Camelia, the Perl 6 bug

IRC log for #parrot, 2008-09-12

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 GeJ joined #parrot
00:03 mberends joined #parrot
00:05 mberends joined #parrot
00:09 AndyA joined #parrot
00:10 wknight8111 I've looked at Mastering Algorithms with Perl too
00:10 japhb joined #parrot
00:10 wknight8111 and The Art of Agile Development (don't tell chromatic!)
00:11 slavorg joined #parrot
00:13 pmichaud I sold most of my library when I moved out of Corpus Christi...  :-|
00:14 particle it'll probably be very wet soon anyway
00:14 pmichaud indeed.
00:14 pmichaud it took me two trips to the used book store to get rid of them all.
00:15 pmichaud oh well.  It did help pare down the library a bit.  Now I need to get more books, but I'm having trouble finding any that I really like (programming or otherwise)
00:16 wknight8111 pmichaud, you have any good recommendations?
00:17 * particle referenced "programming the perl dbi" today
00:17 particle i do really like "perl testing: a developer's notebook"
00:17 wknight8111 i've never even heard of that
00:18 particle ian and chromatic did a good job with that
00:19 wknight8111 i need to get me a job at O'Reilly and write books all the time. That would be awesome
00:20 particle they're looking for technical authors
00:20 particle talk to c
00:20 wknight8111 yeah, i'm sure he's not tired yet of getting emails from me :)
00:25 particle ok, i've gotta go ride my bike. it's way to nice out
00:25 GeJ I got Perl Testing almost as soon it was published. Good book, although it wasn't really what I was expecting.
00:26 GeJ I was more looking for testing methodology, and Perl Testing is focusing on the CPAN modules available in the Test namespace.
00:27 GeJ So I got TDD from Kent Beck to get the theory, and regularly use Perl Testing for the tools.
00:31 GeJ is there still a Learning Perl 6 in the works?
00:32 wknight8111 I haven't heard anything abou tit
00:32 wknight8111 I'm actually getting ready to start writing my own Perl 6 book
00:33 cotto_work Will it be ready by Christmas?
00:33 GeJ brian_d_foy and Schwern. source was available via svn.
00:33 GeJ But it looks like allison's server doesn't respond to checkout requests anymore.
00:34 tetragon joined #parrot
00:34 Hinrik does anyone know the people behind this one? http://www.amazon.com/Beginning-Perl-Novice-Pr​ofessional/dp/1590598334/ref=pd_bbs_sr_1?ie=UT​F8&s=books&qid=1221179642&sr=8-1
00:35 Hinrik anyone in here, that is
00:36 GeJ Richard Dice is the president of TPF, isn't he?
00:37 pmichaud my wife is asking what book I'll do.  It should probably be parrot compiler tools and/or compiler design (using PCT and Rakudo as a model), but I doubt they'd be big sellers
00:37 pmichaud maybe I'll write the book on Perl 6 regular expressions, though :-)
00:37 GeJ well, I'd buy it
00:38 GeJ scratch that, "them".
00:38 pmichaud Perl 6 regular expressions would be a good book to write
00:38 pmichaud that one would do well, especially if it covered parsing strategies and led into writing compilers
00:38 Hinrik or just add the PCT to p6pe and improve the regex coverage? :P
00:39 pmichaud p6pe is really out of date. :-)
00:39 pmichaud but yes, pct would make for a good new section of p6pe
00:39 pmichaud actually, p6pe should probably just become "Parrot Essentials and the Parrot Compiler Toolkit", leaving the p6 stuff to other books
00:39 Hinrik exactly, that would make it less out of date
00:40 pmichaud so, it would cover NQP (which would be simple to cover) and PCT
00:40 pmichaud and it wouldn't have to have a bunch of Perl 6 stuff in it
00:40 wknight8111 if you need an editor, I have a lot of spare time
00:40 Hinrik it wouldn't be hard to separate, as the perl6/parrot stuff is each given a half of the book as it is
00:40 wknight8111 and a decent vocabulary
00:41 pmichaud maybe I'll just write up NQP, PCT, and basic compiler construction, and then we can see if it's worth its own book or if it should just be a part of a Parrot book (or maybe both)
00:41 bacek joined #parrot
00:41 pmichaud I would think that NQP and PCT would be a book at least the size of "Perl Testing", and roughly in the same genre/category (here's how to use some existing modules)
00:42 wknight8111 Have you seen the Parrot book I've been writing? it has some decent PCT/NQP sections
00:42 Hinrik separating p6pe into two more detailed books sounds pretty reasonable
00:44 wknight8111 "decent", if I may toot my own horn :)
00:44 wolverian joined #parrot
00:44 Hinrik wknight8111: is your work available somehwere?
00:45 dalek joined #parrot
00:46 wknight8111 http://en.wikibooks.org/wiki/Parrot
00:46 dalek r31007 | chromatic++ | trunk:
00:46 dalek : [PMC] Skipped forced recycling of context in RetContinuation PMC.  This seems
00:46 Hinrik nice
00:46 dalek : to work against exception handling, where invoking a RetContinuation to return
00:46 dalek : the point of throw *shouldn't* get rid of an active context.  See RT #58772
00:46 dalek : (Moritz Lenz) for one symptom -- an infinite loop in context marking when
00:46 dalek : reusing the freshly recycled context and setting its caller to itself.  Oops.
