Camelia, the Perl 6 bug

IRC log for #parrot, 2009-11-27

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:07 dalek winxed: r231 | julian.notfound++ | trunk/winxed (3 files):
00:07 dalek winxed: reintroduce -e option mistakenly dropped from stage 0, add -e option to stage 1
00:07 dalek winxed: and fix -e handling in driver
00:07 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=231
00:14 bluescreen joined #parrot
00:16 dalek TT #1337 created by pmichaud++: deprecate runtime/parrot/library/Protoobject.pir
00:26 eternaleye joined #parrot
00:50 plobsing joined #parrot
01:07 zak_ joined #parrot
01:33 cghene joined #parrot
01:39 lucian joined #parrot
01:47 bogen joined #parrot
01:59 zak_ joined #parrot
02:11 cconstantine joined #parrot
02:28 tetragon joined #parrot
03:21 tetragon joined #parrot
03:51 plobsing hi #parrot
03:54 Tene hi plobsing
03:54 plobsing who should I talk to about eval.pmc's freeze/thaw not working?
04:06 chromatic I think you just volunteered!
04:07 plobsing i am going to try and fix it, I just want to know if its already a known issue
04:07 plobsing I didn't see anything in trac
04:10 chromatic There are a few tickets for related issues.
04:47 eternaleye joined #parrot
04:58 JimmyZ joined #parrot
05:36 Ryan52 joined #parrot
06:27 JimmyZ_ joined #parrot
07:06 jan_ joined #parrot
07:12 he__ joined #parrot
07:15 uniejo joined #parrot
07:16 eternaleye joined #parrot
07:39 cotto 39.75/1.5
07:39 purl 26.5
07:43 dalek parrot: r42814 | cotto++ | trunk/tools/dev/pprof2cg.nqp:
07:43 dalek parrot: [pprof2cg] final fixes on nqp port.  Now it works *and* is 26.5x slower than the perl version.
07:43 dalek parrot: review: https://trac.parrot.org/parrot/changeset/42814/
07:46 moritz cotto: now you have a very good reason to optimize parrot :-)
08:09 JimmyZ Does that mean parrot is 26.5 slower than perl 5?
08:13 cconstantine joined #parrot
08:23 moritz no
08:26 kthakore moritz: having fun this morning? :)
08:26 moritz too tired
08:26 kthakore lol
08:27 kthakore moritz: i was refering to the question of parrot v.s perl5 speed
08:27 riffraff joined #parrot
08:28 iblechbot joined #parrot
08:29 kthakore moritz: i was refering to the question of parrot v.s perl5 speed
08:29 kthakore oops
08:29 kthakore stupid iphone
08:54 elmex joined #parrot
09:20 fperrad joined #parrot
09:22 fperrad_ joined #parrot
09:32 mikehh joined #parrot
09:52 theory joined #parrot
10:20 masak joined #parrot
10:28 fperrad seen vadrer
10:28 purl I haven't seen 'vadrer', fperrad
10:31 payload joined #parrot
10:40 theory joined #parrot
10:51 dalek parrot: r42815 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
10:51 dalek parrot: [distutils] fix loop in _clean_win32_installer
10:51 dalek parrot: review: https://trac.parrot.org/parrot/changeset/42815/
10:55 payload joined #parrot
11:43 iblechbot joined #parrot
12:17 uniejo joined #parrot
12:18 JimmyZ joined #parrot
12:18 mariano__ joined #parrot
13:04 mariano__ joined #parrot
13:16 lucian joined #parrot
13:21 xenoterracide joined #parrot
13:29 xenoterracide joined #parrot
13:30 xenoterracide joined #parrot
13:41 tetragon joined #parrot
13:43 cconstantine joined #parrot
13:45 xenoterracide joined #parrot
13:48 xenoterracide joined #parrot
13:52 xenoterracide joined #parrot
13:55 xenoterracide joined #parrot
14:01 plobsing joined #parrot
14:28 he__ joined #parrot
14:29 riffraff joined #parrot
14:30 cconstantine joined #parrot
14:35 riffraff joined #parrot
15:02 Psyche^ joined #parrot
15:03 lucian joined #parrot
15:37 Coke purl owner?
15:37 purl owner is hachi, see also #purl
15:40 dalek rakudo: 2b93c78 | (Solomon Foster)++ | src/setting/Rat.pm:
15:40 dalek rakudo: Rework most of the Rat math functions to use / to construct Rats rather than Rat.new.
