purl I haven't seen 'dukeleto__', Coke
18:06 Coke curses.
18:06 theory I see dead people.
18:06 dalek partcl-nqp: 552913c | (Will Coleda)++ | src/PmTcl/commands/string.pm:
18:06 dalek partcl-nqp: add basic args check to [string]
18:07 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/552913ca2a0fa7603d774a8fc5d1d313719f1703
18:07 dalek partcl-nqp: 485a993 | (Will Coleda)++ | src/PmTcl/commands/string.pm:
18:07 dalek partcl-nqp: add [string bytelength];
18:07 dalek partcl-nqp: alphabetize subcommands
18:07 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/485a9938c1b68001f3b7d5d94b05aba2898771d3
18:07 dalek partcl-nqp: bf6a58c | (Will Coleda)++ | src/PmTcl/commands/string.pm:
18:07 dalek partcl-nqp: accept all valid, fully-qualified subcommand names for [string]
18:07 dalek partcl-nqp: even if they do nothing; error on invalid subcommands, return '' for now
18:07 dalek partcl-nqp: on the stubs.
18:07 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/bf6a58c93a682d19e54d9d1b611cb823387ae139
18:07 dalek partcl-nqp: 04106f2 | (Will Coleda)++ | src/PmTcl/commands/string.pm:
18:07 dalek partcl-nqp: add [string first]
18:07 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/04106f2f95fbfd4c80a00bf7ab45e4eb98d08b21
18:07 Andy joined #parrot
18:07 japhb msg pmichaud I'm trying to follow the build process.  What is the purpose of the 3-stage bootstrap?  The only differences I can find in the PIR between stage1 and stage2 are subids and a seemingly trivial reordering of "!cursor_caparray" arguments.  Is the 3-stage bootstrap another holdover of building from PGE/old NQP?
18:07 purl Message for pmichaud stored.
18:08 theory msg dukeleto I'm back from vacation. Ping me when you get a chance so we can work out your pgTAP issues.
18:08 purl Message for dukeleto stored.
18:13 Coke Jay Scherrer?
18:17 Tene_ msg pmichaud How did lazy rakudo rethink go on sunday?
18:17 purl Message for pmichaud stored.
18:24 Coke koleda is a sun god!?
18:24 Coke huh. Who knew.
18:24 purl knew is offensive to some people?
18:24 Coke no, knew is <reply>
18:24 purl okay, Coke.
18:26 pmichaud_ japhb: (3-stage bootstrap).... stage0 is the last known working compiler
18:26 pmichaud_ it's used to build stage1
18:27 Coke koleda?
18:27 purl it has been said that koleda is a sun god!?
18:27 pmichaud_ stage1 is a fresh compiler built from the updated sources.  however, since stage0 was used to build stage1, we can't be sure that the stage1 PIR source is completely correct w.r.t. the latest source
18:27 Coke no, koleda is probably koliada on wikipedia.
18:27 purl okay, Coke.
18:27 pmichaud_ stage2 is the result of using the latest nqp source to build itself
18:27 pmichaud_ Tene_: still working on it a bit in my head.  I plan to be implementing tonight.
18:28 Tene_ I'm excited to see it.
18:28 hercynium joined #parrot
18:31 japhb pmichaud_, OK, I kinda guessed something along those lines.  Hmmm, I guess the fact that stage1 and stage2 were identical except subids and the like is just because the changes since the last stage0 compile didn't affect anything NQP-rx uses in its own source.
18:31 pmichaud_ correct.
18:40 DrForr joined #parrot
18:40 mikehh All tests PASS (pre/post-config, smoke (#30291), fulltest) at r42825 - Ubuntu 9.10 i386 (g++ with --optimize)
18:41 DrForr joined #parrot
18:42 dalek partcl-nqp: d9e1e35 | (Will Coleda)++ | src/PmTcl/commands/main.pm:
18:42 dalek partcl-nqp: Insure string is a string.
18:42 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/d9e1e356646805f7abd930559fa66a26c300e05d
18:48 dalek parrot: r42826 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
18:48 dalek parrot: [distutils] add steps 'sdist', 'sdist_gztar' & 'sdist_zip'
18:48 dalek parrot: review: https://trac.parrot.org/parrot/changeset/42826/
18:49 Infinoid Coke: Would you mind trying http://github.com/Infinoid/svn-bisect/commi​t/8601c020fe04c3612329f211faaeec835e11a6ab ?
18:50 Coke nopaste.snitch.ch is dead.
18:51 Coke er, snit.ch
18:51 Coke Infinoid: is it lastest on master?
18:51 Infinoid yeah
18:52 * Coke is distracted by ST:TOS:Season 1:BR
18:53 Infinoid Terrible situation.