00:46 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31007
00:47 cotto_work alright!  The punchline.
00:47 * pmichaud bookmarks for later reading
00:53 grim_fandango joined #parrot
01:01 dalek r31008 | chromatic++ | trunk:
01:01 dalek : [PMC] Made PMCProxy's does() initialize newly-created PMCs before trying to
01:01 dalek : call their vtable entries -- without this, you get segfaults in PMCs such as
01:01 dalek : Class which expect that their instance data actually exists (reported by Clark
01:01 dalek : Cooper in RT #53156).
01:01 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31008
01:03 pmichaud time for soccer, bbl
02:01 grim_fandango joined #parrot
02:34 Tene 'kay, let's see if I screw things up here.
02:34 Tene Committing a merge of master into lazyrakudo.
02:34 Tene s/master/trunk/
02:35 Tene Git didn't mention any conflicts, so I'm pretty sure it went okay.
02:39 Tene Still committing...
02:42 Tene Huh, it's stuck on t/pmc/fixedintegerarray.t
02:42 dmknopp joined #parrot
02:42 Tene hi dmknopp
02:42 dmknopp hello
02:42 purl hi, dmknopp.
02:42 dmknopp hello all
02:42 purl It's a crazy world, but hello to you too!
02:42 * Tene cancels the commit, tries again
02:43 Tene dmknopp: you mail your CLA yet?
02:43 dmknopp it is a crazy world and its been a crazy day!
02:43 dmknopp i havent had time, i am going to send it tomorrow
02:43 dmknopp work has been hectic this week.
02:48 dmknopp so I compiled the gen_*.pir files down to .pbc files and had cardinal.pir load_bytecode on the .pbc files. I had hoped this would help speed up run time, but the difference was negligible
02:48 dmknopp :-/
02:49 dalek r31009 | tene++ | lazyrakudo:
02:49 dalek : Merge branch 'master' into branch 'lazyrakudo'
02:49 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31009
02:49 Tene Yeah, that's automatically done when cardinal.pbc is compiled from cardinal.pir
02:49 Tene So it wouldn't make a difference
02:49 clunker9_ joined #parrot
02:49 dmknopp ahh no wonder
02:49 Tene jonathan: check out r31009 for sanity
02:51 Tene The speed issue is dominated by parsing time.
02:52 Tene The grammar is just very complex, and PGE hasn't had any speed optimization at all.
02:52 dmknopp do you know of any ideas to speed it up?
02:52 dmknopp *nod
02:52 Tene A few ideas.  Simplify the grammar.
02:52 dmknopp hehe
02:52 dmknopp is that an option?
02:52 Tene If you can make things fail early, or prevent backtracking, it can speed it up
02:53 Tene That's not so easy
02:53 dmknopp well i have about ~15 new tests, mostly for strings and most of them are working
02:54 Tene Nice.
02:54 dmknopp I definitly want to get it checked in by next release... next tuesday I think
02:54 Tene Yeah.
02:54 dmknopp i was looking into the Range
02:54 Tene If you can give me a patch, I'll commit them right now.
02:55 dmknopp i m having trouble getting the grammar to parse , right now it just does an addition op code on the range
02:55 dmknopp ohh yeah ill get you a patch
02:55 dmknopp one sec
03:02 s1n particle: i was having trouble setting up an auth.perl.org account, pmichaud told me to talk to you
03:20 Tene Hmm... I still haven't seen that commit hit the list.  Has anyone else?
03:39 dalek r31010 | tene++ | trunk:
03:39 dalek : [cardinal]
03:39 dalek : * Many new tests from dmknopp++ (many of them passing).
03:39 dalek : * Many assorted fixes from dmknopp++.
03:39 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31010
03:46 tetragon joined #parrot
05:27 mberends joined #parrot
06:18 cotto_home Is Igor hanging out here?
06:18 cotto_home aka Rafael Sanchez
06:26 dalek r31011 | tene++ | trunk:
06:26 dalek : Update MANIFEST and svn metadata for new test files for Cardinal.
06:26 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31011
06:29 uniejo joined #parrot
06:53 cotto_home Is newbie guidance best kept off-list?
06:53 moritz @tell wknight8111 in that wikibook you use capitalized paths (./Parrot languaages/Perl6/...) that won't work on linux
06:53 moritz purl, msg wknight8111 in that wikibook you use capitalized paths (./Parrot languaages/Perl6/...) that won't work on linux
06:53 purl Message for wknight8111 stored.
06:57 cotto_home I hope the answer is no.
06:57 moritz cotto_home: don't think so
07:01 moritz purl, msg cjfields I'd say go ahead and commit the tests, and include some comments and a link to the RT and the p6l thread
07:01 purl Message for cjfields stored.