15:40 dalek rakudo: This means that Rat math will automatically fall back to Num math if the intermediate values cannot be represented as Ints in Rakudo.  (Previously it would yield a cryptic error message because Rat.new insisted on Int arguments.)
15:40 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/2​b93c7883ab3df5ca79d45d9cb1d4e7d42d81bc6
15:55 darbelo joined #parrot
15:55 bogen joined #parrot
16:14 bogen joined #parrot
16:15 payload joined #parrot
16:54 dukeleto mikehh: it would be nice if you could put the name of the test that is being translated in the commit message, in addition to the TT
17:12 dalek parrot: r42816 | dukeleto++ | trunk/t/pmc/filehandle.t:
17:12 dalek parrot: [t][cage] De-backslash-itis t/pmc/filehandle.t
17:12 dalek parrot: review: https://trac.parrot.org/parrot/changeset/42816/
17:41 bogen joined #parrot
18:30 Coke msg hachi I know I asked you to have purl ignore dalek, but there's karma information in there which is more important than the occasional "change" message. Can you have purl unignore dalek?
18:30 purl Message for hachi stored.
18:30 Coke (there.)
18:31 mikehh dukeleto: it is generally in the first line of the message unless there is more than one
18:32 mikehh dukeleto: for example look at the commit you just did
18:34 dukeleto mikehh: that is from the irc bot
18:34 mikehh ah ok
18:34 dukeleto mikehh: that is not what shows up when i use "svn log" or "git log" and you cannot see what file is being modified without looking at the commit
18:35 Coke you don't need to include thefilename in the comm9it message.
18:35 Coke it's part of the commit info\
18:36 dukeleto whatever. it was just my suggestion. it makes it hard to see what is going on with 10 commits with the same commit message and just different TT #'s
18:37 mikehh dukeleto: I generally haven't put the file there because it is in the email or dalek's commit message
18:37 dukeleto each commit message should be a unique butterfly
18:39 mikehh dukeleto: it becomes difficult if you have removed say, trailing spaces, from 3 or 4 files in one commit
18:39 ascent joined #parrot
18:39 mikehh dukeleto: unless you commit them separately
18:43 mikehh dukeleto: however I will bear it in mind - I used to do that, but thought it was redundent
18:52 dukeleto in my opinion, when one file has been re-written in a new language in a commit, it is useful to name that file in the commit message. in one-line log mode in svn/git, it is not redundant
18:54 Coke you'll get people complaining either way. =-)
19:09 PacoLinux joined #parrot
19:09 joeri joined #parrot
19:12 bogen joined #parrot
19:35 preflex joined #parrot
19:37 lucian joined #parrot
19:38 cotto joined #parrot
19:51 nopaste joined #parrot
19:52 eternaleye joined #parrot
20:21 dalek parrot-plumage: 068ae26 | japhb++ | :
20:21 dalek parrot-plumage: [TESTS] Fix qx() test for Win32
20:21 dalek parrot-plumage: review: http://gitorious.org/parrot-plumage/parrot-plumag​e/commit/068ae26761eae0e77dfc180b359c4353ce70c5cf
20:37 eternaleye joined #parrot
20:37 joeri joined #parrot
20:42 dalek nqptap: fbb061d | dukeleto++ |  (12 files):
20:42 dalek nqptap: Convert over to PIR-based Configure and build system
20:42 dalek nqptap: Parrot's script for generating a language shell was used and tweaked.
