Camelia, the Perl 6 bug

IRC log for #parrot, 2010-08-20

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:00 Coke goes with slash.
00:00 cotto_work You can use it.  It'll just be empty.
00:01 Psyche^ joined #parrot
00:03 ruoso joined #parrot
00:09 dukeleto cotto_work: could be, but maybe not. I don't like how the master branch has merge commits, so that will change
00:09 Paul_the_Greek joined #parrot
00:09 cotto_work ok
00:09 Paul_the_Greek Evening, folks.
00:09 purl Evening yourself, and see how you like it.
00:09 cotto_work hio Paul_the_Greek
00:09 Paul_the_Greek Damn you, purl
00:09 purl Damn ME!?  I think my karma's a bit higher than yours, bub.
00:10 Paul_the_Greek Hey cotto.
00:10 Paul_the_Greek Can you do me a huge favor, cotto?
00:10 cotto_work You can clearly see why botkicking is a popular passtime around here.
00:10 Paul_the_Greek Who whipped up purl, anyway?
00:11 pmichaud 23:33 <cotto_work> dukeleto: will your parrot mirror on github most likely become the offical repo in its current shape?
00:11 pmichaud fwiw, I converted the 'parrot' account to an organizational account a couple of days ago.
00:12 dalek plparrot: 39e900f | leto++ | t/sql/plperl6.sql:
00:12 dalek plparrot: Use a rule instead of a regex in PL/Perl6 grammar example, bacek++
00:12 dalek plparrot: review: http://github.com/leto/plparrot/commit/3​9e900fe9a2be52fe17009fe8a82c9dc256653e8
00:12 dalek plparrot: 01b4a27 | leto++ | t/sql/test.sql:
00:12 dalek plparrot: Add a test for concatenating input args in PL/PIR
00:12 Paul_the_Greek Favor: Could someone nopaste complex.c generated from complex.pmc?
00:12 dalek plparrot: review: http://github.com/leto/plparrot/commit/0​1b4a279f910d2161b01a26c162388b46a305840
00:12 dalek plparrot: ae12d06 | leto++ | t/sql/test.sql:
00:12 dalek plparrot: Add some PL/PIR tests for conditionally concatenating input strings
00:12 dalek plparrot: review: http://github.com/leto/plparrot/commit/a​e12d064324d3be2ef727f624b1ba1f88c4a6701
00:13 nopaste "cotto" at 192.168.1.3 pasted "here you go" (3474 lines) at http://nopaste.snit.ch/22923
00:13 dukeleto pmichaud: cool!
00:14 Paul_the_Greek Thank you, sir.
00:14 nwellnhof ++pmichaud
00:14 cotto_work pmichaud++
00:14 dukeleto pmichaud: could you add me as a member so that I can create a parrot repo and make a better github mirror ?
00:14 dukeleto pmichaud: username = leto
00:16 nwellnhof Will parrot completely switch to git anytime soon?
00:17 pmichaud dukeleto: what sort of membership is needed?
00:18 pmichaud dukeleto: anyway, you're added to the "Parrot" team
00:18 cotto_work nwellnhof:  yes.  One of the roadblocks is getting a nice way to redirect trac urls to svn revisions to the right place on github, which I'm working on.
00:18 Paul_the_Greek Aaarrrgh! It also generates pmc_complex.h
00:19 whiteknight joined #parrot
00:19 dukeleto pmichaud: i think any membership will do, public is fine
00:19 nwellnhof Any other roadblocks?
00:19 pmichaud dukeleto: let me know if you end up needing anything further
00:20 cotto_work we just need to nail down the timeline and make sure enough people agree about the specifics
00:20 dukeleto pmichaud: will do! i am furiously working on some slides for a talk I am giving about PL/Parrot in less than 2 hours, but I will definitely play with the parrot org on github later
00:20 pmichaud dukeleto: no rush
00:20 pmichaud good luck on your talk
00:21 pmichaud oh, I was supposed to tell you how many tests are passing... let me see if I can determine that quickly
00:21 dukeleto pmichaud: thanks! i will be showing off using p6 grammars from postgres
00:21 Paul_the_Greek cotto_work:  I humbly request that you nopaste pmc_complex.h
00:21 dukeleto pmichaud: i just put ~80%, so no big deal. unless it is more than that, and I can update it :)
00:21 pmichaud I still don't know how many tests are in the suite.  ~80% is undoubtedly correct
00:21 nopaste "cotto" at 192.168.1.3 pasted "pmc_complex.h" (95 lines) at http://nopaste.snit.ch/22924
00:21 pmichaud it's certainly not an over-estimate
00:22 Paul_the_Greek cotto++
00:24 cotto_work If you have a question about the madness that is pmc2c, someone here is likely to be able to answer it.
00:27 Paul_the_Greek No problem with pmc2c. I'm trying to figure out when automatic PMC attributes are allocated.
00:27 cotto_work NotFound would be best to ask about that.
00:28 cotto_work He's the one who's responsible for auto_attrs, which does what you're talking about.
00:28 Paul_the_Greek Where is he, time-wise?
00:29 cotto_work spain
00:29 Paul_the_Greek Too late now.
00:30 cotto_work nothing a bit of acking wouldn't fix
00:30 Paul_the_Greek acking?
00:30 cotto_work using ack
00:30 nwellnhof Paul_the_Greek: It seems they're allocated in src/pmc.c
00:31 Paul_the_Greek Sorry, don't know what ack is.
00:31 cotto_work nwellnhof: how'd you get roped into Parrot?
00:31 Paul_the_Greek Let me check pmc.c ...
00:31 cotto_work ack?
00:31 purl it has been said that ack is http://betterthangrep.com or a grep-like tool for code. or at http://www.betterthangrep.com/
00:31 cotto_work ^
00:32 nwellnhof cotto: being a long time Perl5 user, basically
00:32 Paul_the_Greek Oh, that ack. I thought you were talking about contacting NotFound.
00:32 dukeleto does anybody know where I can find a hi-res PNG of the Parrot logo ?
00:32 dukeleto i think it is lost in my email somewhere
00:33 dukeleto found it!
00:34 whiteknight in perl5, is there is known-good regex or something to convert "this_is_an_identifier" into "ThisIsAnIdentifier"?
00:34 whiteknight I could try to invent that wheel, but I suspect a much better version already exists somewhere
00:37 dukeleto whiteknight: have you looked in Regexp::Common ?
00:37 nwellnhof cotto: and maybe implementing XSLT on parrot sometime
00:37 whiteknight no. I didn't know that module existe
00:37 whiteknight d
00:38 Paul_the_Greek nwellnhof:  Allocated there, but auto_attrs is handled somewhere else.
00:39 nwellnhof auto_attrs?
00:39 chromatic see get_new_pmc_header()
00:40 Paul_the_Greek It determines whether a PMC's attribute block is allocated automatically.
00:40 whiteknight dukeleto: I pieced something together. Not pretty
00:40 chromatic src/pmc.c line 450
00:40 Paul_the_Greek I'm at the beach without the source, which is why I sound so helpless.
00:40 cotto_work You're never without the source: http://trac.parrot.org/parrot/browser/trunk
00:41 Paul_the_Greek I should say, I'm without the ability to search the source and generated files in any reasonable manner.
00:42 cotto_work Ah.  It takes a while before you instinctively know where to look for things.
00:43 Paul_the_Greek That's for sure. Enough files named pmc* to confuse the likes of me.
00:46 spinclad joined #parrot
00:47 * cotto_work puts writing a "Navigating the Parrot Source" doc on his todo list
00:47 nwellnhof \QUIT
00:49 Paul_the_Greek cotto_work:  Yikes. Just recommend editors with good regex searches.
00:50 Paul_the_Greek Actually, this is rather amusing. I've now guessed and looked at about 15 files without finding where auto_attrs is handled.
00:52 nwellnhof joined #parrot
01:02 cotto_work Paul_the_Greek: when a PMC uses auto_attrs, attr_size gets set.  The code that checks ->attr_size in src/pmc.c allocates/frees space for the attr struct.