07:02 itegebo while learning how to add new languages, I'm stuck on getting more info about NQP
07:03 itegebo I see compilers/nqp/README.pod
07:04 dalek r31012 | rurban++ | trunk:
07:04 dalek : [pdd] merge r30995 for install draft: added long-term goals and competing layout variants
07:04 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31012
07:04 dalek r31013 | moritz++ | trunk:
07:04 dalek : [cage] fixed codetest failures in cardinal
07:04 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31013
07:04 moritz itegebo: try http://en.wikibooks.org/wiki/Parrot
07:04 moritz itegebo: also it's a subset of Perl 6, so some Perl 6 resources might help you
07:05 itegebo moritz: thanks a lot - got too stuck in the distro files
07:05 itegebo that wikibook was exactly what i was looking for
07:06 * moritz only found it today too ;)
07:07 cotto_home msg kid51 http://rt.perl.org/rt3/Tic​ket/Display.html?id=44041 would be another good ticket for one of the new contributors.
07:07 purl Message for kid51 stored.
07:21 iblechbot joined #parrot
07:45 tetragon joined #parrot
07:54 mberends joined #parrot
08:03 masak joined #parrot
08:11 Ademan joined #parrot
08:31 dalek r31014 | tene++ | trunk:
08:31 dalek : [cardinal]
08:31 dalek : * Clean up infix:==(Bool,Bool)
08:31 dalek : * Remove some duplicate code
08:31 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31014
08:31 dalek r31015 | tene++ | trunk:
08:31 dalek : [cardinal]
08:31 dalek : * Fix chomp, chop, chomp!, chop!, and the test for these.
08:31 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31015
08:32 Infinoid joined #parrot
08:34 masak Tene: how feature-complete is cardinal?
08:35 masak ...in terms of the smallest important thing that can't be done yet :)
09:03 barney joined #parrot
09:40 tetragon joined #parrot
10:07 dalek r31016 | allison++ | pdd27mmd:
10:07 dalek : [pdd27mmd] Eliminate morphing in BigInt's 'add'.
10:07 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31016
10:13 bacek joined #parrot
10:13 bacek g'localtime
10:19 bacek joined #parrot
10:39 dalek r31017 | allison++ | pdd27mmd:
10:39 dalek : [pdd27mmd] Add an extra check whether CPointer's pointer has been set, before
10:39 dalek : deciding whether to mark it.
10:39 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31017
11:04 moritz uhm, is anbody trying to fix the broken svn?
11:05 dalek r31018 | moritz++ | trunk:
11:05 dalek : [rakudo] added tests for literals with underscores to spectest_regression
11:05 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31018
11:19 dalek r31019 | bernhard++ | trunk:
11:19 dalek : [codingstd] Satisfy pdd_format.t.
11:19 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31019
11:20 dmknopp left #parrot
11:22 bacek joined #parrot
11:25 Tene moritz: svn is broken?
11:25 Tene masak: cardinal parsing is very slow
11:25 masak but that's more of a compile-time problem, right?
11:25 Tene masak: all of the builtin methods on the 'Array' class are complete, afaik.
11:26 Tene String is closer.  Hash is not very complete.
11:26 dalek r31020 | bernhard++ | trunk:
11:26 dalek : [codingstd] Eliminate a POD error and some warnings.
11:26 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31020
11:26 moritz Tene: every time I svn ci I get an error "200 OK"
11:26 moritz Tene: the commit does make it to the server, though
11:26 barney moritz: same here
11:26 Tene In terms of syntax... we don't handle symbols (I don't know what they are, really), multiple assignment, splat...
11:27 Tene masak: passing blocks to functions needs to be reworked a bit before I can add support for 'yield'
11:27 Tene I need to complete the issues around blocks, procs, lambdas, etc.
11:28 masak sounds promising, anyway
11:28 Tene masak: mostly, look in languages/cardinal/t.  Very little beyond what's there will work.
11:28 masak oki
11:29 masak as it should be :)
11:29 Tene The biggest thing that needs doing is adding methods to the standard classes.
11:29 Tene And getting tests.
11:29 Tene I should try to work towards running the rubyspec tests.
11:29 masak mm
11:30 kid51 joined #parrot
11:32 Tene masak: Also, cardinal supports loading external and precompiled libraries.
11:32 masak useful.
11:32 tetragon joined #parrot
11:33 Tene I think that only cardinal and rakudo do right now.
11:37 Tene I think I can get t/assignment and t/yield passing tomorrow afternoon
11:38 moritz so the two might be good candidates for cross HLL interoperation
11:40 Tene Yes.
11:41 Tene I'm going to try working towards HLL support in PCT on Sunday.
11:41 Tene maybe saturday
11:44 masak would be cool to see some Ruby/Perl6 hybrid code :)
11:45 masak rubycamels!
11:52 Tene Hmm.  I have an implementation of gather/take.
11:52 Tene But it's 6AM.
11:53 masak show us! :)
11:54 nopaste "tene" at 166.70.38.237 pasted "gather/take in pir for masak" (50 lines) at http://nopaste.snit.ch/14031
11:55 masak cool.