20:43 dalek nqptap: This configuration does not yet compile on the NQP that Parrot
20:43 dalek nqptap: has in trunk right now:
20:43 dalek nqptap: (kadath)(~/git/nqptap master )$ parrot-nqp Configure.pir
20:43 dalek nqptap: Confused at line 4, near ".include '"
20:43 dalek nqptap: current instr.: 'parrot;HLL;Grammar;panic' pc 519 (src/stage0/HLL-s0.pir:336)
20:43 dalek nqptap: called from Sub 'parrot;NQP;Grammar;comp_unit' pc 4922 (src/stage0/NQP-s0.pir:138)
20:43 dalek nqptap: called from Sub 'parrot;NQP;Grammar;TOP' pc 965 (src/stage0/NQP-s0.pir:368)
20:43 dalek nqptap: called from Sub 'parrot;Regex;Cursor;parse' pc 282 (src/stage0/Regex-s0.pir:181)
20:43 dalek nqptap: called from Sub 'parrot;HLL;Compiler;parse' pc 24 (src/stage0/HLL-s0.pir:32)
20:43 dalek nqptap: called from Sub 'parrot;PCT;HLLCompiler;compile' pc 388 (src/PCT/HLLCompiler.pir:287)
20:43 dalek nqptap: called from Sub 'parrot;PCT;HLLCompiler;eval' pc 937 (src/PCT/HLLCompiler.pir:532)
20:43 dalek nqptap: called from Sub 'parrot;PCT;HLLCompiler;evalfiles' pc 1224 (src/PCT/HLLCompiler.pir:676)
20:43 dalek nqptap: called from Sub 'parrot;PCT;HLLCompiler;command_line' pc 1410 (src/PCT/HLLCompiler.pir:763)
20:43 dalek nqptap: called from Sub 'parrot;NQP;Compiler;main' pc -1 ((unknown file):-1)
20:43 dalek nqptap: As far as I know, NQP does not yet have a version numbering scheme and
20:43 dalek nqptap: this makes the situation icky. pmichaud++ has described a good one, it
20:43 dalek nqptap: just has not been implemented yet.
20:43 dalek nqptap: review: http://github.com/leto/nqptap/commit/fb​b061dcceb8a3b9717be873bc9756245173b563
20:45 donaldh joined #parrot
20:48 japhb dukeleto, why would you use parrot-nqp to run a .pir file, instead of using parrot itself?
20:49 dukeleto japhb: i am sorely confused
20:49 * dukeleto blames enormous amounts of turkey
20:50 japhb dukeleto, your nqptap commit.  You used the command ' parrot-nqp Configure.pir'.  That makes no sense.  You should use 'parrot Configure.pir' instead.
20:51 japhb And if you're going the PIR route, why not skip Configure entirely and just switch to fperrad's setup.pir system instead?
20:53 kthakore dukeleto: why blame turkey ... it never did nothing to you :9
20:53 kthakore dukeleto: don't be a turkey hater ... it only wants your love
20:56 dalek parrot: r42817 | fperrad++ | trunk/t/library (4 files):
20:56 dalek parrot: [t][cage] factorize load_bytecode 'lib.pbc'
20:56 dalek parrot: review: https://trac.parrot.org/parrot/changeset/42817/
21:01 mariano__ joined #parrot
21:12 dukeleto japhb: how do (and why should) I use fperrad's system? i used mk_lang_shell
21:12 japhb He just committed in here 20 min ago.
21:12 japhb fperrad: Are you still around?
21:13 japhb dukeleto, It's actually amazingly cool-simple.
21:13 fperrad japhb, hello
21:14 japhb fperrad: dukeleto wants to know how to convert nqptap to setup.pir.
21:14 japhb I figured you could help.  ;-)
21:15 fperrad japhb, you are right
21:17 dalek nqptap: bf1c683 | dukeleto++ | .gitignore:
21:17 dalek nqptap: Previous commit message was post-Thanksgiving-food-coma delirium. Improved .gitignore
21:17 dalek nqptap: review: http://github.com/leto/nqptap/commit/bf​1c683eb909a58564c0c0bb3b2c7074cde7b47c
21:18 japhb "I'm delirious.  Oh, and I edited something."  :-)
21:18 donaldh_ joined #parrot
21:32 allison joined #parrot
21:43 iblechbot joined #parrot
21:53 MoC joined #parrot
21:55 dukeleto fperrad: how do i convert to setup.pir? what features does it have over what mk_lang_shell gives me?
21:58 fperrad dukeleto, mk_lang_shell generates a skeleton for a compiler built over PGE & NQP
22:01 fperrad dukeleto, currently mk_lang_shell generates 2 ways for building
22:01 fperrad * parrot Configure.pir (with Makefile.in) which is currently broken (see TT #1279)
22:01 fperrad * parrot setup.pir
22:03 dalek parrot-plumage: 3cda70e | japhb++ | :
22:03 dalek parrot-plumage: Move depsolving to Plumage::Dependencies: first working code
22:03 dalek parrot-plumage: review: http://gitorious.org/parrot-plumage/parrot-plumag​e/commit/3cda70e853c740624e22b8f433f8e40e8bc5af6d
22:03 dalek parrot-plumage: b5d2614 | japhb++ | :
22:03 dalek parrot-plumage: [plumage] Improve comments in HACKS section
22:03 dalek parrot-plumage: review: http://gitorious.org/parrot-plumage/parrot-plumag​e/commit/b5d261485f94f773a63e426cf06f55638fc5ab64
22:05 fperrad dukeleto, setup.pir contains data which describes your project (basically pathnames, dependencies, ...)
