Parrot 2.1.1 Released! | parrot.org/ | Tasks: PCC deprecations branch, HLL subclassing and MMD branch
Set by moderator on 19 February 2010.
dukeleto Whiteknight: what does "mode-switching cod" taste like? ;) 00:05
Whiteknight dukeleto: very inefficient, and stringy 00:07
dalek rrot: r44334 | mikehh++ | trunk/t/perl/Parrot_Test.t:
trying fix a test that got messed up removing cuddled else
 00:19
rrot: r44335 | mikehh++ | trunk/t/perl/Parrot_Test.t:
fix the test - editor auto removed trailing spaces even when needed
Whiteknight plobsing: do you have an example of a multi-method I could look at? 00:26
dalek rrot: r44336 | mikehh++ | trunk/t/oo (3 files):
fix codetest failures - trailing whitespace
 00:35
Whiteknight I can't find any examples in the repo 00:36
or are you talking about multi vtables? 00:38
00:38 cotto_work joined 00:51 Austin joined
dalek rrot: r44337 | mikehh++ | trunk/src/gc/gc_ms.c:
fix C function docs
 00:52
rrot: r44338 | mikehh++ | trunk/t/op (4 files):
fix codetest failures - trailing whitespace
rrot: r44339 | mikehh++ | trunk/t/compilers/imcc/syn (3 files):
fix codetest failures - trailing whitespace
00:55 abqar joined
Whiteknight plobsing: I definately dont see any cases of METHODs using NCI 00:56
dalek rrot: r44340 | mikehh++ | trunk/t/codingstd (2 files):
fix codetest failures - trailing whitespace
 01:08
rrot: r44341 | mikehh++ | trunk/t/manifest (5 files):
fix codetest failures - trailing whitespace
cotto hio 01:17
Coke plobsing: some components have auto-assign.
dalek rrot: r44342 | mikehh++ | trunk/t/oo (6 files):
fix codetest failures - trailing whitespace
 01:25
rrot: r44343 | mikehh++ | trunk/t/compilers (3 files):
fix codetest failures - trailing whitespace
rrot: r44344 | mikehh++ | trunk/t (2 files):
fix codetest failures - trailing whitespace
Whiteknight mikehh++ 01:38
mikehh hey I still only about half way
I am I should say and I haven't even looked at the perl critic stuff 01:39
Whiteknight an enterprising soul with sed and a list of failures could probably resolve all that pretty quick 01:40
mikehh got to be careful - had some problem,sd with trailing spaces already and my editor is set to automatically remove them 01:41
I run make test every so often as well as the failing codetests 01:42
Whiteknight my editor too
mikehh the main problem was we were not testing the .t files for a lot of the codetests and Coke++ put them in so we got a hugh number of failures 01:45
going to work on it a bit longer and then I need a break 01:46
Whiteknight mikehh++
by the end you'll be able to trade in all your karma for a troll-shaped pencil eraser and a slinky 01:47
mikehh :-}
plobsing Whiteknight: the first case I saw was in src/pmc/packfileconstanttable.c 01:52
Whiteknight: search for "IJPS" 01:53
dalek rrot: r44345 | mikehh++ | trunk/t/library (4 files):
fix codetest failures - trailing whitespace
 01:57
Whiteknight ah, I see it now 02:02
I was searching for METHOD, not MULTI
I really hate pmc2c
mikehh ok that screwed up t/library/getopt_obj.t 02:03
let's see if I can fix it
Whiteknight build fails for me in extra_nci_thunks.nci 02:10
dalek rrot: r44346 | mikehh++ | trunk/t/pmc (10 files):
fix codetest failures - trailing whitespace
 02:14
rrot: r44347 | jkeenan++ | trunk/t/tools/pmc2cutils/05-gen_c.t:
[codingstd] No trailing whitespace.
rrot: r44348 | jkeenan++ | trunk/t/tools/pmc2cutils/04-dump_pmc.t:
[codingstd] No trailing whitespace.
rrot: r44349 | jkeenan++ | trunk/t/steps/init/hints/darwin-01.t:
[codingstd] No trailing whitespace.
rrot: r44350 | jkeenan++ | trunk/t/run/options.t:
[codingstd] No trailing whitespace.
rrot: r44351 | jkeenan++ | trunk/t/perl/Parrot_Test.t:
[codingstd] No trailing whitespace.
rrot: r44352 | jkeenan++ | trunk/t/op/stringu.t:
[codingstd] No trailing whitespace.
rrot: r44353 | jkeenan++ | trunk/t/tools/install (2 files):
[codingstd] No trailing whitespace.
rrot: r44354 | jkeenan++ | trunk/t/op/sysinfo.t:
[codingstd] No trailing whitespace.
kthakore oh wow! 02:17
go kid51
get that ws
get it good
kthakore brings out popcorn 02:18
Whiteknight nevermind, build seems to be working now
kthakore Whiteknight: horray!
Whiteknight nevermind nevermind, it done broked again
kthakore Whiteknight: oh noes
the ups and down of a code repo 02:19
Whiteknight ./parrot_nci_thunk_gen --dynext --output=src/extra_nci_thunks.c <src/nci/extra_thunks.nci
Ignored signature 'vp' on line 83 (previously seen on line 66)
Ignored signature 'ip' on line 194 (previously seen on line 58)
Ignored signature 'pp' on line 199 (previously seen on line 198)
Ignored signature 'ip' on line 200 (previously seen on line 58)
Ignored signature 'ip' on line 201 (previously seen on line 58)
...and it goes on like that for several dozen lines
plobsing Whiteknight: those should only be warnings. I can try to add an option to silence them if you'd like. 02:23
Whiteknight plobsing: is error, build fails
At least, I think it does 02:24
well nevermind, no it isn't broken
I'm retarded
plobsing it is a very noisy tool atm.
mikehh kid51 - t/library/getopt_obj.t requires the trailing whitespace 02:25
Whiteknight why would it possibly require the trailing whitespace? 02:26
mikehh and I don't want to deal with t/examples/pir.t
in CODE section
for example CODE, 'foo is 1', 'bar is ',.., OUT 02:28
dalek rrot: r44355 | jkeenan++ | trunk/t/op/string_cmp.t:
[codingstd] No trailing whitespace.
 02:31
rrot: r44356 | jkeenan++ | trunk/t (2 files):
[codingstd] No trailing whitespace.
rrot: r44357 | mikehh++ | trunk/t/library/getopt_obj.t:
fix failing test - trailing whitespace is required
02:38 tetragon joined
mikehh kid51: t/perl/Parrot_Test.t and I think t/run/options.t 02:38
dalek rrot: r44358 | plobsing++ | trunk/tools/dev/nci_thunk_gen.pir:
Add --no-warn-dups option to silence warning about duplicate signatures.
 02:48
rrot: r44359 | mikehh++ | trunk/t/perl/Parrot_Test.t:
fix failing test - some trailing whitespace is required
kid51 mikehh: Yes, I agree that there are certain files which need to have trailing whitespace 02:58
Added to test's DATA section.
dalek kudo/master: 7c90f14 | pmichaud++ | src/core/ (4 files):
Refactor some methods in/into Iterator/Iterable.
 03:02
rrot: r44360 | plobsing++ | trunk/config/gen/makefiles/root.in:
use --no-warn-dups option in building of extra_nci_thunks dynext library
 03:04
rrot: r44361 | jkeenan++ | branches/rm_cflags/t/steps/auto/warnings-01.t:
Delete tests which pertained to interface since removed. Add tests for
rrot: r44362 | jkeenan++ | trunk/t/codingstd/trailing_space.t:
In coda, identify some files where there is necessarily trailing whitespace.
mikehh ok I've had enough for the moment - will continue after some sleep 03:06
dalek rrot: r44363 | mikehh++ | trunk/t/run/options.t:
fix failing test - some trailing whitespace is required
 03:20
purl i already had it that way, dalek.