11:55 tetragon joined #parrot
11:56 masak if I read it correctly, it already uses continuations
11:56 Tene To actually add that to rakudo, I really should set up an exception type specifically for take
11:56 Tene Yeah.
11:56 masak right now, to have gather/take in _any_ form would be sort of a wet dream
11:57 Tene Yeah?
11:57 Tene Lemme see what I can do.
11:57 masak yes.
11:57 masak \o/
11:57 masak Tene++
11:59 Tene If I fall asleep at work tomorrow, I'm blaming you.
12:03 masak if so, it was for a good cause :)
12:04 Tene Okay, let's see what happens...
12:10 * kid51 goes to $job
12:10 Tene Hmm.  I need to figure out how to pass the block itself as an argument to 'gather' instead of whatever it returns.  Reading the PAST docs...
12:13 rurban joined #parrot
12:15 Tene masak: give me an example to test.
12:15 Tene It looks like this works.
12:15 masak Tene: what do you need, apart from what's in t/ ?
12:15 Tene fine, I'll look through t/ for something.
12:15 Tene Jerk
12:15 Tene Making me do all the hard work.
12:15 masak sorry :)
12:16 masak I'm busy...
12:16 Tene You're fine.  I'm just teasing.
12:16 masak I know
12:16 masak you've done it before :)
12:16 Tene There doesn't seem to be a test for gather/take in t/spec
12:16 masak that's probably true
12:17 masak since there has been no need to move it there
12:17 Tene lemme look in pugs' t/
12:17 masak alright, an example: my @a = gather { say "2"; take 3 }; say "1"; say @a[0]
12:18 Tene Or maybe I just haven't run spectest_regression in too long.
12:18 Tene t/spec/S04-statements/gather.t
12:18 masak ah, there we go.
12:24 Tene Yeah, most of gather.t passes
12:25 Tene Is there a make target to run the spectests and tell me how many unexpected successes there are?
12:25 masak not a make target, but some script, I believe
12:25 * masak goes looking
12:25 Tene I bet moritz knows.  He's not here, though.
12:26 masak perl tools/test_summary.pl
12:26 Tene Thanks.
12:27 masak though over here it produces quite chaotic output right now
12:27 masak many new "uninitialized value" warnings and a few Null PMC Access errors
12:28 Tene S04-statements/gather.rakudo.​.........................  12   16    7    4    0    5
12:28 braceta joined #parrot
12:29 masak I don't know what that means offhand :)
12:29 masak better peruse source
12:30 Tene 160 test files                                      4816 4887 3314   10  286 1277
12:31 Tene Looks like it doesn't pass completely because of array stringification issues and because it's not lazy.
12:31 masak still VERY good news, by me
12:31 masak Tene++
12:32 nopaste "tene" at 166.70.38.237 pasted "gather/take patch for rakudo. needs review" (102 lines) at http://nopaste.snit.ch/14032
12:33 Tene I don't quite feel comfortable committing it without review from pmichaud.
12:36 masak 's ok. cool, a gather/take implementation!
12:36 * masak is all giddy
12:42 Tene Yeah, it is pretty cool.
12:42 Tene Maybe I should add it to cardinal before adding it to rakudo.  ;)
12:42 Tene anyway, sleep now.  srsly.
12:43 masak well, by patching early, I can use it right away. that's all that matters :)
12:43 masak Tene: sleep well.
12:44 Tene msg pmichaud please review the patch at http://nopaste.snit.ch/14032
12:44 purl Message for pmichaud stored.
12:44 Tene purl++
12:56 dalek r31021 | allison++ | pdd27mmd:
12:56 dalek : [pdd27mmd] Refactor out common code for inserting multi in global MULTI
12:56 dalek : namespace.
12:56 purl somebody said : namespace was now one word, not two.
12:56 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31021
13:02 dalek r31022 | rurban++ | cygwin070patches:
13:02 dalek : [pdd] merge from trunk, fix perlcritic.t
13:02 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31022
13:08 iblechbot joined #parrot
13:19 dalek r31023 | rurban++ | trunk:
13:19 dalek : [pdd] make all pdds pass t/codingstd/perlcritic.t and pdd_format.t: trailing whitespace, Missing properly located perl coda, Hard tabs, line length
13:19 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31023
13:24 dalek r31024 | rurban++ | cygwin070patches:
13:24 dalek : [pdd] merge from trunk, fix more perlcritic.t and pdd_format.t
13:24 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31024
13:32 pako joined #parrot
13:32 pako left #parrot
13:41 rurban Anyone knows why languages/eclectus/lib/Parrot/Test/Eclectus.pm was removed?
13:42 rurban I see in the main log. barney "Remove unused Perl 5 module Parrot::Test::Eclectus"
13:49 NotFound rurban: Maybe because is unused?
13:50 barney Yes, it wasn't used. Parrot::Test::Harness does the job with the 'exec' option
13:50 rurban Just found out how to find such svn rm actions in the svn logs. svn log languages/eclectus told it.
13:51 rurban I'm in the state of merging cygwin070patches to trunk. well, the other way round first.