01:02 Paul_the_Greek Hang on ...
01:02 sorear we should just switch to git
01:03 sorear then Paul_the_Greek can have a full clone, and use git grep
01:03 aloha joined #parrot
01:03 cotto_work aloha: aloha
01:03 bacek joined #parrot
01:03 Paul_the_Greek I'm groovy at home, just helpless here at the beach.
01:04 dafrito Paul_the_Greek: you could use SSH, possibly
01:05 sorear Paul_the_Greek: Why are you helpless at the beach?  And how are you on IRC?
01:05 Paul_the_Greek Yes, that might work.
01:06 Paul_the_Greek I'm on my wife's laptop, which has wifi. But I wouldn't want to download gobs of stuff on it.
01:07 Paul_the_Greek cotto_work:  Ah, I get it. auto_attrs causes attr_size to be set.
01:08 Paul_the_Greek Now my concern is that no one is rounding attr_size to a multiple of pointer size.
01:09 Paul_the_Greek I'll work that out when I get home.
01:18 kid51 joined #parrot
01:18 dalek website: Chandon++ | Hybrid Threads: GSoC Project Results
01:18 dalek website: http://www.parrot.org/content/hyb​rid-threads-gsoc-project-results
01:23 kid51 chandon++
01:24 dafrito cotto_work: I removed the swp file and pushed it to github
01:24 cotto_work gracias
01:25 dafrito cotto_work: did you want it removed from the history as well?
01:26 dalek github-trac: 1ed3b50 | dafrito++ | .gitignore:
01:26 dalek github-trac: Added *~ and *.swp to ignored files
01:26 dalek github-trac: review: http://github.com/cotto/github-trac/commit​/1ed3b50e59093f2f5f74539d8bcfbfcd95d4b756
01:26 dalek github-trac: 4add134 | dafrito++ | github/.github.py.swp:
01:26 dalek github-trac: Removed a mistakenly added *.swp file
01:26 dalek github-trac: review: http://github.com/cotto/github-trac/commit​/4add134e76e78a7c88b169ee41ccef11fc41a1d0
01:26 dafrito We'll just leave it, no big deal
01:27 cotto_work It'd be nice but I don't care much.
01:29 * dafrito sets up a dummy Trac for himself
01:30 Paul_the_Greek Does it mean something special when a static function is named Parrot_xxx?
01:31 cotto_work possibly that it was misnamed
01:32 Paul_the_Greek That came to mind, but I was being polite. :D
01:33 Paul_the_Greek I think we really can make a dent in the time required to create a PMC.
01:33 cotto_work If code is junk, it's not impolite to suggest the possibility.
01:33 Paul_the_Greek But a newbie should be circumspect about doing so.
01:33 Paul_the_Greek Just wait until I've been around for awhile.
01:34 cotto_work If you're more comfortable with that approach, sure.
01:35 Paul_the_Greek For now, yes.
01:36 cotto_work dafrito: have you set up trac before?
01:37 dafrito cotto_work: no
01:37 dafrito though the docs don't seem too bad, I'm just slow ;)
01:38 cotto_work Have you done much Python-related hacking?
01:38 dafrito cotto_work: I've worked with Django a decent amount
01:44 cotto_work ping me if you run into something tricky.  Having recently gone through the process, there's a good chance I can help.
01:44 dafrito awesome, will do :)
01:45 cotto_work and make sure to install the 0.1.7 version of GitPython before you try to install the github plugin
01:45 cotto_work 0.3.0 makes trac sad
01:45 * cotto_work is glad to have a better bus number on this plugin
01:46 cotto_work Minefields are fine if you know where not to step.
01:53 petdance joined #parrot
01:53 davidfetter joined #parrot
02:02 * dukeleto is about to talk at PDXPUG about PL/Parrot
02:06 tcurtis dukeleto++
02:08 Paul_the_Greek dukeleto++
02:12 kid51 msg NotFound Whatever you did today re the failure I pasted yesterday appeared to work! Darwin/PPC make test PASS at r48571.  NotFound++
02:12 purl Message for notfound stored.
02:13 cotto ~~
02:21 whiteknight incoming
02:21 purl duck!
02:21 whiteknight ...and outgoing
02:24 dalek parrot-linear-algebra: 415e387 | Whiteknight++ | t/ (6 files):
02:24 dalek parrot-linear-algebra: implement gemm tests for num and complex matrices
02:24 dalek parrot-linear-algebra: review: http://github.com/Whiteknight/parrot-linear-algebr​a/commit/415e38787b7da9d3dc06cacca5aa4d2a8a0ad5aa
02:24 dalek parrot-linear-algebra: bc3d89d | Whiteknight++ | t/methods/ (5 files):
02:24 dalek parrot-linear-algebra: ...and the files
02:24 dalek parrot-linear-algebra: review: http://github.com/Whiteknight/parrot-linear-algebr​a/commit/bc3d89db949d49f5b89a880a0d48fc4ed23de626
02:24 dalek parrot-linear-algebra: accbccf | Whiteknight++ | t/testlib/ (13 files):
02:24 dalek parrot-linear-algebra: rename some files to match the name of the methods they test
02:24 dalek parrot-linear-algebra: review: http://github.com/Whiteknight/parrot-linear-algebr​a/commit/accbccfa3128b31ff02036bba2012d29cf2c7ad2
02:24 dalek parrot-linear-algebra: 46a7a96 | Whiteknight++ | / (59 files):
02:24 dalek parrot-linear-algebra: finish the majority of the test rework. All methods are encapsulated in their own files, and most tests are still inherited across types. I'm not entirely happy about the sheer amount of boilerplate, but I've been able to auto-generate most of them
02:25 dalek parrot-linear-algebra: review: http://github.com/Whiteknight/parrot-linear-algebr​a/commit/46a7a96c9af27fb969a41be945cedfc712b0f906
02:25 dalek parrot-linear-algebra: c843c5e | Whiteknight++ | t/pmc/nummatrix2d.t:
02:25 dalek parrot-linear-algebra: remove duplicate test
02:25 dalek parrot-linear-algebra: review: http://github.com/Whiteknight/parrot-linear-algebr​a/commit/c843c5efe8cb41fbb0e20d8eaf56a1732f723269
02:47 atrodo My favorite part of git, it seeing trees the merges make
02:55 dafrito cotto: Do I need to install the git plugin for trac as well as the github-trac plugin? Or just the latter?
03:15 cotto just github-trac
03:19 janus joined #parrot
03:33 theory joined #parrot
03:55 darbelo left #parrot
04:02 dalek parrot: r48572 | chromatic++ | trunk/src/pmc/class.pmc:
04:02 dalek parrot: [OO] Optimized Class's isa_pmc() slightly.
04:02 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48572/
04:32 * dukeleto survived the talk
04:33 dukeleto I get the "intermittent core dump during a live demo badge", though ;)
04:36 tadzik joined #parrot
04:37 Chandon joined #parrot
04:42 dukeleto this is the core dump, which looks like StringBuilder corrupting memory: http://gist.github.com/539612
04:51 dukeleto i can't seem to reproduce it
05:17 cotto try giving your presentation again
05:20 dukeleto cotto: :)
05:26 * dukeleto learns about ops2c
05:27 bacek_in_sing joined #parrot
05:34 dukeleto bacek_in_sing: morning, meatbag
05:36 sorear dukeleto: Cool.  Any trolls/hecklers? :)
05:36 sorear "Look at Perl 6, it crashes in live demos... and I left all my tomatoes home"
05:44 dukeleto sorear: no hecklers, they were nice. I ran the query again and it worked :)
05:45 dukeleto sorear: somehow, the exact sequence of commands made it happen, but i can't seem to make it happen again. I blamed neutrinos during the talk
05:46 dukeleto here is a link to the PDF of the slides http://github.com/leto/presentatio​ns/raw/master/2010/PDXPUG/pres.pdf
05:46 dukeleto i think the talk was recorded too
05:51 dafrito cotto: ping
05:53 dalek plparrot: dfe6711 | leto++ | html/docs.html:
05:53 dalek plparrot: Add a link to the PDXPUG slides
05:53 dalek plparrot: review: http://github.com/leto/plparrot/commit/d​fe6711cac21c23feabe4082eb1a064a2df5ace3
05:54 cotto dafrito, pong
05:56 dafrito cotto: I got (what I believe to be, anyway) a few bugfix commits, and an interesting branch. I was wondering how they should be committed?