03:41 mikehh_ joined 03:44 mikehh_ joined 03:54 janus joined 04:16 mikehh joined 04:20 theory joined 04:34 petdance joined
dalek kudo/master: a1ed7ff | (Solomon Foster)++ | src/core/IO.pm:
Implement simple version of sub lines to get minmax.t working.
 04:40
kudo/master: d6129bc | (Solomon Foster)++ | t/spectest.data:
Turn on minmax.t.
rrot: r44364 | mikehh++ | trunk/t/codingstd/trailing_space.t:
add t/run/options.t to files to skip trailing spaces test
 04:42
rrot: r44365 | mikehh++ | trunk/t/dynoplibs (2 files):
fix pir coda
 04:59
rrot: r44366 | mikehh++ | trunk/t/dynpmc (3 files):
fix pir coda
rrot: r44367 | mikehh++ | trunk/t/compilers/imcc/syn (6 files):
fix pir coda
 05:15
mikehh need sleep - bbl 05:20
05:21 bubaflub joined
dalek rrot: r44368 | mikehh++ | trunk/t (8 files):
fix pir coda
 05:32
06:08 JimmyZ joined 07:02 bacek joined
bacek aloha 07:04
dalek rrot: r44369 | bacek++ | trunk (3 files):
Replace more mem_sys_free with mem_gc_free
 07:10
rrot: r44370 | bacek++ | trunk (2 files):
Made really_destroy_runllop_jumppoints static
07:18 riffraff joined
dalek rrot: r44371 | bacek++ | trunk/src (6 files):
Replace more mem_sys_free with mem_gc_free/mem_internal_free
 07:43
07:44 bacek joined 07:49 JimmyZ joined 07:59 iblechbot joined 08:46 barney joined 08:48 jan joined
dalek rrot: r44372 | bacek++ | trunk/src/interp/inter_create.c:
Finalize GC is last thing before killing interp.
 09:04
rrot: r44373 | bacek++ | trunk/src/hash.c:
Switch hash to GC allocations
rrot: r44374 | bacek++ | trunk (3 files):
Rerun headerizer
kudo/master: 94ada1f | moritz++ | (4 files):
track change of macro names of PObj_active_destroy_ macros in parrot

Also bumps PARROT_REVISION to r44371		 09:16
09:25 kurahaupo joined
bacek seen plobsing 09:26
purl plobsing was last seen on #parrot 7 hours, 1 minutes and 18 seconds ago, saying: it is a very noisy tool atm.
bacek msg plobsing (#903) Yes, I am. But we have to split it into 3 parts: a) deprecate keyed access to aggregate PMCs (Hashes, Arrays), b) provide different API for "op interpinfo"; c) deprecate Keys 09:27
purl Message for plobsing stored.
fperrad ping darbelo 09:31
bacek "Hostname darbelo not found"
dalek rrot: r44375 | bacek++ | trunk (5 files):
Pass interp through ops initialization process. Switch ops deinit to GC allocations
 09:37
rrot: r44376 | bacek++ | branches/boehm_gc_2 (123 files):
Merge branch 'master' into boehm2
kudo/master: e9b6253 | moritz++ | t/spectest.data:
moritz-- does not understand the simple format of t/spectest.data; fixing
 09:39
nopaste "fperrad" at 93.10.255.71 pasted "[dbm-pmcs darbelo] patch of setup.pir with libdetect" (83 lines) at nopaste.snit.ch/19742 09:40
fperrad msg darbelo see nopaste.snit.ch/19742
purl Message for darbelo stored.
09:45 AndyA joined
dalek kudo/master: 2c4f69b | moritz++ | t/spectest.data:
re-enable sub-ref.t
 09:51
10:09 cotto joined
dalek rrot: r44377 | mikehh++ | trunk/t/pmc (7 files):
fix pir coda
 10:59
11:00 lucian joined 11:09 payload joined 11:10 payload joined 11:11 payload1 joined 11:12 payload1 joined, payload joined
dalek rrot: r44378 | mikehh++ | trunk/t/pmc (4 files):
fix pir coda
 11:16
rrot: r44379 | mikehh++ | trunk/t/pmc (6 files):
fix pir coda
11:17 payload1 left
dalek tracwiki: v159 | Austin_Hastings++ | WikiStart 11:28
tracwiki: trac.parrot.org/parrot/wiki/WikiSta...ction=diff
tracwiki: v160 | Austin_Hastings++ | WikiStart
tracwiki: trac.parrot.org/parrot/wiki/WikiSta...ction=diff
rrot: r44380 | mikehh++ | trunk/t/pmc (7 files):
fix pir coda
 11:32
rrot: r44381 | mikehh++ | trunk/t/pmc (9 files):
fix pir coda
rrot: r44382 | mikehh++ | trunk/t/pmc (6 files):
fix pir coda
 11:49
11:50 payload joined, payload1 joined 11:54 lucian joined
dalek kudo/master: 4071ac3 | moritz++ | src/pmc/objectref_pmc.template:
catch another instance of PObj_active_destroy_SET noticed by quester++
 11:57
rrot: r44383 | mikehh++ | trunk/t/library (7 files):
fix pir coda
 12:05
rrot: r44384 | mikehh++ | trunk/t/oo (2 files):
fix pir coda
Austin msg Tene I've put up a wiki page (trac.parrot.org/parrot/wiki/ExceptionRefactor) with some stuff about Exceptions 12:15
purl Message for tene stored.
dalek tracwiki: v1 | Austin_Hastings++ | ExceptionRefactor 12:17
tracwiki: trac.parrot.org/parrot/wiki/Excepti...ction=diff
mikehh phew - only 10 files with pir coda problems left, then of course the perlcritic stuff 12:19
Austin mikehh++ 12:20
I'm glad Coke made that change(s). Thanks for grunting out the cleanup.
dalek rrot: r44385 | mikehh++ | trunk/t/op (5 files):
fix pir coda
 12:21
rrot: r44386 | mikehh++ | trunk/t/op (7 files):
fix pir coda
mikehh Austin:I was tempted to automate, but the problems are not all the same, I need to check virtually each one 12:22
Austin Geez.
Are there any common themes, that might be good for a posting to the list? (Like "Here's what we're collectively doing wrong")? 12:23
12:23 bluescreen joined
mikehh still after this is done, things should be much easier for everyone, although the perlcritic stuff is going to be a problem 12:24
I will post to the wiki after this is done, what is required etc. 12:26
anyway - back to the grindstone :-} 12:27
dalek tracwiki: v2 | Austin_Hastings++ | ExceptionRefactor 12:34
tracwiki: trac.parrot.org/parrot/wiki/Excepti...ction=diff
rrot: r44387 | mikehh++ | trunk/t/op (5 files):
fix pir coda
 12:38
mikehh ok now t/codingstd/pir_code_coda.t PASSes - let's see if anything is broken 12:41
12:47 ruoso joined
dalek rrot: r44388 | mikehh++ | trunk/t/op (5 files):
fix pir coda
 12:54
13:03 Austin_Hastings joined, clinton joined 13:07 Austin joined
dalek rrot: r44389 | mikehh++ | trunk/t (2 files):
messed up the copyright in two files
 13:11
mikehh codetest - a whole bunch of perlcritic failures, rest of codetest PASSes 13:23
t/examples/tutorial.t - Failed test: 10 - looks like a trailing whitespace problem
all other tests PASS (pre/post-config, make corevm/make coretest, smoke (#32352), fulltest) at r44389 - Ubuntu 9.10 amd64 (gcc with --optimize)
13:34 nbrown_ joined 13:35 ruoso joined 13:39 whiteknight joined 13:43 riffraff joined
Coke msg mikehh if you're working on perlcritic failures, it /might/ be worth running the test only for certain criteria at a time. 13:49
purl Message for mikehh stored.
mikehh Coke: I haven't really started on that yet - everything else fixed now 13:50
dalek rrot: r44390 | mikehh++ | trunk/t (2 files):
add required trailing whitespace, and add to test skip data
 14:00
Coke I'll do a verbose run of P::C and nopaste it. 14:01
mikehh actually when that's done I want to work on Perl::Critic::Bangs :-}
I've actually got it logged 14:02
Coke the output? hokay. 14:03
msg kid51 thanks for updating that test; there's going to be more churn in warnings.pm, though, sorry. :( 14:10
purl Message for kid51 stored.