22:05 fperrad setup.pir loads the library distutils.pbc which contains all rules needed to build (with PGE, TGE, NQP, NQP-RX, PBC_TO_EXE)
22:10 dukeleto fperrad: ok, how do i switch to use setup.pir then?
22:11 japhb fperrad: do you have a "setup.pir howto" document somewhere?
22:12 fperrad dukeleto, there some examples (links) in the runtime/library/distutils.pir
22:13 fperrad I try to convert Kea
22:15 dukeleto fperrad: thanks!
22:16 dukeleto fperrad: that would be a great example for me, and then I can try to do it to nqpTAP
22:17 dukeleto we need a new column in the parrot language list: configure/build language
22:20 dalek parrot-plumage: d9626c7 | japhb++ | :
22:20 dalek parrot-plumage: [LIB] Convert line missed during Plumage::Dependencies refactoring
22:20 dalek parrot-plumage: review: http://gitorious.org/parrot-plumage/parrot-plumag​e/commit/d9626c727442e173a55065924ef7067ca83536e0
22:22 theory joined #parrot
22:26 dukeleto fperrad: you have a commit bit on the kea github repo
22:27 nopaste "fperrad" at 79.81.118.39 pasted "[Kea] convert to setup.pir (for dukeleto)" (122 lines) at http://nopaste.snit.ch/18882
22:28 dukeleto fperrad: can you commit that to the master branch of kea?
22:29 dukeleto fperrad: and add yourself to CREDITS as well, please :)
22:30 dukeleto fperrad: Kea is Artistic License 2.0, in case you are wondering
22:36 dukeleto any one else who wants a commit bit on Kea, just ask me
22:38 japhb Anyone around happen to know how to tell if an optional param in NQP-rx has been supplied or not (:opt_flag in PIR)?  I *don't* mean "default to an undefined value".  I actually need to know if it was supplied.
22:42 dngor joined #parrot
22:43 Tene japhb: the answer is "Don't do that.  Find a different way to do it."
22:43 Tene japhb: the generated pir does have a :opt_flag, but it's not used for anything other than setting defaults.
22:44 Tene japhb: does nqp support block defaults, though?
22:44 Tene looks like it does, maybe...
22:45 Tene oh, no, that just puts the actual block into the variable.
22:46 Tene japhb: I'd do something like making a new class and setting that as the default.  class NotSupplied { ... }; sub foo($a=NotSupplied) { if $a ~~ NotSupplied { ... } }
22:46 Tene I bet something like that could work.
22:46 Tene I don't know what you're trying to do here exactly, though, so I couldn't say more.
22:46 cconstantine joined #parrot
22:46 Tene Well, I could, but it wouldn' tbe useful. :)
22:56 fperrad dukeleto, pushed on Kea, now, you have the following commands :
22:56 fperrad - parrot setup.pir build
22:56 fperrad - parrot setup.pir clean
22:56 fperrad - parrot setup.pir test
22:56 fperrad - sudo parrot setup.pir install
22:56 fperrad - sudo parrot setup.pir uninstall
22:56 fperrad - parrot setup.pir plumage > plumage/kea.json
22:57 dalek kea: 05aa9ff | unknown++ |  (4 files):
22:57 dalek kea: convert to setup.pir (distutils)
22:57 dalek kea: review: http://github.com/leto/kea/commit/05a​a9ff1303ddd0e4dd9c1e5dab0192ae6fb64b1
22:57 dalek kea: c58c501 | unknown++ |  (4 files):
22:57 dalek kea: remove old stuff
22:57 dalek kea: review: http://github.com/leto/kea/commit/c58​c501098b52dac79492acaadf32468c8f74827
23:06 dalek TT #1338 created by vkon++: 3 C functions for tcl/tk library
23:09 japhb Tene, thank you for the ideas
23:14 japhb Tene: and FWIW, I was implementing a reduce() function, which classically has two related definitions, one with an initial value supplied, and one without.  And you can't just treat the one without an initial value as the same as one with an undefined initial value.  In full Perl 6, I would have solved that with MMD, but since that's not available ....
23:14 dukeleto fperrad: thanks!
23:14 Tene japhb: Ah.
23:16 dukeleto fperrad: how do i tell setup.pir to use my custom test harness? which does it use by default?