05:57 cotto I'm sure there are plenty of bugs lurking.
05:57 dukeleto dafrito: are they related? do they have have TT's attached?
05:57 dafrito I got my setup pretty much good to go, since it runs without errors, though I haven't set up test data yet
05:57 dukeleto dafrito: separate patches for each bugfix are good
05:58 cotto you can commit the bugfixes directly and push the branch
05:58 dafrito dukeleto: It's for the github-trac plugin, but yeah, they're in separate commits
05:58 cotto yes, separate patches would be preferred (though I'm not especially consistent on that front)
05:59 dafrito I squashed a few together since they weren't very useful on their own, but I'd like to think they're atomic ;)
06:00 * dukeleto is working on the parrot github org: http://github.com/parrot/parrot
06:00 dafrito Should I push them to a separate branch first? Or mail them? I don't really want to pollute master with any mistakes
06:00 uniejo joined #parrot
06:02 cotto how about a separate branch
06:08 dukeleto messing with ops is messy
06:10 cotto yes
06:10 dafrito Okay, they're pushed. They live in "bugfixes" and "admin", respectively. http://fpaste.org/YQwm/ shows a quick log and diffstat of the two branches
06:12 dukeleto Incompatible versions of `core_ops' oplib. Found 2.6.0 but loaded 2.7.0
06:12 dukeleto what is that about?
06:13 dukeleto i think i see the light. I have a plain parrot install and a rakudo install of parrot
06:13 ruoso joined #parrot
06:18 cotto *sigh*
06:20 cotto I'll fix that.
06:23 cotto done
06:26 dukeleto check it out! http://github.com/parrot/parrot <-- we have a proper github mirror now. just need to setup a cronjob to update it
06:27 dukeleto sans dumb merge commits too!
06:27 dafrito dukeleto++ :D
06:28 dalek github-trac: 74df012 | dafrito++ | github/github.py:
06:28 dalek github-trac: Rename a variable to reveal our shadowed underscore
06:28 dalek github-trac: review: http://github.com/cotto/github-trac/commit​/74df012378dce90da84b4f68fa477038ded86e1e
06:28 dalek github-trac: 463d721 | dafrito++ | github/github.py:
06:28 dalek github-trac: Added trac.resource.ResourceNotFound import
06:28 dalek github-trac: review: http://github.com/cotto/github-trac/commit​/463d721119d35a737180ccc54d93d65876c2d1ce
06:28 dalek github-trac: 9546aee | dafrito++ | github/github.py:
06:28 dalek github-trac: Corrected the syntax used to translate our error message
06:28 dalek github-trac: review: http://github.com/cotto/github-trac/commit​/9546aee1cba2381f540411128d20c89f77fabbfb
06:28 dalek github-trac: dc0c83b | dafrito++ | github/github.py:
06:28 dalek github-trac: Made svn_revmap use both fields as the primary key
06:28 dalek github-trac:
06:28 dalek github-trac: This avoids integrity errors since revisions are not necessarily unique.
06:28 dalek github-trac: Specifically, we run into problems if a single SVN commit has affected
06:28 dalek github-trac: multiple branches. This can occur if unusual things have occurred in
06:28 dalek github-trac: SVN, such as a deletion of the branches/ or tags/ directories.
06:28 dalek github-trac: review: http://github.com/cotto/github-trac/commit​/dc0c83b2ebe3ecfa9bb726b011cd948ace1071a2
06:28 dalek github-trac: e2b3484 | dafrito++ | setup.py:
06:28 dalek github-trac: setup.py now explicitly requires GitPython 0.1.7
06:28 dalek github-trac: review: http://github.com/cotto/github-trac/commit​/e2b348422ab5efcf2e60f340d5d5e53b0df7f436
06:32 jan joined #parrot
06:33 dafrito The admin branch provides an "admin" command. It sets up a clean virtualenv, and handles the monotony of picking up and auto-updating the dependencies
06:33 dalek parrot: r48573 | cotto++ | trunk (2 files):
06:33 dalek parrot: bootstrap ops and bump PBC_COMPAT for 2.7.0 release
06:33 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48573/
06:33 dalek parrot: r48574 | NotFound++ | trunk/t/pmc/packfile.t:
06:33 dalek parrot: improve Packfile get_string_keyed test
06:33 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48574/
06:34 cotto dafrito++ for that.  I was pretty sure I'd be the only one hacking on that plugin.
06:35 cotto What's the motivation?
06:35 purl i heard the motivation was brownies
06:35 cotto other than brownies
06:36 cotto just to make testing simpler?
06:36 dafrito Yeah, mostly
06:37 dafrito I like it because I can have a clean environment that's not secretly dependent on some package I have installed
06:37 dafrito Plus I can have multiple versions of things running around without having to worry about breaking stuff
06:39 dafrito But it's mostly a style thing, I think. I could just as well keep it on my own branch. I had a similar script lying around already, so it wasn't too hard to move it over
06:39 cotto That's a good idea.  I was pretty disappointed when I saw that Python eggs don't include automatically-run tests.
06:40 dafrito It uses tracd instead of apache, so it's not a drop-in replacement for yours. though it wouldn't be hard to support both methods
06:40 cotto I'm using tracd for development too
06:41 dafrito Oh, awesome.
06:41 dukeleto How do i put a NAN into a PMC from C?
06:41 dafrito It should work out of the box, assuming you have python2.6 and virtualenv: ./admin build && ./admin run
06:44 dukeleto src/ops/core_ops.c:19210: error: void value not ignored as it ought to be
06:44 dukeleto that is a new error for me. Didn't know that gcc knew the subjunctive case
06:47 sorear dukeleto: can't you just poke 0.0/0.0?
07:05 cotto How do I review staged commits?
07:05 tcurtis git diff --cached?
07:06 * tcurtis may be either misunderstanding or misremembering.
07:06 moritz no, that's right
07:06 cotto git diff HEAD seems to do it
07:23 dalek parrot: r48575 | NotFound++ | trunk/t/pmc/packfile.t:
07:23 dalek parrot: refactor and improve Packfile get_integer_keyed_str test
07:23 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48575/
07:28 cotto dafrito, do you have a revision map that I could download instead of generating one?
07:34 dafrito cotto, yeah, just a sec
07:37 bacek_in_sing joined #parrot
07:37 dafrito cotto, http://files.dafrito.com/revmap
07:42 cotto thanks
07:43 dafrito dukeleto: Will you provide a SVN tarball for your parrot/parrot repo, like the other ones you've provided on the wiki?
07:46 cotto If it's on github, there's no reason why a simple clone wouldn't be sufficient, is there?
07:46 lucian joined #parrot
07:48 dafrito I don't think a clone would get us the SVN metadata, which means our revmap won't work
07:49 dafrito we can regenerate it, though, so it's not a big problem. It's just a bit time-consuming
07:54 dafrito I wrote a quick cronjob to regenerate http://files.dafrito.com/revmap every 20 minutes, so it should be up-to-date most of the time
07:56 cotto It's not vital to have an up-to-the-minute copy.