Coke rant: the warnings probe errors out if a legitimate warning is found based on the currently select set of warnings. causing the remainder of the probes to fail. 14:40
rant: ranting about things that aren't true. 14:43
14:47 iblechbot joined 14:55 kjeldahl joined
Coke I am going to replace @foo?bar@ with either @foo::bar@ or @foo;bar@ . preferences? 15:10
(this is just in a branch for now.) 15:11
also, I need a c programmer. =-)
15:11 theory joined 15:27 bubaflub joined 15:39 whiteknight joined 15:48 Psyche^ joined
Tene Can someone make an exceptions_refactor' branch in svn for me? 16:03
pmichaud Tene: svn copy svn.parrot.org/parrot/trunk svn.parrot.org/parrot/branches/exc...s_refactor 16:04
Tene pmichaud: Thanks. 16:05
Coke C question: 16:08
trac.parrot.org/parrot/browser/trun.../test_c.in
how can I make this pass with gcc's -Wold-style-definition -Wstrict-prototypes 16:09
?
Austin_away What's the gripe? 16:10
Coke that test file fails to compile cleanly with those gcc options. 16:11
dalek nxed: r425 | julian.notfound++ | trunk/examples/pirado.winxed:
absolute minimal but working pasm compiler example
Coke (which prevents us from using those options for testing other files.)
Austin I'm thinking it wants int main(void), since you don't use 'em, or maybe int main(int argc, const char **argv)
Coke I'll try void...
Tene Austin: I made a branch, and I'll start committing later today, maaaaybe. I'm teaching an online class this week, so I've got a bit of time free during the day, but I've got other obligations tonight. 16:12
Coke Austin: booyah. thanks.
dalek rrot: r44391 | tene++ | branches/exceptions_refactor:
Create a branch for exceptions refactoring work.
Austin bows.
NotFound Coke: int main(int argc, char**argv)
Coke: use void only if that prove isn't used for c++ compilers. 16:13
Really dosen't harm, but someone may want to use c++ warnings for unneeded void. 16:14
Coke NotFound: if I switch to that, then it fails a bunch of other warnings. =-) 16:15
(with gcc). testing g++...
NotFound Thinking better... is highly unlikely that any parrot file compiles clean with such option,
Coke AIGH! 16:16
just reverted the wrong file. :|
(*&@#$*(&#@$(*@&$(*@#&$
Tene Coke: ouch. :( 16:17
NotFound Calm, captain Haddock.
Austin Wow. 44 degrees fabulously Farenheit...
NotFound whiteknight: ping 16:18
whiteknight NotFound: pong
NotFound whiteknight: Have you seen my last Winxed commit? 16:19
whiteknight NotFound: no, I don't watch the feed really
link?\\
NotFound code.google.com/p/winxed/source/detail?r=425 16:20
Coke crap. I can't get anything back with vi or svn here. 16:21
count this as my vote to switch to git. :P 16:22
(&*#$.
NotFound It can compile only ops withour args or with a int or string const arg, but works and generated valid pbc 16:23
dalek rrot: r44392 | coke++ | branches/rm_cflags/config/auto/warnings/test_c.in:
Fixup program used for testing warnings.

marking them as invalid (and therefore unused on any other files.) Also, the parrot include was superfluous.
  Austin++		 16:28
16:34 patspam joined
whiteknight NotFound: very nice! 17:01
NotFound whiteknight: thanks 17:03
whiteknight NotFound: if we had an Operation class that contains an Opcode PMC, an optional label name, and a list of arguments, we could construct an array of Operations and then compile it in two passes (first pass to convert labels into offsets, and second pass to convert to bytecode 17:04
In the second pass, we do Operation.compile(), and store the result in the PackfileRawSegment object 17:05
17:05 bluescreen joined
NotFound whiteknight: don't forget the constant generation 17:06
whiteknight NotFound: that's true. When we add the parameters to the Operation, if the arg is a constant, we can store it in a Constant segment immediately
or, maintain a list of constants, weed out duplicates, and then move everything to a PackfileConstantSegment 17:07
in any case, with the right classes, this could be very easy
Maybe we could write a good compilation library in PIR that things like PCT could use
NotFound Yeah, we are on the good way now.
whiteknight: we can write in in winxed following that model, and later move the key parts to pir. 17:08
17:09 cotto_work joined
NotFound (maybe that "we" is "me") X-) 17:09
whiteknight NotFound++
cotto_work hello
purl what's up, cotto_work.
17:17 davidfetter joined
bubaflub hola #parrot; can i get someone to hold my hand through setting the svn:properties correctly on a file? 17:19
i'm converting some tests to PIR and need to make a new file to hold em (and move the old perl5 tests to _old.t) 17:20
17:21 payload joined, payload1 joined
Austin what do you need, bub? 17:23
bubaflub Austin: just want to make sure i set the svn:properties correctly
since i usually use git-svn
i've got a local svn copy
Austin okay 17:24
Are there some other files that aren't new in that dir?
(Oh, and windows or unix?)
bubaflub mac os x, so unix-ish-ish
Tene bubaflub: what I always do is just commit, and then in an svn checkout, there's a test that I run that tells me what to do. 17:25
something about metadata, I think
Austin So the unix-ish way to look is 7 svn proplist foo.t 17:26
Tene t/distro/file_metadata.t I think it was
Austin And the way to set is 7 svn propset foo.t
Yikes.
Dollar signs didn't go through.
The way to set is 7 svn propset <<name>> <<value>> foo.t 17:27
PerlJam almost always uses "propedit" rather than "propset"
Austin So you might do something like 7 svn propset svn:keywords "Author Date Id Revision" foo.t 17:28
(And yes, the svn:keywords really has a colon in it)
bubaflub Tene: your way does seem the path of least resistance 17:29
Tene bubaflub: exactly why I use it. :) 17:30
17:31 ruoso joined
dalek kudo/master: 069df7a | pmichaud++ | src/core/Seq.pm:
Initial version of Seq.sort . Seems to have issues with
 17:32
kudo/master: 9b33a8d | pmichaud++ | src/core/Seq.pm:
Refactor Seq.sort to resolve some arity < 2 bugs (and separate
17:37 riffraff joined 17:41 davidfetter joined
dalek TT #1445 closed by pmichaud++: Add ResizableStringArray.get_number VTABLE function 17:41
bubaflub ok Tene, i've got my files ready to commit on my svn copy.
should i just commit and then run the codingstd tests? 17:42
dalek rrot: r44393 | pmichaud++ | trunk (2 files):
[core]: Add get_number vtable to FixedStringArray PMC. Resolves TT #1445.