23:25 dukeleto i answered my question
23:26 dalek kea: ad8b4be | dukeleto++ | README:
23:26 dalek kea: Update README to reflect new build system from fperrad++
23:26 dalek kea: review: http://github.com/leto/kea/commit/ad8​b4bed592810f53bcbb82aedf04dbbba3b2696
23:30 japhb dukeleto or fperrad: drop me an email (or send me a repo link) for the updated kea.json, and I'll update plumage.
23:32 dukeleto japhb: i think i can do that now
23:32 japhb dukeleto, OK, perfect
23:37 dalek kea: 6032262 | dukeleto++ |  (2 files):
23:37 dalek kea: Add a plumage metadata file and correct the README
23:37 dalek kea: review: http://github.com/leto/kea/commit/603​2262ebd9d8cb4e7fd45bc656a75de7cb2c6de
23:39 japhb dukeleto, kea.json is wrong ... it claims Perl 6 in 'provides' and 'keywords'.
23:39 plobsing joined #parrot
23:39 dukeleto japhb: oops!
23:41 dukeleto japhb: incoming
23:41 purl duck!
23:42 japhb :-)
23:42 dukeleto msg fperrad parrot setup.pir plumage does not work for me
23:42 purl Message for fperrad stored.
23:43 dalek kea: 05c9b84 | dukeleto++ | plumage/kea.json:
23:43 dalek kea: Fix the 'provides' and 'keywords' fields in the Plumage metadata file based on japhb++'s comments
23:43 dalek kea: review: http://github.com/leto/kea/commit/05c​9b84e22755737cad4a2f1282a6dbd0b541685
23:43 dalek kea: 447b8ac | dukeleto++ | .gitignore:
23:43 dalek kea: Improve .gitignore
23:43 dalek kea: review: http://github.com/leto/kea/commit/447​b8ac4b573b3054cb2e5404eb6f0c457fa6c86
23:48 japhb dukeleto, I thought you switched to just switched to setup.pir for building?  The instructions say make ....
23:49 dukeleto japhb: i thought setup.pir does not use Makefile?
23:49 dukeleto msg fperrad i would like to add custom steps to setup.pir, how hard is that?
23:49 purl Message for fperrad stored.
23:50 dukeleto =head1 DESCRIPTION
23:50 dukeleto No Configure step, no Makefile generated.
23:51 dukeleto i want to add a configure setup with a small readme-like printout. i like that step, even if there is nothing to be actually configured
23:51 dukeleto you can point users to documentation, the irc channel, mailing lists, etc...
23:51 japhb dukeleto, that's what I'm saying, kea.json says 'make' for the instruction types for most steps, instead of 'parrot_setup'
23:52 japhb dukeleto, FWIW, that's conceptually a different step ... being able to give the user a message is something we ought to treat specially, without hacking it in as a side effect of some normal stage
23:53 dukeleto japhb: i just gave you a commit bit to kea, can you fix it? i had to hand-write that file because generating it wasn't working
23:53 dukeleto japhb: then i want a "readme" step in plumage
23:53 japhb dukeleto, give me a second to finish what I'm working on, then yes.
23:54 japhb And a readme step is a reasonable request ... but the question is, what does it do and when does it happen?
23:55 dukeleto japhb: it is an optional step that can be run at any time
23:56 workbench joined #parrot
23:56 japhb dukeleto, optional sure.  Run at any time seems LTA, since it could happen in the middle of stuff that scrolls it off the screen.
23:57 dukeleto japhb: optional *first* step works for me
23:57 japhb I'm almost thinking that Plumage should collect the readme results from all projects and dependencies, and display them all at once.  Sort of like APT does.
23:57 japhb dukeleto, then it would have to pause the install -- for each dependency as well.  A single plumage install command could be paused several times.
23:58 dukeleto japhb: why paused? it just prints stuff out
23:58 dukeleto japhb: no pausing, just printing. if it is an automated dependency then i guess don't print it
23:58 dukeleto the upper level readme may all the end user actually wants to see
23:58 japhb dukeleto, and when the build scrolls the screen past, how does the user see the readme ...
23:59 dukeleto japhb: i am thinking about developers typing in the various "parrot setup.pir xxx" commands
23:59 tetragon joined #parrot
23:59 dukeleto japhb: i don't care if the readme scrolls by, i just want it printed out
23:59 dukeleto japhb: it is very short

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

Parrot | source cross referenced