08:03 fperrad joined #parrot
08:13 Casan joined #parrot
08:30 dalek parrot: r48576 | NotFound++ | trunk/t/pmc/packfileconstanttable.t:
08:30 dalek parrot: exercise PackfileConstantTable mark vtable in tests
08:30 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48576/
08:47 dalek parrot: r48577 | NotFound++ | trunk/t/pmc/packfileannotations.t:
08:47 dalek parrot: exercise PackfileAnnotations mark vtable in tests
08:47 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48577/
09:09 tadzik joined #parrot
09:14 ilbot2 joined #parrot
09:14 Topic for #parrot is now Parrot 2.7.0 "Australian King" Released! | parrot.org Log:  irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | close 13 tickets, merge outstanding branches, help NotFound clean up PBC tests
09:18 PacoLinux left #parrot
09:31 dalek winxed: r597 | NotFound++ | trunk/winxedst1.winxed:
09:31 dalek winxed: define and use ModifierList constructor in stage 1 compiler
09:31 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=597
09:36 bacek_in_sing left #parrot
09:39 masak joined #parrot
09:58 nik joined #parrot
09:59 nik is now known as nwellnhof
10:06 dalek winxed: r598 | NotFound++ | trunk/winxedst1.winxed:
10:06 dalek winxed: refactor Namespace classes in stage 1 compiler
10:06 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=598
10:12 jan left #parrot
10:20 dalek rakudo: 81b4490 | moritz++ | t/spectest.data:
10:20 dalek rakudo: run our first S19 test file
10:21 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/8​1b4490b73d436dbaeb555eb33de03ae911e224a
10:21 dalek winxed: r599 | NotFound++ | trunk/winxedst1.winxed:
10:21 dalek winxed: emit end libs comment only when needed
10:21 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=599
10:22 tadzik left #parrot
10:26 jsut_ joined #parrot
10:31 jsut left #parrot
10:32 dalek rakudo: 6a3ec07 | moritz++ | docs/release_guide.pod:
10:32 dalek rakudo: [release_guide] volunteers for November (masak++) and Dezember (smash++)
10:32 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/6​a3ec07159f56fa3ca0c7cd280c9353e4d631836
10:32 dalek rakudo: e45bf69 | moritz++ | docs/release_guide.pod:
10:32 dalek rakudo: [release_guide] mathw++ has volunteered for next month; add a few more planned release dates
10:32 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/e​45bf698a51b90073e967ac156db816b00f27b64
11:01 ruoso left #parrot
11:35 dalek parrot: r48578 | NotFound++ | trunk/compilers/imcc/parser_util.c:
11:35 dalek parrot: reset context for imcc compiling to default values, TT #1694, plobsing++
11:35 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48578/
11:41 nwellnhof I just ran $ perl tools/dev/mk_manifest_and_skip.pl
11:41 nwellnhof And got "Sorry, this script is not compatible with git-svn"
11:42 nwellnhof How do I update MANIFEST if I'm using git-svn?
11:42 NotFound nwellnhof: by hand
11:43 nwellnhof OK, thanks
11:50 moritz with an svn checkout
11:50 moritz or you could patch tools/dev/mk_manifest_and_skip.pl to work with git-svn
11:51 moritz that would be a worthy contribution
12:06 whiteknight joined #parrot
12:08 whiteknight Chandon++
12:09 dalek parrot: r48579 | nwellnhof++ | trunk (4 files):
12:09 dalek parrot: Account for fixed size pools in memory_allocated
12:09 dalek parrot: Also add tests for TT#945 and TT#1465. Both are still TODO but I
12:09 dalek parrot: have a fix in TT#1737.
12:09 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48579/
12:23 nwellnhof My first real commit!
12:23 moritz congratulations
12:25 NotFound nwellnhof++
12:25 masak nwellnhof++
12:27 whiteknight nwellnhof++
12:27 whiteknight and a commit to the oft-neglected GC no less!
12:42 mikehh left #parrot
12:48 ruoso joined #parrot
12:52 ttbot Parrot trunk/ r48580 MSWin32-x86-multi-thread make error http://tt.taptinder.org/file/cmdout/372242.txt ( http://tt.taptinder.org//bui​ldstatus/pr-Parrot/rp-trunk/ )
12:56 nwellnhof left #parrot
12:59 dalek parrot: r48580 | NotFound++ | trunk (7 files):
12:59 dalek parrot: experimental Parrot_getpid function, TT #1564
12:59 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48580/
12:59 dalek parrot: r48581 | nwellnhof++ | trunk/t/op (2 files):
12:59 dalek parrot: Get GC stats before possibly triggering another GC
12:59 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48581/
12:59 dalek parrot: r48582 | coke++ | trunk/README:
12:59 dalek parrot: Mention what we are before we mention what you need.
12:59 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48582/
13:05 ttbot Parrot trunk/ r48582 MSWin32-x86-multi-thread make error http://tt.taptinder.org/file/cmdout/372298.txt ( http://tt.taptinder.org//bui​ldstatus/pr-Parrot/rp-trunk/ )
13:06 Coke NotFound: that's you in r48580
13:07 NotFound Coke: I'm on it
13:07 Coke config/gen/platform/generic/pid.c(24) : fatal error C1083: Cannot open include file: 'unistd.h': No such file or directory
13:07 Coke NMAKE : fatal error U1077: '"C:\Program Files\Microsoft Visual Studio 9.0\VC\BIN\cl.EXE"' : return code '0x2'
13:07 Coke Stop.
13:07 Coke NotFound++
13:16 dalek parrot: r48583 | NotFound++ | trunk (5 files):
13:16 dalek parrot: add a pid.c for win32 platform and fix some other mistakes from r48580, TT #1564
13:16 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48583/
13:26 uniejo left #parrot
13:27 Coke NotFound++ # fixed.
13:29 Coke ugh. now that I can explore namespaces again in partcl, I can see the litter from the nqp conversion. bah.
13:36 Casan left #parrot
13:39 Coke NotFound: should the functional get_pid given a warning on the experimental nature (rather than just a comment?)
13:40 NotFound Coke: good question
13:40 purl Yeah, it is. I'm stumped.
13:41 Coke NotFound: also, doesn't windows have _getpid() ?
13:41 Coke (looking at the thread on that ticket)
13:41 Coke also, NotFound++. =-)
13:42 NotFound Coke: I don't have a windows system at hand, so I stay on the safe side ;)
13:44 jnthn 'tis the GetCurrentProcessId API call on Windows.
13:45 jnthn #include <Windows.h> / DWORD pid = GetCurrentProcessId();
13:45 nwellnhof joined #parrot
13:46 Paul_the_Greek joined #parrot
13:46 Paul_the_Greek Good morning, folks.
13:47 whiteknight hello Paul_the_Greek
13:47 whiteknight Paul_the_Greek: Where are you located, abouts?
13:48 moritz Greece? :-)
13:49 whiteknight in Greece he isn't "Paul the Greek", he's just "paul"
13:49 whiteknight most people there are "...the Greek"
13:49 whiteknight plus, it isn't morning in Greece right now :)
13:50 dalek parrot: r48584 | NotFound++ | trunk (2 files):
13:50 dalek parrot: add a TODO test for TT #1446
13:50 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48584/
13:50 Paul_the_Greek is now known as Paul_C_Anagnostopou
13:50 Paul_C_Anagnostopou See, the bastard chops it off.
13:51 Paul_C_Anagnostopou is now known as Paul_Anagnostopoulo
13:51 Paul_Anagnostopoulo Even without the initial!
13:51 Paul_Anagnostopoulo is now known as Paul_the_Greek
13:51 Paul_the_Greek Chopping off my name went out with punched cards.
13:51 whiteknight yeah, that must be frustrating
13:52 whiteknight I haven't had that problem since playing the legend of zelda as "Andre"
13:53 Coke Paul_the_Greek: that's why my nick is Coke, btw.
13:53 Paul_the_Greek Doesn't happen too often anymore. I'm easy.
13:53 Coke because "CokeGod" was ETOOLONG.
13:53 Paul_the_Greek How long is your real name, Coke?
13:54 Coke shorter than yours. ;)
13:54 Coke 21 chars with spaces.
13:54 Paul_the_Greek I'm having fun playing with Git, but its messages are a bit too Unixey.
13:55 whiteknight Coke: includes middle name?
13:55 Paul_the_Greek I'm 20 with spaces, but 32 if I use my full middle name.
13:55 Coke hai.
13:55 Coke Paul_the_Greek: yes, full middle name.