 17:50
18:01 iblechbot joined 18:14 bluescreen joined
Tene bubaflub: that's what I'd do. 18:14
bubaflub Tene: cool. i'll run t/codingstd/* first and then commit, run that test you mentioned, then fix whatever crops up
dukeleto 'ello 18:16
18:17 gaz joined
bubaflub mornin' dukeleto 18:17
dukeleto bubaflub: morning 18:18
bubaflub: you may want this gist.github.com/312514
bubaflub dukeleto++
dukeleto Austin: that gist is a useful bash function to set svn props for parrot
bubaflub Tene: i'm running that test on trunk and nothing has exploded
(which normally would be a good thing) 18:19
dukeleto: how's the RTEMS stuff going? 18:20
Tene bubaflub: for the codingstd stuff and svn metadata stuff that I don't really get, I just trust the tests. I know from experience that if I screw up, someone will notice and bug me about it or fix it for me, and that if I worry about it myself, I just won't get anything done, so I prefer to default to asking forgiveness.
dalek rrot-data-structures: dd34546 | (Martin Kealey)++ | (4 files):
Commit edits
rrot-data-structures: c1a225f | (Martin Kealey)++ | t/pmc/fixedpmcstack.t:
Tidy up push_eh/pop_eh flow
rrot-data-structures: fa8001d | (Martin Kealey)++ | (9 files):
Merge branch 'master' of git@github.com:Whiteknight/parrot-data-structures
rrot-data-structures: 39a9558 | (Martin Kealey)++ | src/pmc/fixedpmcqueue (3 files):
typo fixes, now compiles
bubaflub Tene: in the words of Luther, "Sin boldly."
whiteknight kurahaupo++ 18:23
dalek rrot: r44394 | bubaflub++ | trunk/t/compilers/imcc/reg (2 files):
convert some tests in t/compilers/imcc/reg/spill.t to PIR
 18:24
purl I don't know how to convert some tests in t/compilers/imcc/reg/spill.t to PIR.
Coke does imcc spill anymore?
kurahaupo Whiteknight: thanks. Any way to tell dalek to credit my nick? 18:25
whiteknight kurahaupo: Yes, I think we can add you in the CREDITS file in the repo 18:26
I'm not sure if there is another way, you would have to ask Infinoid
bubaflub Coke: you referring to the test i just converted?
kurahaupo Ok, thanks. @DAYJOBS calling, gotta go
Coke bubaflub: yes. 18:28
I think those tests are from a time when there were only 32 registers. 18:29
bubaflub Coke: no wonder they all pass.
Coke I mean, we still have to test "can we pass N args"... but that's not really the same thing as spilling.
bubaflub Coke: ah i see. yeah, i somewhat blindly converted some Perl 5 pir_output_is to a separate PIR test 18:30
Coke might be worth pinging the list; maybe we can just kill some of those tests.
also, when you add a file, run tools/dev/mk_manifest_and_skip.pl 18:31
(or hand edit the MANIFEST* files)
bubaflub whoops. i'll convert the updated MANIFEST now
Coke finally, when you copy a file, use 'svn cp'. don't just 'cp && svn add' 18:33
that preserves history on the file. now when anyone looks at spill_old, they can't go back beyond r44394.
bubaflub Coke: ok. 18:34
Coke: i'll be converting some more tests; do you have a recommendation on where i should start hacking? 18:35
(i.e. which tests would be most useful)
Coke bubaflub: not especially. 18:38
I'm more of a rant-after-the-fact guy. =-) 18:39
particle bubaflub: run the tests with HARNESS_TIMER=1 and look for the slowest ones
bubaflub Coke: haha. ok. i'll shoot the dev-list an email to see if there are any immediate needs. if not, i'll just grab so low-hanging fruit.
particle: ok, i'll try it on fulltest 18:40
particle capture the output so you can grep/sort/whatever
Coke bubaflub: not necessary. fulltest runs N cores, if you do the conversion, you'll probably speed up all cores.
dalek rrot: r44395 | bubaflub++ | trunk/MANIFEST:
updated MANIFEST file for previous commit
Coke in fact, I'd do 'coretest' - those tests are run more often than any other tests.
bubaflub++ 18:41
18:41 bacek joined
bubaflub Coke: ok. So a `make coretest HARNESS_TIMER=1 > out.txt` should do it? 18:41
dukeleto ooh, never heard of HARNESS_TIMER 18:43
Coke enabled that years ago and forgot about it. =-) 18:44
dukeleto bubaflub: we should have a wiki page somewhere that lists tests that need to be converted, with some notes about each test file. want to start a wiki page?
bubaflub: i can help you fill it out
bubaflub: also, look for files that have a foo-old.t, those are tests that still need to be converted
Coke dukeleto: seems like a command to find all the non PIR .t files would be more useful than a list.
bubaflub dukeleto: i know there is a wiki page that lists files that have ### TODO ### in them
particle i set HARNESS_TIMER=1 HARNESS_OPTIONS=j9:f HARNESS_VERBOSE=1 in my .bashrc
bubaflub particle: ok, i'll duplicate that setup. 18:45
dukeleto Coke: i hear ya, but some non PIR .t files are not supposed to be translated
Coke dukeleto: I imagine that list is much shorter than the list of all files. =-)
dukeleto Coke: so the wiki could say stuff like "this test must stay in PASM" etc
Coke: i hear ya
Coke speak up, sonny.
Coke finally does 'set backup' in his vimrc on feather. :( 18:46
dukeleto Coke: i also had the idea of seeing what % of our test suite is in Perl/PIR for each release, so we have some way of knowing where we stand
Coke sheds a single tear for the lost code that he will have to duplicate this evening.
japhb "Pour some out for the binary homies"? 18:49
bacek yawns 18:56
Good morning
bubaflub dukeleto: that sounds good. i believe one of the codingstd tests already figures out which tests are perl and which tests are PIR. we can use that to generate our table and then annotate which ones have to remain perl, which ones are being worked on, etc. 19:03
particle search for ^#!.*perl$ 19:06
19:09 whiteknight joined
Coke bubaflub: the tests ask Parrot::Distribution for "all perl files" or "all pir files". 19:09
cotto_work good morning bacek
clock?
purl cotto_work: LAX: Tue 11:09am PST / CHI: Tue 1:09pm CST / NYC: Tue 2:09pm EST / LON: Tue 7:09pm GMT / BER: Tue 8:09pm CET / IND: Wed 12:39am IST / TOK: Wed 4:09am JST / SYD: Wed 6:09am EST /
Coke for .t files, it scans the shebang, as particle says.
bubaflub Coke: ok. i'll work on a lil somethin-somethin that'll give us some stats 19:10
19:11 fperrad_ joined
bacek morning cotto 19:11
yeah... It's 6am... 19:12
whiteknight is in hell 19:14
I've spent over two hours looking for a USB->RS232 converter that works AND that I have a driver for
davidfetter ugh 19:18
darbelo Silly whiteknight, you should know by now that no USB->RS232 converter works.
whiteknight I know! It's a nightmare
And yet I continually have to work with products that only support an RS232 interface, and no laptop in the last 5 years has apparetly been made with an RS232 port 19:19
like it would kill anybody to add in a USB chip 19:21
darbelo I keep an old 486 at home just for working with a handful of old devices that I can't connect to a modern PC.
dalek m-dynpmcs: b588cbe | darbelo++ | setup.pir:
Add library detection to setup.pir, patch by fperrad++.
 19:22
19:25 lichtkind joined
kthakore NotFound: hi 19:26
NotFound: how can I add you fix for env var for libs for NCI to SDL_Image?
NotFound: github.com/kthakore/parrotSDL
NotFound: almost have every thing compiling just need Image working 19:27
19:28 riffraff joined 19:29 pjcj_ joined, payload1 joined 19:30 payload joined 19:34 hercynium joined 19:42 bkuhn joined 19:43 joeri joined
dukeleto joeri: hola 19:43
purl hi, dukeleto.
cotto_work purl, go jump in a lake 19:44
purl cotto_work: i'm not following you...