13:51 gryphon joined #parrot
14:06 dalek r31025 | julianalbo++ | trunk:
14:06 dalek : fix for Parrot_add_library_path
14:06 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31025
14:16 grim_fandango joined #parrot
14:41 rurban NotFound: see my fix wrong Parrot_locate_runtime_file documentation
14:41 dalek r31026 | rurban++ | trunk:
14:41 dalek : fix wrong Parrot_locate_runtime_file documentation
14:41 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31026
14:49 Ademan joined #parrot
14:52 dalek r31027 | rurban++ | trunk:
14:52 dalek : [perl6] fix perl6doc usage message
14:52 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31027
14:56 dalek r31028 | rurban++ | trunk:
14:56 dalek : [pipp] do not load config.pbc twice
14:56 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31028
15:15 rurban NotFound: strange HEADERIZER error with your  Parrot_add_library_path:
15:16 nopaste "rurban" at 212.183.84.235 pasted "src/library.c:511: error: conflicting types for 'Parrot_add_library_path" (24 lines) at http://nopaste.snit.ch/14034
15:17 rurban Il retry now with a make clean.
15:18 rurban still the same. gcc-3.4.4
15:21 rurban I'm lost now.
15:37 rurban Oops, forgot to svn up for r31025. thanks.
15:37 dalek r31029 | rurban++ | trunk:
15:37 dalek : RT#48260: add get_search_paths docs
15:37 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31029
15:39 Theory joined #parrot
15:44 rurban My next todo is make test with an already installed libparrot. This is a big one...
15:44 rurban Requires a ticket probably.
15:45 NotFound rurban: I don't have that problem. Maybe an svn failure, try to update.
15:46 rurban I forgot to svn up in include only in src, that's why
15:46 rurban It's fixed for me now.
15:46 NotFound Ok
15:46 rurban I just added a short doc patch to yours
15:54 NotFound rurban: I dislike the black magic approch of those functions, but is better to have it documented, of course.
15:55 rurban Ah, trailing whitspace in src/library.c 518
15:57 dalek r31030 | rurban++ | trunk:
15:57 dalek : [core] trailing space in library.c
15:57 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31030
16:00 rurban kid51: set P0["configuration_steps"], "ARRAY(0x17faa70)" doesn't seem right to me
16:01 NotFound rurban: there is a problem with load_bytecode, by the way. If I follow your suggestion of using 'Xlib' instead of 'Xlib.pbc', and the .pbc is present, parrot try to use the .pbc as source.
16:02 peepsalot joined #parrot
16:02 rurban bad. this should make a ticket
16:02 NotFound The reason is that the logic is: if it has no ".pbc" at the end, then is source.
16:03 rurban load_bytecode may load pir and pasm source, but on a .pbc extension it should detect the bytecode
16:03 NotFound But it checks the extension *before* of searching for the file.
16:04 rurban At least that is my opinion, and it would make life uch easier.
16:04 rurban #ps or ticket?
16:05 rurban I have to fix a pmc/object.pmc error in my tree first and then I can write a patch.
16:05 NotFound IMO load_bytecode must be bytecode, by simple logic of the name. No change of search paths just because of file extension.
16:05 NotFound What to do when not found, must be chooser later.
16:06 NotFound If the main parrot executable wants to use another search logic, must do it by himself.
16:06 particle parrot should not rely on file extension to determine filetype
16:07 NotFound particle: actually, it does.
16:07 particle i know it does
16:07 particle it should not
16:07 NotFound But the problem I see now is not file type, but search path choosed.
16:07 particle this part of parrot is still in the prototype phase
16:08 rurban The detection by extension is IMHO not so bad as the not-checking at all.
16:08 NotFound It choose search path depending of the filename extension.
16:08 NotFound load_bytecode actually select what it wants to do, by his name.
16:09 NotFound The name of the opcode, I mean.
16:16 rurban pmc/object.pmc error found and fixed. All good, but nci. So over to the next library problem...
16:26 rurban One memory concern: Do we really need the huge generated src/nci.c? statically provide methods for all sigs seems to be a huge heap overkill tome. Normally one does generate the nci call on the fly (FFI)
16:26 particle that's what compilers/ncigen/ does
16:26 particle src/nci.c is due to be deprecated
16:26 particle tewk needs to better document ncigen
16:27 particle so people like notfound can start to use it instead
16:27 rurban ncigen does it dynamically? I thought is is till stazic, but just C parsing sugar
16:28 rurban C => PIR
16:29 particle yes, it's dynamic
16:29 rurban I see. It generated an ffi from c sources, and does the dlfunc calls dynamically. good
16:29 particle you no longer need call_list.txt
16:30 rurban did someone tested ncigen against nci? memory vs speed would be interesting
16:30 rurban (uh, my english gets worse and worse)
16:31 particle no, that hasn't been tested yet
16:31 particle i agree, it would be interesting
16:32 rurban I assume it will be a little slower (5%), but need about 30% less memory
16:35 rurban NotFound: But the library.c docs say it shoudl work like discussed. "start guessing now. this version tries to find the lowest form of the code, starting with bytecode and working up to PIR"
16:36 rurban And the series of extensions it tries is correct. First .pbc then pasm past pir
16:41 dalek r31031 | rurban++ | trunk:
16:41 dalek : [core] improve try_bytecode_extensions docs
16:41 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31031
16:49 particle rurban: what's a .past file? there is no such thing
16:50 rurban I really have no idea. leo around?