13:55 Paul_the_Greek Paul Constantine Anagnostopoulos
13:56 Paul_the_Greek My folks gave me a break with the first name.
13:56 * moritz basically never uses his middle name
13:56 Paul_the_Greek Good database test, though.
13:56 Paul_the_Greek What is it, moriz?
13:56 moritz Paul_the_Greek: "Andreas"
13:56 * Coke wonders if any of his direct relatives still use the cyrillic.
13:57 Paul_the_Greek Ooh, nice.
13:57 moritz .oO( "my last name is a four-letter word!" )
14:06 bubaflub joined #parrot
14:18 smash purl: seen ovid
14:18 purl ovid was last seen on #dbix-class 94 days, 21 hours, 32 minutes and 23 seconds ago, saying: Cheers for the url.  [May 17 16:46:23 2010]
14:19 smash purl: ty
14:19 purl np, dh! hand!
14:20 Coke smash: votings soon?
14:21 smash Coke: yes, i'm hoping tomorrow
14:23 Coke smash++
14:23 Coke HULK?
14:23 purl Hulk SMASH!!!
14:41 dalek rakudo: 7e966fa | moritz++ | src/Perl6/Grammar.pm:
14:41 dalek rakudo: don't carp on smart-matching against types that start with Bool::True::Something
14:41 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/7​e966fae324ce38b67e2f6d4091a813b4f0b8023
14:41 davidfetter joined #parrot
14:50 Paul_the_Greek left #parrot
14:57 dalek parrot: r48585 | nwellnhof++ | trunk (6 files):
14:57 dalek parrot: More adjustments to GC timing
14:57 dalek parrot: Apply the rest of my GC timing changes (TT#1737), but keep GC_ALWAYS_SKIP
14:57 dalek parrot: for constant pools. Decrease the dynamic threshold to 25% from 50%.
14:57 dalek parrot: This also fixes TT#945 and TT#1465.
14:57 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48585/
14:59 Coke left #parrot
15:04 dalek TT #1737 closed by nwellnhof++: Timing of GC runs
15:04 dalek TT #1737: http://trac.parrot.org/parrot/ticket/1737
15:06 Paul_the_Greek joined #parrot
15:08 ash_ joined #parrot
15:09 aloha joined #parrot
15:09 bacek joined #parrot
15:16 jan joined #parrot
15:17 dalek rakudo: 7b00312 | moritz++ | src/Perl6/Grammar.pm:
15:17 dalek rakudo: add a parsing comment, as advised by pmichaud++
15:17 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/7​b003126f10b38322dd2c62d794999fe767f9132
15:18 khairul joined #parrot
15:21 ash_ left #parrot
15:24 theory joined #parrot
15:30 plobsing_work joined #parrot
15:33 Andy left #parrot
15:34 davidfetter left #parrot
15:40 NotFound left #parrot
15:41 aloha left #parrot
15:41 Andy joined #parrot
15:43 bacek left #parrot
15:46 Coke joined #parrot
15:53 Paul_the_Greek left #parrot
15:54 NotFound joined #parrot
16:09 Paul_the_Greek joined #parrot
16:10 Paul_the_Greek left #parrot
16:12 tcurtis joined #parrot
16:12 chromatic joined #parrot
16:13 brianwisti joined #parrot
16:14 whiteknight Only 4 candidates? I thought there were more
16:14 whiteknight I swore I saw somebody nominate chromatic, Coke, and Allison
16:14 smash whiteknight: you need to actually accept your nomination
16:15 davidfetter joined #parrot
16:15 smash so, just the fact of being nominated is not enough to stand for election
16:15 whiteknight oh
16:16 smash whiteknight: anyway let me double check that everyone is listed
16:16 theory left #parrot
16:17 allison whiteknight: aye, we're retiring
16:17 allison whiteknight: (though, of course available as resources and advisors for the next generation)
16:17 smash allison: greetings skipper
16:18 allison smash: hi!
16:21 masak left #parrot
16:30 whiteknight allison: retiring? all three of you?
16:30 whiteknight allison: and are you going to be continuing your role as architect?
16:31 allison whiteknight: yes
16:31 allison We felt like it was time for new blood in the foundation.
16:31 smash retiring only from the board of directors i guess
16:32 allison smash: oh, yes, definitely, only from the board
16:32 * smash nods.
16:32 allison we're all continuing in our project roles
16:32 allison foundation != project
16:32 whiteknight allison: If I had known so many vacancies were opening up, I might have been more aggressive in nominating new people
16:32 whiteknight a board of 4 people (Assuming all nominees are elected) isn't very much
16:33 allison we've only been 5 from the beginning
16:33 whiteknight I thought it was 6?
16:33 allison we added Patrick last year
16:33 smash whiteknight: you'll still have me as a deputy  ;-)
16:34 allison smash: didn't you get nominated to the board?
16:34 * smash shakes his head.
16:35 Coke you had but to ask. =-)
16:35 allison smash: ah, blast, sorry I was going to do it but thought someone else already had :(
16:36 smash Coke, allison: i'm fine just doing the bucocratic work
16:36 smash *burocratic
16:36 Coke whiteknight: I assumed that anyone who was crazy enough to run would be elected. 4 is fine.
16:36 Coke smash: I hate that word. ;)
16:36 Coke (because it's so hard to (@#*&$ spell.)
16:36 whiteknight bureaucratic
16:37 smash Coke: (bureaucratic) a necessary evil
16:37 allison whiteknight: one of the things that needs to happen is for things like the membership committee to be formed
16:37 allison whiteknight: so, it's not the board alone taking care of things
16:38 whiteknight allison: a good written list of thigns that the current board thinks the next board should be doing would be a great start
16:38 whiteknight since i'm sure you all are familiar with the ongoing, unfinished, and unstarted work that the board needs to do
16:39 chromatic tcurtis has passed his GSoC project.
16:39 moritz tcurtis++
16:39 moritz chromatic++ # for mentoring
16:39 moritz I must say that working with tcurtis++ was real fun, and that I got very good feedback from him, and all my feature requests implemented :-)
16:40 allison whiteknight: definitely
16:40 Coke nwellnhof: does 48585 give any speed/memory boost?
16:42 allison darbelo: ping
16:43 tcurtis moritz++ It was very fun having someone actually using something I wrote. :) Thanks for all the feedback you provided.
16:43 Coke allison: you doing school again, or just work now?
16:43 allison Coke: just work now
16:44 whiteknight allison: what was your degree/major?
16:44 Coke that is happy and sad. =-)
16:44 whiteknight (I don't think i ever asked that before)
16:44 allison Coke: I tried to do too much last year
16:44 Coke allison: you!?
16:44 allison Coke: I'll get back to it again, down the road :)
16:44 allison Coke: heh :)
16:44 allison whiteknight: PhD in Computer Science, on concurrency in dynamic language virtual machines
16:45 whiteknight nice. And perhaps a very pertinent subject
16:45 allison Coke: I'm really looking forward to doing just 1 job and Parrot :)
16:46 allison Coke: and, actually, they work pretty well together, as one of the core requirements for the job is to be on IRC 8 hours a day :)
16:46 whiteknight allison: Where were you going to school? I've been looking on and off for a good place to go for my own PhD eventually
16:46 Coke allison: nice gig.
16:46 Coke whiteknight: it was a bit out of the way. :P
16:46 allison whiteknight: I did a year at Bristol, was ultimately planning to finish off at Cambridge
16:47 * Coke has pretty much given up on getting a phd at this point. :(
16:47 allison whiteknight: but, yes, what Coke says :) I'd look someplace closer to home
16:47 Coke or even a masters, sadly.
16:47 whiteknight Coke: I was smart in that regard. I got the MS before the wife and the kid
16:47 Coke has 3 and a half other people to get into college first.
16:47 Coke whiteknight: good move.
16:47 purl checkmate
16:48 Coke purl: AIGH
16:48 purl Coke: what?
16:48 allison Coke: life has a way of intervening. And really, at a certain point in your career, the degree doesn't add much to the resume.
16:48 * Coke ponders tricking allison into hacking on partcl.