19:47 riffraff joined
lichtkind dukeleto: seen your talk 19:47
dalek m-dynpmcs: 0a97ed0 | darbelo++ | setup.pir:
Cleanup after lib detection.
 19:50
19:50 ruoso joined, chromatic joined
dukeleto lichtkind: was it decent? 19:52
19:52 patspam joined
chromatic #ps in 38 19:52
dukeleto will be there
lichtkind lichtkind: yeah relaxed and it gave some realistic impressions from inside the parrot project 19:53
chromatic: please check www.perlfoundation.org/perl6/index.cgi?chromatic and feel free to add 19:55
NotFound kthakore: hard to say what's the better way. Specifying individually each library will be tedious, but if you set only de directory you must seach for possible library names.
kthakore NotFound: any clues?
purl Dee Dee Dee!
lichtkind dukeleto: you want also your page at perl 6 wiki?
kthakore NotFound: on how I can do that? 19:56
NotFound: any regex in parrot?
NotFound kthakore: a good way can be to try to locate them during the install, using pkgconfig or whatever, and put them in a config file, or directly in the code.
kthakore ok 19:58
chromatic Reading now, lichtkind. 19:59
kthakore what is the paste location for here/
?
19:59 Andy joined
kthakore paste? 19:59
purl hmmm... paste is (see: nopaste) or like glue but a little safer to sniff. or nopaste.snit.ch:8001/ or scsys.co.uk:8001/ anywhere shadowpaste is or mmm, sticky paste or You there! Eating the paste. or <see> 2 girls, 1 paste or App::Nopaste
dukeleto lichtkind: that would be great! 20:03
dalek rrot: r44396 | bacek++ | branches/boehm_gc_2/src/hash.c:
Use internal free
dukeleto kthakore: do you need info on regexes in parrot? 20:04
lichtkind dukeleto: then tell me what you did for perl 6 :)
rrot: r44397 | bacek++ | branches/boehm_gc_2 (5 files):
Expose strdup via GC Subsystem
rrot: r44398 | bacek++ | branches/boehm_gc_2/compilers/imcc (11 files):
Switch IMCC to gc_strdup
rrot: r44399 | bacek++ | branches/boehm_gc_2/src/packfile.c:
Switch to Parrot_gc_strdup
rrot: r44400 | bacek++ | branches/boehm_gc_2/src (2 files):
Initialize and disable Boehm GC in main to prevent affecting other GC implementations
rrot: r44401 | darbelo++ | trunk/config/init/hints/openbsd.pm:
Make OpenBSD's liker look in /usr/local/lib even without ICU.
rrot: r44402 | bacek++ | branches/boehm_gc_2 (76 files):
Merge branch 'master' into boehm2
dukeleto lichtkind: i wrote the original implementation of roots() for Rakudo, and I have written many tests for spec test suite
nopaste "kthakore" at 199.198.223.108 pasted "SDL test fails Null PMC access in isa() ???" (19 lines) at nopaste.snit.ch/19746 20:07
kthakore NotFound: any clue ^^ ? 20:08
20:08 patspam joined
kthakore NotFound: why 20:09
does the NULL pmc happen on isa when there is no use off that in StopWatch.pir ?
NotFound kthakore: beacause some thing not expected no be null is, usually. 20:11
kthakore oh ...
ok
what is a pmc ?
purl well, a pmc is responsible for deciding whether to extend itself or not. or a parrot thing, kind of like "magical holds-one-of-anything variable"? or parrot magic cookie or Parrot Magic Cookie or pARROT mAGIC cOOKIE or Poly Morphic Cracker (for the Parrot) or a big problem for optimizations or a compiled pm
NotFound purl: good bot 20:12
purl :)
kthakore huh ?
oh
ok
holds anything var
got taht
purl: good bot
purl :)
kthakore purl: go play in traffic
purl wanders off to dent some cars.
dukeleto kthakore: you can think of a PMC as a general Parrot object 20:13
kthakore dukeleto: ok
20:13 patspam joined
kthakore are there specific objects? 20:14
so far I ahve seen $P* , $I* and $N*
pointer? integer? and numbers?
dukeleto kthakore: those are registers
kthakore ok
dukeleto kthakore: P is for PMC, I for integer, N for numeric and S for string
kthakore ah ok
cool
I learn more cleaning up stuff 20:15
;)
20:15 kurahaupo joined
kthakore I really need time to go trough the PIR tutorials 20:16
but I have only another hour for parrotSDL
then back to work and perlSDL 20:17
dukeleto kthakore: well, this is the best place for all of your questions. keep 'em coming! 20:20
kthakore dukeleto: when it say pc 18 (SDL/StopWatch.pir:49) 20:21
where is the problem happening
49 there is nothign there
in the file
or on line 18 20:22
dukeleto kthakore: it should be line 49, but occasionally our parser gets line numbers wrong. it should be within a few lines of that 20:23
kthakore ok
chromatic We need to fix line numbers... japhb?
20:23 bubaflub left
kthakore chromatic: please ... so lost 20:24
NotFound kthakore: now you start know why I've do only some fixes ;)
japhb chromatic, you rang?
kthakore NotFound: yeah ...
purl doo wop doo wop // shang a langa shanga langaa
kthakore NotFound: jeez
dukeleto: ok so here is my tought process
20:24 patspam joined
lichtkind chromatic: thanks 20:24
chromatic japhb, you mentioned working up examples of line numbering and annotations a while back. 20:25
kthakore dukeleto: it says NULL PMC in isa() and parrot;SDL::StopWatch;__onload
so something in init?
dukeleto: right?
and around line 49
Austin kthakore: So where's the code?
kthakore there is createclass:
dukeleto kthakore: nopaste it
purl nopaste it is, like, See /who is nopaste for the URL
dukeleto nopaste?
purl hmmm... nopaste is at nopaste.snit.ch/ (ask TonyC for new channels) or paste.scsys.co.uk/ or App::Nopaste or tools/dev/nopaste.pl or at www.extpaste.com/ or paste.scsys.co.uk (for #catalyst, #dbix-class, #moose and others) or gist.github.com/ or paste or gtfo or tools/dev/nopaste.pl or trac.parrot.org/parrot/browser/tru...nopaste.pl
kthakore Austin: github.com/kthakore/parrotSDL 20:26
NotFound kthakore: the SDL classes use some deprecated ways of initializing, you probably need to locate and call explicitly his initialization functions,
kthakore ok
it says get_class 'SDL::LCD' tehre 20:27
japhb Oh, right, long ago chromatic wanted me to write tests for line number issues. Then I got (figuratively) run over by work, life, and dying computers. OK. chromatic, you may need to ping me later this week. If I have any tuits, right now they will need to go to help people working on NCI (kthakore and plobsing) and Plumage/distutils (fperrad).
kthakore yay!
Austin Where is this onload function?
kthakore Austin: that is where I am lost 20:28
purl lost is the past tense of to lose or what we call those who are stuck on WINN'T. or "I fell on it in the shower"
kthakore :|
oh line 38
chromatic Will do, thanks.