16:50 particle whatever it was, it's no more
16:51 rurban Do you want a ticket? I dont want to write it because I respect the great past
16:51 rurban :)
16:52 particle sure. i've been around with parrot since 2002, so i know it's not been used for a *long* time
16:52 rurban Ok. I'll add a ticket and patch to remove it.
16:54 mberends all: Pod6Parser at http://nopaste.snit.ch/14035 and Pod6_to_man at http://nopaste.snit.ch/14036
16:56 cotto_work oooooh.  new shiny!
16:58 mberends Pod6_to_xhtml coming tomorrow. Tested only on Debian sid.
17:00 nopaste "rurban" at 212.183.84.235 pasted "perlcritic problem with c-file-style: "parrot" on library.c" (20 lines) at http://nopaste.snit.ch/14037
17:01 rurban Can I ignore that? It should be fixed imho. It is awful long to wait for a full make test.
17:02 mberends perl6doc script included inside Pod6_to_man.pm - make separate file to run
17:09 davidfetter joined #parrot
17:14 rurban_ joined #parrot
17:15 Ademan joined #parrot
17:16 dalek r31032 | jonathan++ | trunk:
17:16 dalek : [rakudo] Re-enable setting block signature on subs; seems to cause no issues.
17:16 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31032
17:18 dalek r31033 | jonathan++ | trunk:
17:18 dalek : [rakudo] Implementation of =<> that may work. Not completely sure it works, since we never manage to parse this operator yet. Checking it in now, so I don't do it as part of another commit of something else by accident at some point in the future; it'll work once parsing stuff is fixed up.
17:18 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31033
17:19 rurban All tests pass for http://rt.perl.org/rt3/Tic​ket/Display.html?id=58794 May I apply the patch?
17:24 rurban t/native_pbc/header..skipped: ongoing PBC format changes. This should really be settled now.
17:25 tewk joined #parrot
17:42 tewk Remind me how do I tell tweety and piper no to bother me when I join?
17:44 nopaste "rurban" at 212.183.53.22 pasted "t/native_pbc/header.t" (32 lines) at http://nopaste.snit.ch/14038
17:44 cotto_work tewk, on xchat you can go to Window->Ignore List and add Piper!*@*
17:44 cotto_work not sure about other clients
17:44 dalek r31034 | jonathan++ | trunk:
17:44 dalek : [rakudo] As we aren't smart enough in the compiler yet to sort between class/role types and subset types, do it upon construction of signature objects. This gets signature objects in shape for the multi-dispatcher.
17:44 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31034
17:45 tewk irssi user
17:45 cotto_work In that case I can't help you more than you can help yourself. ;)
17:46 tewk I thought there was a command you could privmsg to the log bots but I don't remeber what it is.
17:50 nopaste "rurban" at 212.183.53.22 pasted "src/packfile.c" (13 lines) at http://nopaste.snit.ch/14039
17:50 Zaba joined #parrot
17:51 dalek r31035 | coke++ | trunk:
17:51 dalek : [rakudo] fix a doc typo.
17:51 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31035
18:01 nopaste "rurban" at 212.183.53.22 pasted "native_pbc_header.patch #2 (strill failing) minor <> major mixed up?" (81 lines) at http://nopaste.snit.ch/14040
18:02 dalek r31036 | jonathan++ | trunk:
18:02 dalek : [rakudo] Trailing whitespace fix.
18:02 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31036
18:02 dalek r31037 | jonathan++ | trunk:
18:02 dalek : [rakudo] Subs with no signature still need to get a Signature object.
18:02 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31037
18:05 nopaste "rurban" at 212.183.53.22 pasted "native_pbc_header.patch #3 this is it finally. looks correct and passes." (58 lines) at http://nopaste.snit.ch/14041
18:10 nopaste "rurban" at 212.183.53.22 pasted "fix make test with already installed libparrot.so|dll" (25 lines) at http://nopaste.snit.ch/14042
18:10 rurban I have to go now, bye.
18:12 particle rurban: perlcritic should *not* test .c files
18:14 rurban I know. But it would be useful for a quick&dirty selftest with single file arg to perlcritic.t
18:14 particle perlcritic.t used to have that option iirc
18:14 rurban I'm just bringing my cygwin070paqtches branch up to the trunk level. All tests still pass.
18:15 particle but probably instead you should call perlcritic directly and point to the conf file and file you want to test
18:15 rurban make test with already installed libparrot shared is also almost ready, just a few tests fail. (nopasted) I'll continue on the weekend on that
18:16 rurban From monday on I'm a full month in Germany without my fast machine, just my laptop, so I want to finish most stuff now.
18:17 particle somebody please give mberends a pugs commit bit next time he surfaces
18:17 particle have him commit these scripts there, rather than nopasting here. i don't know what to do with them.