18:53 Coke thankfully I don't own it yet. :|
18:54 Coke Infinoid: run is missing from the list of subcommands.
18:54 Infinoid Well, I got 3/4 of the places I needed to update documentation, at least
18:55 dalek parrot: r42827 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
18:55 dalek parrot: [distutils] fix compatibily Windows & *nix
18:55 dalek parrot: review: https://trac.parrot.org/parrot/changeset/42827/
18:57 * Infinoid notices svn-bisect doesn't exactly scream along on the parrot repo... is it slower than usual, or am I just spoiled?
19:01 * moritz recommends git bisect on a git-svn clone of the parrot repo
19:01 Coke moritz: given that he's trying to test svn bisect...
19:01 Infinoid moritz: That's what I was comparing "svn-bisect run" against (I just implemented that)
19:02 * moritz out of context
19:02 Infinoid On 18000 revisions, git took 11 seconds, svn-bisect was off by at least 2 orders of magnitude
19:02 Infinoid moritz: In general, I'd agree.  In fact, I feel a little bit guilty about publishing a crutch for svn users :)
19:02 Coke Infinoid: I am having trouble getting out of an old bisect that was running.
19:03 Infinoid svn-bisect reset?
19:04 Coke yes. that's what's giving me trouble. =-)
19:04 Infinoid It doesn't work?  Or what?
19:05 Coke it either seems to hang (is it resetting to the original revision number? evil) or gives me guff about unversioned things in the way. I've been deleting the unversioned things and trying the reset again.
19:05 slavorgn joined #parrot
19:05 Coke be nice if there was a "nuke" option. (don't try to put anything back...)
19:06 Infinoid Coke: Yes, it resets to the original version number.  Same as git bisect reset does...
19:06 Infinoid Reset is the one part of the git bisect interface that never really seemed intuitive to me
19:07 Coke so is there a way to drop the bisect in progress without doing the svn update to the old rev #?
19:08 Infinoid heh.  rm .svn/bisect.yaml
19:08 Coke hey, that directory is not for you! =-)
19:08 Infinoid It's better than what svk does.
19:09 Infinoid I could be convinced to make reset less invasive.
19:10 Infinoid At the very least, it should probably emit a message about what it's taking so long for
19:11 Infinoid I'd actually prefer some way to make reset unnecessary.  I often forget to do it, and come back to a checkout a couple of months later and discover that I never cleaned up after the last bisect
19:16 Coke sounds familiar. =-)
19:17 joeri joined #parrot
19:22 Infinoid Coke: Thoughts on nuking by default and adding a --back to get the old (annoying) behavior?
19:28 cotto_work pmichaud, what's your estimate on how much Parrot's lack of a proper lvalue model slows down the code that nqp generates?
19:37 Coke Infinoid: I think having it do nothing is probably a better default than doing what it's doing, so +1
19:37 Coke (perhaps add a warning message to say "we're just leaving you at revision X, you're on your own.)
19:38 dalek parrot: r42828 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
19:38 dalek parrot: [distutils] minor refactor
19:38 dalek parrot: review: https://trac.parrot.org/parrot/changeset/42828/
19:41 mikehh joined #parrot
19:42 Coke ok. 30 "rm -rf path/to/unversioned && svn up" later...
19:43 Coke (still not done)
19:44 simcop2387 joined #parrot
19:49 mikehh joined #parrot
19:55 Coke yay, done.
19:55 Coke now, what was I supposed to test? =-)
19:55 Infinoid svn-bisect run
19:58 Infinoid cool.
19:58 Infinoid infinoid@chirp parrot-svn % svn-bisect reset --back
19:58 Infinoid Resetting your checkout back to r42000.
19:59 Infinoid ...
19:59 Infinoid infinoid@chirp parrot-svn % svn-bisect reset
19:59 Infinoid Cleaned up.  Your checkout is still at rev r42575.
20:00 Coke Infinoid++
20:01 Infinoid I'll push that in a moment
20:01 Coke is there a variant of svn up that actually removes deleted things from working copies?
20:01 Coke (because I suspect this is going to kill me again as I try to find a range to test.)
20:04 Infinoid sadly, I've gotten used to typing "svn st | cut -c 8- | xargs rm -f; svn update"
20:04 Infinoid That doesn't mean the facility doesn't exist, I just haven't been able to find one.
20:07 dukeleto_ 'ello
20:07 Coke i sense a pattern.
20:08 Coke Infinoid: ARGH. stuck in svn up hell again. moment.
20:08 bacek joined #parrot
20:09 Infinoid Coke: The xargs rm trick works.  If it doesn't, add -rf.
20:09 jsut_ joined #parrot
20:10 Infinoid (it will also wipe out any local modifications, of course.)