#ps in 2
kthakore Austin: SDL/StopWatch.pir
Austin Not LCD.pir?
kthakore Austin: code is on githuib
Austin Yeah, I'm looking.
kthakore Austin: um no ... LCD.pir loads properly 20:29
Austin But I don't see isa
kthakore it is SDL::StopWatch
I nopasted the output
will do it again
nopaste.snit.ch/19746 20:30
Austin: there ^^
Austin Got it.
kthakore ok
Austin pc#18 has to be the first line or 2
kthakore from where? line 49?
or on_init?
onload I mean
lichtkind chromatic: i just linked the rakudo star book which hs already an article
Austin The sub 20:31
kthakore ok
there is
class = get class "SDL::StopWatch"
and if_null class, create_class
what I don't get is
why does if_null need a ','
and not if?
but ignoring that it seems to go into create_class: 20:32
Austin Because if_null is an opcode, while if is a macro
NotFound kthakore: if_null is a real opcode, plain if is sintactic sugar. 20:33
kthakore ok
Austin kthak: What happens if you just run that file in parrot directly? (parrot /path/to/file.pir)
kthakore Austin: it uses the runtime library LCD.pir 20:36
Austin Is that bad?
kthakore which looks for the LCD.png in the wrong location
yes
I fix that already
Austin O...kay
kthakore can I use -ISDL/ to use the local SDL?
Austin Yeah
I just want to get the TAP filter out of the way
kthakore ok 20:37
Austin So we can see the actual output
kthakore yeah -ISDL/ still using the runtime
:|
20:37 Andy joined
Austin MAybe you need -LSDL/ as well ? 20:37
kthakore ok
Austin (I for include, L for load byecodes?
*bytecodes
kthakore ok 20:38
...
wth
no go?
is there something like unshift? to put -I on the top? 20:39
Austin What do you mean?
kthakore well it prefers runtime over my SDL local
Austin You're running parrot -ISDL -LSDL .. ? 20:40
kthakore runtime is the first place it looks
yeah
Austin :(
Is it runtime, or is the code maybe saying library/...
in the load directive?
Can you make a symlink " ln -s . library" inside the SDL dir? 20:41
chromatic Ugh, PMCProxy should die.
darbelo NotFound++ # pirado.winxed
Austin chromatic++
kthakore ok
I can try that
chromatic Here's why you can't subclass intrinsic types from PIR.
kthakore where is runtime install ?
chromatic: is that for me?
chromatic The Class PMC looks for vtable overrides defined in PIR.
kthakore, no.
kthakore chromatic: ok
chromatic If one of the parent classes happens to be a PMCProxy with its own vtables defined in C... well, they're not visible. 20:42
Sorry. Out of luck.
Austin Heh
chromatic That's not too bad. I can work around that
... except for the MMD dispatcher. 20:43
Austin Which is part of the whole reason _why_ we want to subclass...
chromatic ... which sees--not the type for which the proxy proxies, oh no!--but the PMCProxy type itself and says "Gosh golly gee, you ain't a Complex PMC, are ya?"
davidfetter anybody know offhand how mod_parrot is handling HLLs? 20:44
NotFound In case some has curiosity, 'pirado' is Spain's argot for fool.
kthakore Austin: crap ... 20:45
Austin ?
kthakore Austin: removed SDL from runtime dirs
forced local SDL
Austin Extreme, but guaranteed to work
And?
kthakore still dies for some reason
Austin Sure.
kthakore hehe 20:46
Austin But that's what we were looking for
Where's it die?
kthakore Austin: LCD loading the image
Austin Mkay
kthakore I put it in same directory
in SDL/LCD.png
Austin Yeah, but is that the cwd?
kthakore in root 20:47
where LCD.pir is in SDL/LCD.pir
Austin Okay, but the code I see (distro) looks for runtime/parrot/library/SDL/LCD.png
kthakore yeah but not parrotSDL
it looks for SDL/LCD.png 20:48
Austin Okay.
kthakore I have LCD.png there
Austin So are you running from SDL/.. ?
kthakore and test can load it
Austin Okay
SDL::Image?
kthakore but parrot run can't
Austin: SDL::Image loads fine
Austin Got to be a path problem, then. 20:49
kthakore ... 20:50
Austin Are you getting a file-not-found error, or something else?
kthakore nope
just null PMC()
Austin nopaste?
purl rumour has it nopaste is at nopaste.snit.ch/ (ask TonyC for new channels) or paste.scsys.co.uk/ or App::Nopaste or tools/dev/nopaste.pl or at www.extpaste.com/ or paste.scsys.co.uk (for #catalyst, #dbix-class, #moose and others) or gist.github.com/ or paste or gtfo or tools/dev/nopaste.pl or trac.parrot.org/parrot/browser/tru...nopaste.pl
kthakore Austin: you have parrot and git?
you can run it too
Austin Installed, and yes.
kthakore ;)
I will nopaste this though 20:51
Austin libSDL not found 20:52
nopaste "kthakore" at 199.198.223.108 pasted "parrotSDL fails more" (31 lines) at nopaste.snit.ch/19747 20:53
Austin Apparently I don't have SDL installed? 20:54
dalek tracwiki: v9 | dukeleto++ | AllisonTasklist
tracwiki: trac.parrot.org/parrot/wiki/Allison...ction=diff
tracwiki: v10 | dukeleto++ | AllisonTasklist
tracwiki: trac.parrot.org/parrot/wiki/Allison...ction=diff
kthakore Austin: sudo apt-get install libsdl1.2 libsdl-image1.2 20:55
Austin That package is not available, but these 18 packages have replaced it... 20:56
kthakore huh?
NotFound bacek: I don't think that using include files from HLL is good way. That disallows the posibility of directly generating code whithout callig a pir or pasm compiler, or forces to add a parser for standarized include files.
Austin I got libsdl1.2debian-all installed, and passed 7 tests.
bacek NotFound, true. 20:57
kthakore Austin: what no errors for you?
bacek But we still need some standard way to pass PASM constants to HLLs
Austin Sure, on the 8th test
nopaste "Austin" at 68.37.46.53 pasted "Partial success" (15 lines) at nopaste.snit.ch/19748
Austin Don't hate me because I'm lucky...
NotFound bacek: yes, in general we lack a way to store named constants, or a standarized way to get them at runtime. 20:58
Getting the value given the name, I mean.
Austin Maybe you need to install more packages....
(apt-get is like violence...) 20:59
bacek NotFound, let's create parrot_constants hash, generated in compiletime.
NotFound, similar to config_hash
NotFound bacek: yes, but also for modules providing his own.
21:00 riffraff joined
kthakore Austin: install libsdl-image1.2 21:00
Austin: I will not hate you if this works
Austin: I have these packages already 21:01
Austin ttF?
purl rumour has it ttF is comming soon - gimme a week - another week you'll have TeX fonts. or This The Fuck
kthakore Austin: sure but
libsdl-ttf1.2
NotFound bacek: maybe better a keyed access in the interpreter. Exposing a hash allow accidental modifications.
Austin Couldn't find package 21:02
bacek NotFound, it's harder to implement. Basically we can't generate this method. 21:03
NotFound Mmm.. true.
kthakore libsdl-ttf2.0-0 21:04
Austin: ^^
NotFound A loadable pbc may be fine. Is easy to give that way a try.
Austin Well, that was harder than it needed to be.
kthakore Austin: yeah ... 21:05
purl doo wop doo wop // shang a langa shanga langaa
bacek NotFound, yeah. We can add constants_hash to IGLOBALS. And populate it automatically in some loadable PBC
Austin Package is installed, but I'm still getting the SDL_ttf not initialized failure on test 9
NotFound Austin: there are lots of libraries runtime-linking unfriendly these days.
kthakore yeah I dunno know ...