18:20 dalek r31038 | rurban++ | cygwin070patches:
18:20 dalek : merge with trunk r31020
18:20 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31038
18:23 rurban The svn server seems to be fixed now.
18:33 sjansen joined #parrot
18:44 dalek r31039 | chromatic++ | trunk:
18:44 dalek : [Rakudo] Fixed POD in misc builtins file.
18:44 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31039
18:46 Debolaz joined #parrot
18:50 cotto_work rurban, ping
19:01 sjansen joined #parrot
19:33 itegebo I'm starting to work through the Squaak_Tutorial, as well as reading perl6/doc/design/syn/S05.html - would anyone have any recommendations for learning the NQP rules a la parrot/examples/tutorial/*.pir?  If not, perhaps that is what I'll work on.
19:37 sjansen joined #parrot
19:38 dalek r31040 | allison++ | pdd27mmd:
19:38 dalek : [pdd27mmd] Underscores are no longer part of the multi name, and multis no
19:38 dalek : longer magically adapt the signature of a multi vtable function to take a
19:38 dalek : different number of arguments.
19:38 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31040
19:42 dalek r31041 | allison++ | pdd27mmd:
19:42 dalek : [pdd27mmd] Multis also get added to the namespace/class so they can be invoked
19:42 dalek : as subs and methods.
19:42 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31041
19:53 pako joined #parrot
20:16 jonathan itegebo: I'm not sure there is a good NQP tutorial yet.
20:17 jonathan It's a subset of Perl 6, though I think the NQP docs mostly say what bits of Perl 6, and rely on you knowing what they mean syntactically.
20:17 jonathan If you're interested in working on an NQP tutorial, I'd guess that will be welcomed. :-)
20:22 cognominal joined #parrot
20:28 Tene itegebo: mostly the tutorial is compilers/nqp/t/
20:42 itegebo Tene: I don't see any tests related to rules in there, so perhaps I could add those - but it at least gives me an example of how to play with NQP (thanks)
20:43 itegebo jonathan: how might I go about getting review once I've got something I think is helpful?
20:43 itegebo should I just follow project/committer_guide.pod?
20:45 particle itegebo: yes.
20:46 jonathan itegebo: Ask folks who write quite a bit of NQP what they think. :-) There are various on this channel.
20:46 particle and docs/submissions.pod
20:46 jonathan But of course, sending it to the list is good too. :-)
20:46 jonathan The channel often depends on who is around.
20:47 * jonathan was planning to write lots of code tonight, then went out for a big curry...
20:48 Tene jonathan: I wrote http://nopaste.snit.ch/14032 last night, waiting for pmichaud to come around ot harass him to review it.
20:49 jonathan (harassing pmichaud)++ ;-)
20:50 jonathan +        MISC_TAKE
20:50 jonathan I'm not completely sure we should be adding language-specific stuff to core Parrot.
20:50 jonathan But if more than one langauge is going to use it that you know of, then fine.
20:51 jonathan eh = 0 # work around the workaround
20:51 jonathan Scary?!
20:52 jonathan +    elsif $sym eq 'gather' {
20:52 Tene Yeah, I don't know what the best thing to do about the exception type is.
20:52 jonathan The very long line after this one, would be nicer split over a few, like many other places in actions.pm
20:52 jonathan Though that's a style point rather than a functionality one.
20:53 jonathan And very, very happy to see:
20:53 jonathan +S04-statements/gather.t
20:53 jonathan I don't see anything immediately wrong with it, other than my concern about what to do about the type.
20:54 Tene Is creating a new PAST::Block(:blocktype('declaration')) the right thing to do there?
20:54 jonathan erm
20:54 jonathan Actually, I suspect $past is already a PAST::Block
20:55 jonathan So you just want to do
20:55 jonathan $past.blocktype('declaration')
20:55 jonathan ;
20:55 jonathan On the line before.
20:56 Tene It isn't always.
20:56 jonathan Ah, you're right
20:56 jonathan For gather I think we always want to end up with a block, though.
20:56 jonathan You could check if we have a block and form one if not.
20:57 jonathan if $past.isa(PAST::Block)
20:57 jonathan That would likely be best
20:57 jonathan I do similar for "where" stuff.
21:24 bacek joined #parrot
21:24 cotto_work has anyone built a recent parrot with msvc?
21:26 jonathan cotto_work: Yes.
21:26 jonathan I regularly build with this.
21:26 jonathan I svn up'd and built with it today.
21:26 Tene jonathan: how's the lazyrakudo branch?
21:27 Tene I never saw the commit from my merge make it to the list
21:27 cotto_work do you know enough about it to guess why I'm getting complaints about an unresolved external symbol __environ?
21:30 jonathan Tene: I was dalek report it.
21:31 jonathan Tene: Not had chance to check it yet - if Rakudo builds and passes the sanity tests then you've not made things any worse, loads of spec tests are expected to fail in that branch though.
21:31 jonathan cotto_work: No - what version of MSVC?
21:31 cotto_work 9 aka 2008
21:31 Tene I'll test that right now
21:32 jonathan cotto_work: Ah, I'm building with an older version.