16:48 Coke allison: oh, I just want to learn more. :)
16:48 Coke and perhaps, slightly, maybe, say "That's DOCTOR Coleda... to you."
16:48 allison Coke: (re: partcl) heh :)
16:49 smash Coke: hahahahahah
16:49 purl LOLCON 5.5 reached.
16:49 icarroll left #parrot
16:49 icarroll joined #parrot
16:49 allison Dr. Coke
16:49 brianwisti left #parrot
16:49 * Coke goes back to hacking on javascript & css & cold fusion.
16:49 allison maybe we can get you a red telephone box with a white swirl on the side :)
16:49 allison (instead of a blue police box)
16:50 whiteknight Coke: ever think about writing a ColdFusion compiler for parrot?
16:51 Coke oooh, I'd rather have the police box. I have the perfect place in my backyard for it. =-)
16:51 Coke whiteknight: often.
16:51 allison Coke: on learning more, that's what I wanted
16:54 whiteknight allison: on the topic of concurrency, I think we could easily have Chandon's green threads merged to trunk by 2.8 or 2.9, and I am confident we could add in OS thread support by 3.6 at the latest
16:54 whiteknight with developer effort, of course
16:55 allison whiteknight: excellent
16:55 whiteknight 2.9 might be a better estimate since he has some more nits to pick and there are some deprecation issues to combat
16:56 allison Coke: (re: partcl again) tempting, I have to say. I'm trying to resist adding anything new to fill the space I've just freed up for at least 3 months. We'll see how long I last. :)
16:57 cotto_work ~~
16:57 smash allison: don't forget to forward me any interesting research projects
17:00 whiteknight allison: we're obviously going to need to make some tough decisions about data synchronization mechanisms and things of that nature before we have thread support merged
17:09 Coke allison: certainly it's fun enough to play around with and get a handle on nqp and pct. =-)
17:09 Coke I can even tell you what my current blocker is. =-)
17:10 moritz do tell
17:10 theory joined #parrot
17:11 nwellnhof Coke: r48585 should give a memory boost at least in some corner cases
17:12 Coke I am trying to get library/init.tcl running again. (need that to be able to run tcltest) ; it cannot parse $::foo (which is kind of like what you'd expect from perl)
17:12 Coke nwellnhof: +1
17:12 purl 1
17:12 Coke nwellnhof++, even
17:13 allison smash: definitely, I'll get you the list
17:13 Coke so, minor change to grammar, maybe, and on action for variables.
17:13 Coke (yah, looks like $::tcl_library is just treated as a literal string.)
17:13 moritz oh, I thought with "blocker" you mean "nqp/pct/parrot feature you need to progress"
17:14 allison whiteknight: yes, it was important not to try to solve everything in the GSoC project, there's only so much one person can do in a summer
17:14 Coke moritz: oh, if you want one of those...
17:14 allison whiteknight: it means looping back around to expand on it later
17:14 Coke http://wiki.github.com/perl6/nqp-rx/requests
17:15 Coke (though I already added a .split for strings in partcl-nqp)
17:15 allison whiteknight: you might be the best person to take point on that one, a combination of experienced core developer, and the person who's spent the most time working with Chandon
17:15 Coke interpolated strings in .panic is the biggest one atm.
17:15 moritz Coke: I might attack .split, or steal it from partcl-nqp
17:16 Coke http://github.com/partcl/partcl-nq​p/blob/master/src/TclString.pm#L89
17:17 Coke IWBNI if there was a way to say "give me an HLL pmc of this core type" from pir.
17:17 Coke (box only gets you 3 types)
17:18 jnthn Coke: Curiously enough, in my meta-model prototype that's how I implemented my box op.
17:18 jnthn Coke: You pass it the thing to box and the type to box it into.
17:19 Coke so how would box an RPA? (in my case, into a TclList)
17:19 jnthn Coke: Nice to have another data point to suggest that's right. :-)
17:19 Coke jnthn: \o/
17:20 jnthn Coke: I didn't quite work out array-ish things in the prototype yet.
17:20 Coke hokay.
17:21 jnthn Coke: I think it'll end up with having an array representation.
17:21 * Coke wonders how to de-pollute the top level HLL namespace in tcl.
17:21 jnthn And you'd just base your type on that and start of by creating the right thingy in the first place. I dunno how feasible that is for every situation though.
17:21 Coke (all the infrastructure, like the grammar and the regex engine, are exposed in top level hll namespace.
17:23 Coke token identifier { <ident> ** '::' }
17:24 Coke - does that allow "::foo" or "foo::" or just "foo::bar" ?
17:24 Coke (or "::")
17:24 moritz just 'foo' or 'foo::bar'
17:25 Coke ah. there's the grammar bug for $::foo, then. =-)
17:25 Coke (not to mention, tcl allows $foo::::::::::bar. :P
17:25 icarroll left #parrot
17:25 moritz should foo:: be allowed?
17:26 jnthn token identifier { [<ident>?] ** '::' } # maybe ;-)
17:26 moritz '::'? <ident> ** '::'
17:27 jnthn moritz: That won't handle the :::::::::::::::::::::::::::::::::::: case in the middle though :-)
17:27 moritz '::'? <ident> ** ['::'+] ?
17:27 Coke sokay, breaking that out into a <colons> rule. Dnake.
17:29 Coke moritz++
17:29 Coke moritz: want a commit bit? ;)
17:30 particle Coke: you're still asking first? just opt people in, like facebook!
17:30 brianwisti joined #parrot
17:39 smash left #parrot
17:57 nwellnhof I just had a look at TT#1603 and made an interesting discovery...
17:57 nwellnhof No parrot string is ever freed in the first GC run after it's created.
17:58 nwellnhof It takes at least a second GC run to free a string.
17:59 nwellnhof This is because Parrot_gc_new_string_header always sets the live flag.
18:00 nwellnhof And it can potentially waste a lot of memory.
18:00 dukeleto 'ello
18:00 chromatic If we have marking correct, removing that flag should work fine.
18:01 Coke if we don't have marking correct, aren't we scrod anyway?
18:01 nwellnhof I just tried it and all tests pass.
18:01 Coke try testing rakudo with it. =-)
18:01 nwellnhof Coke: It could potentially mask GC bugs.
18:02 chromatic If we don't have marking correct, but it's incorrect such that the STRING outlives the incorrectness, we're okay.
18:04 Coke nwellnhof: the current behavior, or your patched version?
18:05 nwellnhof The current behavior could mask other bugs.
18:06 nwellnhof stringbuilder.pmc also sets the live flag.
18:06 chromatic I excised it from all of the code.  So far so good.
18:07 nwellnhof I also removed it from stringbuilder, running tests now...
18:09 nwellnhof Result: PASS
18:11 chromatic This is our best chance to remove it, release timing wise.
18:11 chromatic Commit away.
18:11 nwellnhof OK
18:13 chromatic If you're still digging into the GC, may I recommend looking at the gc_massacre branch?
18:14 dukeleto ruoso: ping
18:15 nwellnhof I had a quick look at it. There are many nice cleanups.
18:15 dalek parrot: r48586 | nwellnhof++ | trunk/src (3 files):
18:15 dalek parrot: Don't set live flag on new strings
18:15 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48586/
18:23 particle nwellnhof: your parrot cla has been received and reviewed, all's well.
18:24 particle so, you've got your commit bit fair and square now :)
18:24 particle welcome!
18:24 purl Welcome to #perl.  You will never find a more wretched hive of scum and pedantry.
18:26 davidfetter purl, this. is. #parrot
18:26 purl davidfetter: i'm not following you...
18:26 * davidfetter wondering where purl is dining this evening
18:28 * Coke would watch 300 again tonight but just got TOS Season 3 on blu ray.
18:29 * Coke eagerly anticipates the remastered classic of TOS 3.1
18:30 dalek nqp-rx: 8788740 | pmichaud++ | src/Regex/P6Regex/ (2 files):
18:30 dalek nqp-rx: Add parsing for &&.