:|
Austin: I don't have enough skills to debug this hunk of junk 21:06
no offense people
Austin :)
kthakore I take all the blame
Austin IS there a perl module that can test my sdl installation?
darbelo kthakore: You adopted it, you get to call it names.
kthakore yesh
NotFound kthakore: don't worry, we were fixing at tiny steps because of that.
kthakore darbelo: hey! but then I abuse my adopted kids too 21:07
Austin: github.com/kthakore/SDL_perl
Austin: github.com/kthakore/Alien_SDL
Austin: the second one will compile everything for you if you want
Austin Okay, what's the plan? 21:08
purl i heard the plan was to use fastcgi
kthakore Austin: test the deps
perl Build.PL; perl Build; perl Build test
Austin I've got sdl, right, so I download one? or both? of those, and do what?
kthakore Austin: teh SDL_perl is for test suite
darbelo dukeleto: ping.
kthakore Austin: Alien_SDL is a dep of SDL-perl
Austin Are they on cpan? 21:09
kthakore Austin: yes but we just did a lot of fixes on redesign
dalek rrot: r44403 | jkeenan++ | trunk/t/compilers/imcc/reg/spill_old.t:
Set SVN eol-style property on file.
kthakore Austin: oh btw github.com/kthakore/SDL_perl/tree/redesign 21:10
Austin: the cpan one is a dev reelase too
Austin fatal: github.com/kthakore/SDL_perl/info/refs not found: did you run git update-server-info on the server?
rrot: r44404 | bacek++ | branches/boehm_gc_2/compilers/imcc (20 files):
More GC allocations in IMCC
NotFound kthakore: maye will be easier to restart from scratch.
Austin git--
kthakore Austin: it is a html link
dalek rrot: r44405 | bacek++ | branches/boehm_gc_2/src/library.c:
Use gc_strdup
rrot: r44406 | bacek++ | branches/boehm_gc_2/lib/Parrot/Ops2c/Utils.pm:
Use mem_gc_free
kthakore NotFound: probably
rrot: r44407 | jkeenan++ | trunk/t/compilers/imcc/reg/spill_old.t:
Add SVN properties.
rrot: r44408 | bacek++ | branches/boehm_gc_2 (2 files):
Pass interp into free_values func
kthakore wow
rrot: r44409 | bacek++ | branches/boehm_gc_2/config/gen/crypto/digest_pmc.in:
Use GC allocations
kthakore this is ....
dalek rrot: r44410 | bacek++ | branches/boehm_gc_2/src/parrot_debugger.c:
Use gc_strdup
rrot: r44411 | bacek++ | branches/boehm_gc_2/t/src/extend.t:
Use GC allocations
kthakore Austin: you need to open those links in a bowser 21:11
21:11 hercynium joined
kthakore git clone url is on there 21:11
Austin yeah, got it
NotFound kthakore: my own idea was to do some prototyping in Winxed.
kthakore Austin: get teh redesign branch fro SDL_perl
NotFound: how did that go for ya?
NotFound kthakore: but I suppose that perl people will find nqp friendlier. 21:12
dukeleto darbelo: pong
kthakore NotFound: I am doing this for parrot peop
bacek NotFound, github.com/bacek/pir 21:13
kthakore NotFound: I don't seem to like taking easy ways out :|
Austin: we are over on #sdl I have to head out
NotFound kthakore: for protoyping?
darbelo dukeleto: Are you still working on Parrot-on-bug? I might have some embedded tuits this week. 21:14
kthakore NotFound: I have to learn pir sometime
NotFound: I already have the arcitecture prototyped in SDL-perl
Austin: I have to go ask help for SDL_perl in #sdl
Austin okay, see you later
kthakore Austin: gluck comrade 21:15
may the red communist start glow on your cold soul
NotFound kthakore: was just an idea, do it as you like better.
kthakore NotFound: np I always do that
;)
dukeleto: NotFound Austin thank you very much 21:16
darbelo red communist?
purl somebody said red communist was www.comics.aha.ru/rus/stalin/1.html
kthakore I probably got a lot further then I though I would
dukeleto: open source is comunnism
NotFound bacek: yes?
kthakore dukeleto: didn't someone tell you? 21:17
dukeleto: it was a joke btw
dukeleto darbelo: yes, I am working on BUG. i am learning about BitBake
kthakore caio
dukeleto kthakore: later!
kthakore japhb: also thanks for help too 21:18
dukeleto: k
dukeleto darbelo: i am attempting to setup a working OpenEmbedded dev environment
darbelo: so that I can write a bitbake recipe for parrot
darbelo dukeleto: Excellent, I was starting on that when I ran out of steam. Let me know if you get a build enviroment working. 21:19
dukeleto darbelo: sounds good. i will create a git repo soon for my BUG-related development and give you a commit bit 21:20
darbelo: all i have now is mostly notes and ideas
21:25 AndyA joined
chromatic Fun question: which should take priority, vtable overrides or MMD? 21:30
Austin In what sense?
darbelo Both! None!
chromatic t/pmc/multidispatch_26.pir
AInt extends Integer, which provides its own add VTABLE entry in C. 21:31
AInt also has a multi with the first candidate first parameter type of AInt. 21:32
Austin AFAICT, that's a sub, not a vtable override. So it shouldn't get called unless you say 'add'($P0, $P1) 21:33
(No :vtable marker on the sub, for instance) 21:34
21:34 AndyA_ joined
chromatic Good point. There's an interesting twist. 21:34
Austin If the sub was avtablle override, then it would be competing with whatever Integer offers, and I guess it would win because $P0 is an AInt. 21:35
I can't recall how mmd handles (+1, -1) versus (-1, 0) 21:36
No, that'd be (-1, -1) vs (-1, 0), which I think means the integer vtable would win 21:37
21:38 AndyA joined
chromatic It doesn't even get to MMD. 21:38
We have two conflicting goals here. 21:39
Austin Why not?
chromatic Hm, that's a good question. I have a different idea.
dalek rrot: r44412 | bacek++ | branches/boehm_gc_2/src/gc/gc_boehm.c:
Don't try to copy data from not-yet-allocated strings.
 21:43
rrot: r44413 | bacek++ | branches/boehm_gc_2 (2 files):
Export Parrot_gc_strdup
21:49 magnachef joined
cotto_work timetobleed.com/garbage-collection-...onference/ 21:53
chromatic, ping 21:54
chromatic pong 21:56
cotto_work Do you have the time to do a quick writeup on how you knew how to fix the memory leak that was reported by the rakudo devs? 21:57
I took a look at it and could only come to the conclusion that, yes, there's a leak. 21:58
chromatic I knew what the patch did. 21:59
I added some debugging statements to check what I thought were the invariants of the patch.
* Did the memory allocated counter get updated properly?
* Did the code run into an overflow? 22:00
cotto_work which patch?
chromatic Oh, I bisected to see where the out of memory occurred.
cotto_work ah
thanks
dukeleto bisect++ 22:01
22:01 shockwave joined
chromatic Mostly it was experience and stubbornness. 22:01
shockwave The PMCs located at <root>/src/pmc are commented, but they are not at docs.parrot.org 22:03
I wonder if there is a script that already processes them. 22:04
darbelo I *think* that's 'make html's job
shockwave darbelo, could be. 22:05
So, it's not possible to do: $N0 = 2.14 + 1, eh.