20:10 Coke svn--
20:15 Coke Infinoid: it might be nice if the harness said whether or not the run script was good/bad/before/after
20:15 Coke (no clue what git does there)
20:16 Infinoid git assumes 0 means good and 1 means bad.  before and after are svn-bisect specific aliases
20:16 Coke just looking for some diagnostic output.
20:16 Infinoid Oh, in the status lines.  Makes sense
20:16 Coke looks good for the bogus test I just ran...
20:17 Coke let me find something that I can actually verify changed. =-)
20:17 Infinoid My test was just checking for a file that was added in some rev
20:18 lucian joined #parrot
20:18 Infinoid Anything more complicated will likely have the same kind of cleanup problems you were seeing
20:18 dukeleto exit 125 means "invalid" in the git-bisect world
20:18 Infinoid Yep, it means "skip" here too
20:28 Coke Infinoid++ # seems to work.
20:29 Coke didn't test the 125 exit code, but i checked for when something was added to a file, and it worked.
20:30 Infinoid cool.  I've just got 2 more cases to clean up, and then I'll call this done and upload 1.0
20:30 Coke Infinoid++
20:30 Coke I think I'm just going to delete my wholesale and put up a blog post about it.
20:30 Coke *mine
20:30 Coke pointing them at yours.
20:31 Infinoid I'm not sure more than 2 people on the planet care about mine :)
20:32 japhb "We all care about you, Infinoid.  That's why we've strapped you in that straight jacket and called the friendly men in the white van."
20:32 Infinoid (hence spamming #parrot instead of making an empty and pathetic #svn-bisect... thanks for your patience all)
20:33 Coke (care about yours) well, you can have the one person who cares about mine, giving you a 50% increase.
20:35 Infinoid on to slightly more tangentally parrot-related things... does anyone have a dalek wishlist?
20:35 moritz tracking branches
20:36 Coke who admins blogs.perl.org?
20:36 moritz for gibhub projects
20:36 moritz Coke: davorg does, I think
20:36 Coke seen davorg?
20:36 purl davorg was last seen on #dbix-class 11 days, 8 hours, 9 minutes and 34 seconds ago, saying: Hmm... Also - "heavily influenced by ActiveRecord::Migrations"  [Nov 19 12:25:32 2009]
20:36 * moritz offered davorg hosting on timtowtdi.org
20:38 japhb Infinoid, IWBNI if dalek did better on gitorious.  It's cutting off commit messages (probably because the source RSS is, but it should be possible to follow the commit link and pull down the full message).  Likewise no information about which files were changed.  There's something else that the other parsers display that the gitorious parser doesn't, but I've forgotten it offhand.
20:38 bluescreen joined #parrot
20:38 Infinoid japhb: Got a rev that it fails to parse?
20:39 Infinoid moritz: All branches, or specified (somehow) ones?
20:39 Coke I would expect specified. (couldn't you get this effect by listing the branches on the Languages page on teh wiki?)
20:39 moritz Infinoid: some specified ones
20:39 japhb Infinoid, I don't think it completely fails, just is LTA.  And pretty much anything from the parrot-plumage feed with a commit message over 40 or 50 chars.
20:40 Infinoid japhb: Does the rss feed cut it off with a "..." or something silly like that?
20:41 japhb Infinoid, yeah, I think so.
20:41 Infinoid Ok, I can work on that
20:41 Infinoid moritz: And branches too.  I know we already started on that stuff (for rakudo-ng I think?)
20:42 moritz we did
20:42 japhb Since new commits are relatively rare, as long as you only followed the commit links when parsing a new commit, you shouldn't hit a performance issue, I would think.  And dalek is already structured to optimize that way, of course.  Infinoid++ for that.
20:43 Infinoid I already have to hit the commit links for some parsers, specifically the ones that don't include a file listing
20:43 Infinoid That's not a problem
20:43 * japhb is trying to imagine what made the gitorious developers think cutting off commit messages was a good idea.
20:43 Infinoid They probably wanted it to fit on their iphone screen
20:44 * Coke cannot login to pause.
20:44 Infinoid There seems to be far too much presentation-related stuff in RSS/ATOM feeds in general
20:44 japhb Infinoid, oh cool, then you'll already be halfway to the next problem -- getting the file listing info.  :-)
20:44 japhb Infinoid, YES, QUITE.
20:47 Infinoid dalek's job would be 3 times easier if it didn't have to parse out escaped html trash :)
20:48 Infinoid Anyways, thanks guys, I'll take a look later today
20:49 japhb I bet
20:54 GeJ joined #parrot
20:55 GeJ Good morning everyone
20:55 japhb o/
20:55 cotto_work hi GeJ
21:01 dalek winxed: r233 | julian.notfound++ | trunk/winxedst1.winxed:
21:01 dalek winxed: operator -- in stage 1
21:01 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=233
21:06 dalek winxed: r234 | julian.notfound++ | trunk/Makefile:
21:06 dalek winxed: make target to build stage 1
21:06 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=234
21:08 confound joined #parrot
21:10 davidfetter jhorwitz?