21:32 cotto_work it makes me feel better that at least there's a meaningful difference that's likely to be causing the brokenness
21:33 jonathan C:\Consulting\parrot\trunk\languages\perl6>cl /H
21:33 jonathan Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 12.00.8168 for 80x86
21:34 cotto_work time for some hunting
21:37 cognominal jonathan, I am playing with scheduler and tasks by adding appropriate Perl6 class but pushing a task dies with share_ro() not implemented in class 'PMCProxy'
21:37 cognominal probably a problem with PMCProxy.
21:41 cognominal Task has a share_ro
21:45 jonathan cognominal: Hard to guess at.
21:46 jonathan cognominal: I'd have to see the code.
21:47 cognominal I will send you a tarball
21:47 jonathan OK
22:01 cognominal jonathan, sent
22:02 jonathan merci
22:03 cognominal good night
22:03 purl sleep well too
22:03 * cognominal reads pmcproxy to figure out its behavior
22:09 cotto_work I won't give it away, but there's a surprise ending.
22:14 jonathan Uh-oh... :-)
22:16 * jonathan has lots of failures in his local Rakudo tree
22:17 jonathan ...but hopes for a happy ending.
22:22 mj41_ joined #parrot
22:35 Limbic_Region joined #parrot
23:06 s1n particle: still need your help with auth.perl.org
23:06 Tene jonathan: what are you working on?
23:09 jonathan Tene: Switching Rakudo over to the new Perl 6 MMD algorithm.
23:09 jonathan It essentially works
23:09 jonathan The problem is dealing with objects that fall out of the Perl 6 object hierarchy.
23:09 jonathan Oh, and junctions
23:10 jonathan The new MMD is a bit too correct for other bits of Rakudo. :-)
23:14 tetragon joined #parrot
23:19 rurban cotto_work: you pinged me?
23:20 Tene I was thinking more about exceptions... we could easily set properties on the exception handler object to set what severity and types of exceptions it's intended for.
23:20 Tene and skip it if the exception doesn't match
23:20 Tene I think I'll try implementing that tonight.
23:21 jonathan So we don't have to even invoke it to find out if it can handle it?
23:21 Tene right
23:21 jonathan That'd probably be cheaper than the try and rethrow approach.
23:21 jonathan Woo. We can nearly haz Perl 6 MMD algorithm.
23:21 Tene and simpler, too.  no need to require EHs to have a check/rethrow prelude
23:21 jonathan Yes
23:22 jonathan That makes sense to me.
23:22 cotto_work rurban, I got an answer.
23:22 rurban good, because I want to got to sleep now :)
23:22 rurban bb
23:32 * jonathan make spectest_regression's and hopes that this will be it.
23:43 dalek r31042 | jonathan++ | trunk:
23:43 dalek : [rakudo] Constraints in a signature should actually be represented as an all Junction, not an array.
23:43 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31042
23:45 dalek r31043 | jonathan++ | trunk:
23:45 dalek : [rakudo] The Perl6MultiSub PMC when checking constraints should do ACCEPTS on them, not just invoke them - it's meant to be a Junction.
23:45 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31043
23:47 dalek r31044 | jonathan++ | trunk:
23:47 dalek : [rakudo] While we likely need a real Exception class at some point, we should map them now so .WHAT etc works on $! after a failed try.
23:47 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31044
23:52 bacek joined #parrot
23:53 bacek morning
23:55 jonathan morning bacek
23:56 bacek hi jonathan. How is your russian visa going?
23:56 jonathan Russia have issued me a visa.
23:56 dalek r31045 | jonathan++ | trunk:
23:56 dalek : [rakudo] If we are calling ACCEPTS in a multi-dispatch to do a type-check, then because blocks are not differentiated from regexes at the moment we get exceptions when trying to store $/. This patch wraps that code up in a handler.
23:56 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31045
23:56 bacek jonathan: congratulations ;)
23:56 jonathan Though I had to courier my passport to the UK and pay an agency there to take it to the embassy for me.
23:57 jonathan So, now I can attend Far East Perl and make my Trans-Siberian journey. Got confirmation of those tickets today too. :-)
23:57 drew joined #parrot
23:57 jonathan So all I'm missing now is the Belarus visa so I can attend BY Perl too, but that is currently in hand.
23:57 jonathan So I don't see any problems there.
23:58 bacek me either.
23:58 nopaste joined #parrot
23:58 bacek btw, I have a question about #56468 (replace PMC_blah_val with VTABLE methods)
23:59 bacek VTABLE PMC *clone() {
23:59 bacek PMC * const dest = pmc_new_noinit(INTERP, SELF->vtable->base_type);
23:59 bacek PObj_custom_mark_SET(dest);
23:59 bacek PMC_str_val(dest) = string_copy(INTERP, SELF.get_string());
23:59 bacek return dest;
23:59 jonathan OK, one moment, just putting in a patch I've been working on for hours in. :-)
23:59 kid51 joined #parrot
23:59 bacek Should we replace PMC_str_val(dest) with VTABLE_set_string_native?

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

Parrot | source cross referenced