I guess that compiler writers are supposed to promote 1 to a float? 22:06
darbelo You could do it in two steps. $N0 = 2.15 ; $N0 += 1 22:07
eh, s/4/5/
eh, s/5/4/
shockwave Sorry. The above works. What doesn't work is: $N0 = 1 + 2.14
darbelo, yeah. It could be done in multiple steps. 22:08
Does anyone know if the fact that "$NO = 2.14 + 1" works but "$N0 = 1 + 2.14" doesn't a feature or a bug? 22:14
- /doesn't a/doesn't is/
chromatic Sounds like a bug, but I'd have to see the bytecode to say for sure. 22:16
darbelo looks to me that it happens becouse we don't have a add_n_i_* opcode. 22:17
shockwave chromatic, I can put a ticket if you like. Finding out the answer to this may help me elimate a 'special case'.
chromatic Sounds like something the PIR compiler can rewrite. 22:18
shockwave chromatic, Do you mean IMCC?
chromatic Yes. 22:19
fperrad ping darbelo 22:20
shockwave Actually, this does work. I was assigning the result to an int register, which is why it was failing. I'll do some more tests. 22:21
darbelo fperrad: pong 22:22
fperrad darbelo PIR version of t/dynpmc/gdbmhash.t fails on Windows 22:23
nopaste "fperrad" at 79.81.92.9 pasted "output of t/dynpmc/gdbmhash.t on Windows" (124 lines) at nopaste.snit.ch/19750
darbelo That's odd. 22:24
dalek TT #784 closed by chromatic++: subclassing breaks multi dispatch
darbelo It can create the file, but not remove it?
nopaste "fperrad" at 79.81.92.9 pasted "patch of t/dynpmc/gdbmhash.t" (22 lines) at nopaste.snit.ch/19751 22:25
fperrad darbelo, POSIX unlink is not immediat, but when all file descriptor are closed 22:27
and win32 unlink is not POSIX
darbelo I thought the OS pmc was supposed to handle that for us. 22:28
fperrad Perl5 unlink is better than OS.unlink 22:29
darbelo Yeah, I'm looking at the OS pmc implementation now... 22:30
It has room for improvement. 22:32
dalek rrot: r44414 | chromatic++ | trunk (2 files):
[lib] Removed "MMD must obviously take care of this!" from i_* VTABLE entries,
rrot: r44415 | chromatic++ | trunk/src/pmc/oplib.pmc:
[PMC] Fixed compiler warnings in OpLib's get_pmc_keyed_int().
mj41 I see occasionally hanging in pbc_disassembler.exe during make test on TapTinder windows clients. 22:36
22:37 Whiteknight joined
darbelo mj41: pbc_disassemble is rather fragile. It segfaults every now and then on my box. 22:37
fperrad darbelo another way is : gdbmhash PMC implements a method close
NotFound mj41: Can you locate a pbc file that fails? 22:38
mj41 NotFound: probably tt.ro.vutbr.cz/report/pr-Parrot/do?...hat+I+mean 22:40
NotFound "Can't spawn cmd.exe Permission denied" doesn't look pbc related. 22:43
mj41 NotFound: I see hanging so I kill pbc_dis...exe in task manager. Trying to compare other test runs on tt.ro.vutbr.cz/report/pr-Parrot/rp-trunk (checkboxe and button). 22:44
kthakore Austin: how did it go?
Whiteknight is upset about missing #ps today 22:48
NotFound Files not closed on exit is a known source of problems, BTW
kthakore hi Whiteknight
dalek rrot: r44416 | darbelo++ | trunk/t/dynpmc/gdbmhash.t:
Make temp file deletion smarter to cover the fact that the OS pmc is dumb. Thanks to fperrad++ for pointing this out and prviding a patch.
Whiteknight hello kthakore
NotFound I think the problems with that are related to order of destruction. 22:49
Whiteknight I missed allison today, apparently. What's going on with the PCC refactors? I'm itching to work on that 22:50
is she leading the charge?
NotFound Whiteknight: she said she hopes to be able to finish it a few weeks.
Whiteknight NotFound: a few weeks? that's quite a long time. I suspect we could knock it out faster 22:51
especially if we're going to call it "high priority" every week
NotFound Well, I think she said two, actually.
shockwave What's PCC?
purl Welcome to Perl Community College, where we, for free, and at your convenience, teach you all of an intro-CS curriculum, such as you should have learned in college or by simply reading on your own. Please also visit the Perl Crisis Clinic, where we do all your job for you. or Parrot Calling Convention or Proof Carrying Code or en.wikipedia.org/wiki/Portable_C_Compiler
NotFound checking the log... 22:52
Whiteknight shockwave: the calling conventions subsustem
shockwave Whiteknight, purl is a pretty good bot.
purl :)
NotFound "and, hopefully I can merge the branch in within 2 weeks"
purl: have a cookie 22:53
purl Sorry, NotFound, I'm on a diet.
Whiteknight NotFound: that's all fine and good, but two weeks is a long time to wait 22:55
these refactors aren't nearly so important as the last PCC refactors, but it's still a lot to put entirely on allison's shoulders 22:56
darbelo Whiteknight: she said she could partition out tasks for people willing to help IIRC.
Whiteknight darbelo: okay, that's more what I wanted to hear!
shockwave Math has so many undesirable properties, IMHO. I hate the fact that addition is cummutative, but negation isn't. Who do I complaint to about that? 22:57
Whiteknight I would like to help with chromatic's namespacey issues too, but I probably can't be as much help there without some serious research into his branch
darbelo irclog.perlgeek.de/parrotsketch/201...#i_2025837 22:58
Whiteknight shockwave: it's a system of convention. Create your own system with it's own rules. Only hard part is proving it is consistent
plobsing shockwave: how can a unary operation (negation) be commutative?
shockwave probsing, lack of context analysis? I meant subtraction. 22:59
Whiteknight where is Pynie, googlecode? 23:00
NotFound plobsing: you can't prove that is not commutative :) 23:01
Whiteknight bah, pynie doesn't even build 23:02
darbelo Whiteknight: really? 23:03
I saw commit go by just today. 23:04
Whiteknight darbelo: Just checked out, looks like some files were moved and the build script not updated 23:05
dalek rrot: r44417 | darbelo++ | trunk/t/dynpmc/gdbmhash.t:
Give each test it's own db file.
rrot: r44418 | mikehh++ | trunk/t/perl/Parrot_IO.t:
fix perlcritic failure - remove hard tabs
nxed: r426 | julian.notfound++ | trunk/examples/pirado.winxed:
enclose debug statements within "if (DEBUG)" instead of commenting out in pirado
 23:14
darbelo Whiteknight: nopaste.snit.ch/19752 23:16
Does that work for you? 23:17
dalek rrot: r44419 | mikehh++ | trunk/t/codingstd/c_cppcomments.t:
fix perlcritic failure - remove hard tabs
 23:22
23:23 kid51 joined
kthakore hi kid51 23:25
Whiteknight purl msg kurahaupo FPQ2 and FPQ3 both fail tests with floating point exceptions 23:30
purl Message for kurahaupo stored.
dalek rrot-data-structures: c36ded1 | Whiteknight++ | benchmarks/fixedpmcqueue (2 files):
add benchmarks for FPQ2 and FPQ3. Fix test counts for both. Add both to setup.pir so they build, and benchmarks.sh so the benchmarks run. Both types exhibit Floating point exceptions and fail tests/benchmarks
 23:31
dukeleto oooh, i need to see how PDS does benchmarking. maybe i can leverage euler_bench 23:32
chromatic Whiteknight, there's not much to say about tt_389. 23:36
dalek rrot: r44420 | mikehh++ | trunk/t/distro/meta_yml.t:
fix perlcritic failure - missing perl coda
 23:38
Whiteknight chromatic: last I asked, you had open questions 23:49
chromatic All that branch needs to do now is store methods defined on core types in the single PMCProxy for each core type. 23:50
Whiteknight ok
chromatic (well and then change the default PMC's find_method() VTABLE to look in that proxy)
dalek rrot: r44421 | mikehh++ | trunk/t/dynpmc/rational.t:
fix perlcritic failure - missing perl coda
 23:54
rrot: r44422 | mikehh++ | trunk/t/compilers/imcc/syn/hll.t:
fix perlcritic failure - missing perl coda
darbelo bacek: ping
23:59 patspam joined

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!