18:30 dalek nqp-rx: review: http://github.com/perl6/nqp-rx/commit/8​788740b185a5de82dd5819a24b5be426b2879c5
18:30 dalek nqp-rx: 41267b0 | pmichaud++ | / (2 files):
18:30 dalek nqp-rx: Add && conjunctions to regexes.
18:30 dalek nqp-rx: review: http://github.com/perl6/nqp-rx/commit/4​1267b0f913ec91e516384f135795aaaedaaaca6
18:30 dalek nqp-rx: 253aeeb | pmichaud++ | src/stage0/ (4 files):
18:30 dalek nqp-rx: Update bootstrap.
18:30 dalek nqp-rx: review: http://github.com/perl6/nqp-rx/commit/2​53aeeb6c68ad816cf75445eb0403741068788f3
18:30 cotto_work &&?
18:30 purl && is higher precedence than and
18:31 cotto_work What does that do in a regex context?
18:32 dalek parrot-linear-algebra: c171c07 | Whiteknight++ | src/pmc/ (3 files):
18:32 dalek parrot-linear-algebra: several bug fixes to make a non-trivial example work. GEMM calls weren't respecting lazy transposes, so force that to happen in several places. Also both Gemm methods were returning the wrong value and not respecting forced conversions on parameter C
18:32 dalek parrot-linear-algebra: review: http://github.com/Whiteknight/parrot-linear-algebr​a/commit/c171c078e509cfbb639091e2dbcd618d82905cea
18:33 dalek TT #1748 created by brianwisti++: http://parrot.org/release/developer points to 2.6.0 rather than 2.7.0
18:33 dalek TT #1748: http://trac.parrot.org/parrot/ticket/1748
18:38 cotto_work seen gerd
18:38 purl gerd was last seen on #parrot 66 days, 6 hours, 15 minutes and 30 seconds ago, saying: I will start to copy the tar file to the ftp server.  [Jun 15 12:22:48 2010]
18:39 cotto_work That's not promising, given that he's the next two release managers.  Hopefully he'll show up.
18:42 Coke cotto_work: && is like &, except that it forces left-to-right evaluation.
18:49 ruoso dukeleto, pong
18:50 dukeleto ruoso: gsoc evals are due today at 7pm pacific
18:50 dalek parrot: r48587 | pmichaud++ | trunk/ext/nqp-rx/src/stage0 (4 files):
18:50 dalek parrot: [nqp-rx]:  Update with implementation of && in regexes.
18:50 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48587/
18:50 Coke cotto_work: looks like & is like |, except you have to match both sides at the same time in the same position.
18:51 dukeleto ruoso: can you make sure pawel does it?
18:51 dalek TT #1748 closed by cotto++: http://parrot.org/release/developer points to 2.6.0 rather than 2.7.0
18:51 dalek TT #1748: http://trac.parrot.org/parrot/ticket/1748
18:51 dalek TT #1121 closed by nwellnhof++: system.c trace_mem_block small bug
18:51 dalek TT #1121: http://trac.parrot.org/parrot/ticket/1121
18:51 Coke which sounds useless to me, but perhaps for something like : { <uppercase> & <vowel> } that works.
18:51 dukeleto ruoso: he has not done his eval yet
18:52 dukeleto msg plobsing_work GSoC evals are due today at 7pm UTC
18:52 purl Message for plobsing_work stored.
18:53 dukeleto ruoso: it is actually 7pm UTC!
18:53 dukeleto ruoso: so they are actually already late! darn.
18:53 dukeleto ruoso: if you can get him to do it soon, that would be awesome
18:53 dukeleto allison: ping! could you do your GSoC mentor eval? I think they are just slightly late now, but are still editable
18:54 dalek nqp-rx: b139200 | moritz++ | / (2 files):
18:54 dalek nqp-rx: [setting] implement split(), mostly stolen from partcl-nqp, Coke++
18:54 dalek nqp-rx: review: http://github.com/perl6/nqp-rx/commit/b​1392000d6b6e81aefaee9ee597b724e3061ea96
18:55 Coke \o/
18:57 dukeleto msg allison please submit your gsoc eval ASAP, thanks!
18:57 purl Message for allison stored.
18:57 ruoso dukeleto, hm... he said to me he did fill the final eval already
18:58 Coke moritz: if there are 4 chunks in that nqp-rx test, why are there only two |'s when you join it?
18:58 dukeleto ruoso: my list says that he didn't. can you ask him to resubmit?
18:59 ruoso he doesn't seem to be online right now...
18:59 plobsing_work dukeleto: aaaaaaaah!!! submitted.
19:00 allison dukeleto: I have it nearly finished, I though I had until 19:00?
19:00 dukeleto plobsing_work: thanks :)
19:00 dukeleto allison: 7pm UTC, which was 2 hrs ago, i think
19:01 cotto_work clock?
19:01 purl cotto_work: LAX: Fri 12:01pm PDT / CHI: Fri 2:01pm CDT / NYC: Fri 3:01pm EDT / LON: Fri 8:01pm BST / BER: Fri 9:01pm CEST / IND: Sat 12:31am IST / TOK: Sat 4:01am JST / SYD: Sat 5:01am EST /
19:01 allison dukeleto: crap, that's unhelpful of them
19:01 plobsing_work It's noon here and I thought I was at UTC-7
19:01 allison it says "7:00pm" at the top of the page, nothing about UTC
19:01 cotto_work 1900 UTC was a few minutes ago according to date --utc
19:01 allison so, I assumed "Google time"
19:02 dukeleto i only know this because my irc client shows UTC time
19:02 ruoso they usually inform time in california time
19:02 * tcurtis wishes his date command understood --utc.
19:02 dukeleto http://socghop.appspot.com/document/sho​w/gsoc_program/google/gsoc2010/timeline <-- this says 19:00 UTC
19:02 ruoso okay...
19:03 * dukeleto assumed it was 7p pacific, until he looked again
19:05 allison dukeleto: just tried to submit and they wouldn't let me
19:05 allison I get an error "this page is inactive"
19:05 dukeleto damn them.
19:05 Coke tcurtis: "TZ=utc date"
19:06 tcurtis Coke: thanks.
19:07 dukeleto allison: did it tell you to email a certain address or anything?
19:07 allison dukeleto: nope, just a few bare red words on an empty page
19:08 lucian_ joined #parrot
19:08 allison ah, I see it does say "UTC" in the fine print
19:09 lucian left #parrot
19:10 pmichaud 18:51 <Coke> which sounds useless to me, but perhaps for something like : { <uppercase> & <vowel> } that works.
19:10 pmichaud actually, I implemented it because I needed it.
19:10 pmichaud We wanted to be able to detect when the next identifier was exactly 'Bool::True'
19:10 dukeleto allison: i am sending an email to Carol and cc'ing you
19:10 pmichaud so with && it becomes     [ 'Bool::True' && <.longname> ]
19:11 Coke Ah. I would have probably just checked in the action.
19:12 Coke but yah, having it in the grammar itself is nice.
19:12 pmichaud there's not really an action for this
19:12 pmichaud because the thing that follows is a <.panic: ...>
19:12 pmichaud i.e., if we find 'Bool::True' at this point, we throw an error/warning.
19:12 pmichaud (although TimToady++ did mention after this that perhaps it should be a semantic check instead of a syntactic one)
19:14 moritz Coke: 3
19:14 moritz oh
19:14 dukeleto allison: email sent, i asked for an extension for you and pawel
19:15 dukeleto ruoso: please try to ping pawel, if you can
19:15 dukeleto ruoso: is he on irc or IM anywhere?
19:15 moritz Coke: hm, the test passes regardless of the test string... that's bad!
19:15 * moritz looks into it
19:17 moritz nqp: say("a" eq "b", "c")
19:17 p6eval nqp: OUTPUT«0c␤»
19:18 moritz nqp: say("a" eq "a", "c")
19:18 p6eval nqp: OUTPUT«1c␤»
19:18 moritz hum
19:18 moritz Coke: fixed now
19:19 ruoso dukeleto, nope... he's offline
19:19 tcurtis left #parrot
19:21 pmichaud one of the recent parrot changes breaks rakudo
19:21 pmichaud I'll try to bisect the commit
19:21 dukeleto pmichaud: oh noes!