21:10 purl jhorwitz is "mod_parrot-guy" or mailto:jeff@smashing.org
21:12 payload joined #parrot
21:33 japhb I'm thinking of making Plumage automatically aware of the installed parrot changing, so that it knows to upgrade all installed projects.  I'd also like to make Plumage work with multiple parrots on the same system (for people that want to have last stable release, last release, HEAD, and a branch or two all installed at once).  Any ideas on how to determine a unique key for a parrot build?  Best I've got so far is some variant of `parrot_c
21:33 japhb onfig --dump | sort | sha1sum`.  Any other suggestions?
21:33 nopaste joined #parrot
21:33 TonyC joined #parrot
21:40 cotto_work japhb, why not add a guid to parrot_config as part of the build process?
21:41 japhb cotto_work, that's not a bad idea.  Certainly more efficient to be checking all the time.
21:41 bacek_ joined #parrot
21:42 cotto_work hail, bacek
21:42 bacek Good morning
21:42 bacek aloha cotto_work
21:42 bacek seen notfound
21:42 purl notfound was last seen on purl 22 hours, 57 minutes and 0 seconds ago, saying: <private message>
21:43 japhb Anyone around running Mac OS X?
21:44 cotto_work japhb, why?
21:44 cotto_work I have a seldom-used system, but I haven't build Parrot on it yet.
21:44 japhb cotto_work, Wondering if OS X uses UUIDs, GUIDs, or something entirely else due to rampant NIH.
21:46 cotto_work no idea
21:50 Coke japhb: I have os x.
21:50 Coke but I can't answer that question.
21:50 japhb heh
21:51 dalek parrot: r42829 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
21:51 dalek parrot: [distutils] add a step bdist, and win32-inno-installer becomes bdist_wininst
21:51 dalek parrot: review: https://trac.parrot.org/parrot/changeset/42829/
21:51 dalek parrot: r42830 | bacek++ | trunk/t/pmc/complex.t:
21:51 dalek parrot: Rewrite skipped test for #TT1318 using C<todo>.
21:51 dalek parrot: review: https://trac.parrot.org/parrot/changeset/42830/
22:01 dalek parrot: r42831 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
22:01 dalek parrot: [distutils] clean tarball & zip
22:01 dalek parrot: review: https://trac.parrot.org/parrot/changeset/42831/
22:02 NotFound bacek: hi
22:04 fperrad ping chromatic
22:04 purl I can't find chromatic in the DNS.
22:05 chromatic fperrad, pong
22:05 fperrad chromatic, could you look at my comment in TT #1279 and confirm if it's a bug or not.
22:05 fperrad If not a bug, I just rename my library.
22:06 chromatic Will do.
22:07 chromatic It's working as designed right now, but it's easy to argue that it's surprising.
22:07 chromatic The easiest solution is to rename your library.
22:09 dalek winxed: r235 | julian.notfound++ | trunk/Makefile:
22:09 dalek winxed: make target for stage 1 testing, PacoLinux++
22:09 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=235
22:25 nopaste "fperrad" at 78.113.87.157 pasted "[Pheme] test report" (55 lines) at http://nopaste.snit.ch/18905
22:25 fperrad chromatic, pheme is now available via Plumage, but few tests fail (see http://nopaste.snit.ch/18905)
22:28 chromatic I'll take a look.  Thanks.
22:34 hercynium joined #parrot
22:39 dalek winxed: r236 | julian.notfound++ | trunk/ (2 files):
22:39 dalek winxed: add a test for stage 1, PacoLinux++
22:39 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=236
22:53 Andy joined #parrot
23:09 Coke chromatic: partcl-nqp should provide you with more alternatives to profile things.
23:09 Coke also cotto_work
23:11 Coke chromatic: I know this isn't your thing any more, but is there a way to post a comment on: http://www.oreillynet.com/onlamp/blog/200​7/07/cpan_module_review_appsvnbinar.html ?
23:18 chromatic I think there is, but I'm not sure how.
23:19 Coke chromatic: k. that's pretty much the only place on the net that refers to that module.
23:21 * Coke reviews his own module.
23:22 chromatic Oh wow, that's the third blogging platform they abandoned, out of... I dunno, they're on number five or six now?
23:25 chromatic Coke, I doubt I even have a login there anymore.
23:33 Coke chromatic: give parrot.org time, we'll get there.
23:59 allison joined #parrot

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

Parrot | source cross referenced