19:21 dukeleto pmichaud: what kind of error do you get?
19:22 moritz pmichaud: changes to the nqp-rx setting don't require updating bootstrap files, right?
19:23 pmichaud moritz: well, you'd want to update the setting .pir file, I suspect.
19:23 dalek nqp-rx: 4d31c5b | moritz++ | t/setting/04-regex.t:
19:23 dalek nqp-rx: precedence fix for regex test, Coke++
19:23 dalek nqp-rx: review: http://github.com/perl6/nqp-rx/commit/4​d31c5b84a3518d9e1341dd42251ed959b9e9fdc
19:26 moritz pmichaud: how is nqp-setting.pir generated?
19:26 pmichaud moritz: by nqp-rx's makefile, iirc
19:26 pmichaud and then it's copied into src/stage0, same as the other bootstrap files.
19:27 pmichaud I might be wrong about that.
19:28 moritz pmichaud: in the nqp-rx repo, neither 'make' nor 'make bootstrap-files' generates it... unless my Makefile is out of date
19:29 * moritz tries to reconfigure
19:30 pmichaud moritz: checking.  it's possible that it's created from nqp-setting as part of the parrot build process.
19:31 moritz pmichaud: ah, that's possible
19:31 Coke moritz++
19:42 moritz rakudo on svn HEAD says
19:42 moritz PackFile_unpack: This Parrot cannot read bytecode files with version 8.2.
19:42 moritz PackFile header failed during unpack
19:43 cotto_work I bumper PBC_COMPAT last night.  You probably need to clear out all old pbc files.
19:43 cotto_work s/bumper/bumped/
19:44 moritz trying rm -rf parrot_install now
19:50 moritz cotto_work: would it be very hard to tell which file produces that error?
19:51 chromatic That seems doable, moritz.
19:51 moritz that would be great
19:51 chromatic Give me a few minutes and I can look at it.
19:51 * moritz gives chromatic a few minutes
19:51 cotto_work moritz: any pbc files
19:52 moritz cotto_work: but which?
19:52 purl which is probably bad because it means that a role can generate different methods for different classes or in the main pod: http://search.cpan.org/~frew/DBIx-Class-0.0​8121/lib/DBIx/Class.pm#GETTING_HELP/SUPPORT or silly, since Time::Piece does some of that and then goes funky with timezones or drew's domain registrar. and not the cheapest one either. or very helpful to get rid of lag or epic
19:52 moritz purl: forget which
19:52 purl moritz: I forgot which
19:52 Coke moritz: looks like not easily.
19:52 Coke at the point that is called, we've already sucked it into memory.
19:53 Coke and are working on a Packefile_Header which doesn't contain a filename.
19:53 jnthn You could keep the filename in a contextual var...oh, wait.
19:53 Coke (this is fixable, of course.)
19:54 Coke src/packfile.c:1074
19:54 jnthn .oO( right tool, wrong language...)
19:54 jnthn Coke: Not sure the filename belongs in the header structure
19:55 jnthn Coke: Since it shouldn't really be in the file either...and is mis-leading if a file gets renamed or relocated or some such
19:55 jnthn There should be some place to hang it though.
19:55 * jnthn isn't sure how far along the packfile refactoring towards PMCs got.
19:55 whiteknight left #parrot
19:56 Coke jnthn: me either. I just did 10 seconds of ack'ing.
20:11 fperrad left #parrot
20:13 theory_ joined #parrot
20:13 theory left #parrot
20:13 theory_ is now known as theory
20:19 jsut joined #parrot
20:19 jsut_ left #parrot
20:27 chromatic left #parrot
20:31 theory left #parrot
20:32 theory joined #parrot
20:34 nopaste "coke" at 192.168.1.3 pasted "do-nothing jqgrid?" (18 lines) at http://nopaste.snit.ch/22949
20:48 Paul_the_Greek joined #parrot
20:48 Paul_the_Greek Say, do we have Parrot t-shirts or mugs?
20:51 perlite left #parrot
20:56 perlite joined #parrot
21:18 tcurtis joined #parrot
21:20 allison Paul_the_Greek: not with the new logo
21:21 Paul_the_Greek Time for news t-shirts or mugs.
21:24 cotto_work dafrito: Sorry I got distracted last night.  I don't mind the extra admin stuff as long as it stays out of the way.
21:24 cotto_work Feel free to merge.
21:30 Coke whoops, that nopaste was not meant to go here. sorry. =-)
21:31 dafrito cotto_work: Sure, I'll probably squash the branch and merge it. It's purely an optional testing thing; it's not intended to replace other build processes in the repo
21:35 ruoso left #parrot
21:37 cotto_work dafrito: can you add svn revision wiki syntax to TODO?
21:38 cotto_work That'll need to be done manually since there won't be an actual svn vc backend.
21:40 dafrito cotto_work: done
21:42 cotto_work also, a graceful uninstall (though I don't know if Trac provides a way to do that)
21:42 cotto_work thanks
21:43 dafrito You mean like, cleaning up the tables and stuff?
21:43 cotto_work yes
21:44 cotto_work I hate leaving junk behind if it's avoidable
21:44 dafrito hm, lemme look into it. You'd hope they would
21:44 cotto_work You'd hope.
21:46 cotto_work I'm a little surprised that the plugin doesn't already provide wiki syntax.  It'll be easy to add.
21:47 bubaflub left #parrot
21:48 cotto_work Do you have any idea why there's no "browser" link in the navbar?
21:48 cotto_work I suspect that it's a config issue.
21:49 dafrito I don't, but yeah, that's my thought
21:50 cotto_work Once the code is ready, I want to make sure everything's well documented for the osuosl guys.
21:52 cotto_work This may be what we want: http://trac-hacks.org/wiki/NavAddPlugin
21:56 lucian_ left #parrot
21:58 nwellnhof left #parrot
22:16 Andy left #parrot
22:22 darbelo joined #parrot
22:26 arnsholt In PAST, can I call an get an immediate Block to be called with an argument?
22:29 dalek github-trac: edf3697 | dafrito++ | TODO:
22:29 dalek github-trac: Added SVN revision wiki syntax to the TODO
22:29 dalek github-trac: review: http://github.com/cotto/github-trac/commit​/edf36971d91b0e1fb7349e3cfe81ce5f216690bd
22:29 dalek github-trac: c88388d | dafrito++ | / (3 files):
22:29 dalek github-trac: Added admin tool to setup a virtualenv
22:29 dalek github-trac: review: http://github.com/cotto/github-trac/commit​/c88388d83d6fadf3b320345339b94cf7ecf1cb12
22:32 cotto_work dukeleto: parrot/parrot only shows one branch.  Is that intentional?
22:32 cotto_work (on github)
22:33 brianwisti left #parrot
22:35 darbelo It seems to be missing older commits, if that extends to the ones that created the branches they won't show up.
22:36 cotto_work http://github.com/parrot/parrot/commit/e​0d9c4d1cfe9b4456637d0f32d8feacc535a53bb is pretty old (r1234)
22:37 darbelo Hm. I tried to paginate through github's "commit history" and failed.
22:38 darbelo The tags are also missing.
22:38 purl okay, darbelo.
22:38 darbelo purl: the tags?
22:38 purl the tags are a must have.
22:46 Paul_the_Greek left #parrot
22:51 dukeleto cotto_work: still working on getting all the other branches there, not intentional
22:51 dukeleto darbelo: haven't worked on tags/branches yet
22:54 cotto_work ok
22:57 * dukeleto will work on the github mirror later tonight/this weekend
22:57 * dukeleto just updated master, tho
23:00 darbelo left #parrot
23:01 Paul_the_Greek joined #parrot
23:23 cotto_work left #parrot
23:27 cotto_work joined #parrot
23:45 davidfetter left #parrot

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

Parrot | source cross referenced