Parrot 2.6.0 | parrot.org Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | fix 'make html' (talk to Coke), update tutorial (talk to tcurtis)
Set by moderator on 21 July 2010.
00:02 snarkyboojum joined
NotFound Confirmed that the build problems on Windows is that \\ are not escaped since r48179 00:16
Someone more skilled with perl re than me should look at it. 00:17
chromatic Go ahead and revert it. I think it's the wrong approach altogether. 00:18
NotFound Too late for me today, going to bed.
00:49 zostay joined 01:19 rurban_ joined 01:26 dduncan joined
Coke do we have any winders folks that could help pmichaud with fixing 'ufo' for windows? 01:32
kthakore ufo? 01:35
purl ufo is a major influence on chip's 'future' imagery or a great 70s hard rock band from England that didn't get nearly enough acclaim.
sorear ufo is also the Perl 6 MakeMaker prototype 01:36
purl okay, sorear.
kthakore sorear: ufo? weird name though 01:37
sorear the perl 6 people love wierd names
dalek kudo: 2808a52 | (Solomon Foster)++ | src/core/Int.pm:
Add Int versions of infix:ļæ½==ļæ½, infix:ļæ½!=ļæ½, and infix:ļæ½<ļæ½ which go
 01:53
cotto ~~
chromatic Takers welcome for incoming.... 01:54
PerlJam chromatic: incoming what? 02:02
purl duck!
dalek tracwiki: v2 | chromatic++ | FixingPIRVtableOverrides 02:04
tracwiki: added task list
tracwiki: trac.parrot.org/parrot/wiki/FixingP...ction=diff
02:08 snarkyboojum joined
bubaflub dukeleto: i just pushed some more stuff to the repo, all the tests in t/configure now pass 02:18
dukeleto bubaflub++ # nice work! 02:19
bubaflub: have you tried a "make coretest" ?
bubaflub dukeleto: i'm running it now
dalek tracwiki: v2 | Util++ | splint 02:20
tracwiki: Add formatting to Splint Wiki page
tracwiki: trac.parrot.org/parrot/wiki/splint?...ction=diff
dukeleto bubaflub: sweet. gist the results when you get them 02:22
bubaflub dukeleto: bahahahahhaa 02:27
All tests successful.
purl Time for a well deserved vacation, bubaflub!
bubaflub Files=215, Tests=6554, 76 wallclock secs ( 1.90 usr 1.04 sys + 30.40 cusr 21.00 csys = 54.34 CPU)
Result: PASS
dalek website: gerd++ | Documentation and information collection 02:34
website: www.parrot.org/scratch/documentatio...collection
02:36 janus joined
cotto crosses one thing off his todo list 02:41
literal All tests successful. 02:42
purl cotto: all tests successful =is= <reply>Time to write more tests!|<reply>Time for a well deserved vacation, $who!|<reply>$who: that's because you wrote only one test, slacker!|<reply>That's great, $who! Now the only bugs that remain are the ones you haven't tested for yet!|<reply>pass(); ok(1); is(1,1); done_testing();|<reply>Are you feeling lucky?
Util_ gerd++ # Good job on the doc cluster
dalek rrot: r48193 | cotto++ | trunk/docs/project/support_policy.pod:
update support policy with deprecation notifications
 02:43
02:43 bubaflub joined 02:46 LoganLK joined
ttbot Parrot trunk/ r48193 MSWin32-x86-multi-thread make error tt.taptinder.org/file/cmdout/359928.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 02:50
cotto that'd better not be my fault
02:55 pmichaud joined
bubaflub cotto: i don't think so, those windows errors have been popping up all day 02:56
cotto My change didn't touch any code.
If that change broke the build, ...
bubaflub cotto: NotFound says it's r48179 (irclog.perlgeek.de/parrot/2010-07-28#i_2621088) 02:59
that broke the build
cotto Time for some yak shaving. 03:10
03:12 fred joined
bubaflub dukeleto: so i think i should refactor all of my uses of FindBin into a separate module so we get something a bit more beautiful than the current patch 03:14
chromatic Feel free to revert r48179. 03:17
03:21 tcurtis joined
dukeleto bubaflub: that sounds good 03:22
bubaflub dukeleto: yeah, i'll think about how to best do that. it needs to be a singleton that works similar to FindBin but has a few more features
03:51 plobsing joined 03:54 snarkyboojum joined
dalek rrot: r48194 | tcurtis++ | branches/fixing_pir_vtable_overrides:
Create a branch for implementing chromatic's task list from FixingPIRVtableOverrides on wiki.
 04:05
04:14 somebody_ joined
dalek tracwiki: v14 | chromatic++ | LoritoDesignQuestions 04:34
tracwiki: filled out lots of little details
tracwiki: trac.parrot.org/parrot/wiki/LoritoD...ction=diff
cotto Yay. Some who's not me is editing that page! 04:41
chromatic++ 04:42
chromatic I added a short section at the top to explain the ultimate power.
cotto "supported" should be defined more carefully there. It sounds like the VM should support those things (CPS, mmd, gc, etc) directly. 04:48
05:04 bubaflub joined 05:19 bubaflub joined
dalek tracwiki: v15 | chromatic++ | LoritoDesignQuestions 05:24
tracwiki: cotto said "Wait, what supports what?"
tracwiki: trac.parrot.org/parrot/wiki/LoritoD...ction=diff
cotto yes, I did 05:29
chromatic, what'd be a better fix than this: 05:30
nopaste "cotto" at 192.168.1.3 pasted "avoid prematurely collecting some pf-related strings" (52 lines) at nopaste.snit.ch/22362
cotto Without that, the names of the pf segments get collected and end up being printed as garbage (or better) by the time control flow reaches the debug segment. 05:31
chromatic Will look in a moment. 05:32
cotto thanks 05:33
05:40 bacek_really_at_wor joined
bacek_really_at_wor aloha, humans 05:40
cotto aloha, bacek_really_at_work_but_slightly_truncated
bacek_really_at_wor cotto, :) 05:41
chromatic cotto, are those strings really PMCs? 05:42
also, why do they get collected prematurely?
cotto they're not PMCs but that function only pretends to care
I guess because nothing's marking them. 05:43
chromatic Right, I figured out the first question after I asked.
Fixing the marking seems better to me. 05:44
another_bacek migrating to Packfile PMCs is even better 05:45
cotto I knew you were going to say that. 05:46
I'd prefer to get something fixed tonight though.
chromatic Why don't they get GCd prematurely when run through parrot, not pbc_dump?
Or does nothing in parrot use them?
sorear I suppose that's better than bacek_at_war 05:47
another_bacek chromatic: maybe -t4 use them. But unlikely.
cotto who says they don't?
chromatic bacek_destroy_all_humans
cotto I doubt anything normally looks at them
chromatic coke does occasionally complain that -tn doesn't work.
afk, loud television 05:48
cotto I CAN SHOUT SO YOU CAN STILL HEAR ME IF IT HELPS
much better fix now 05:55
and I broke the build 05:56
after committing 05:57
all unbroked 06:00
ttbot Parrot trunk/ r48195 i386-linux-thread-multi make error tt.taptinder.org/file/cmdout/359979.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ )
cotto stupid bird 06:01
dalek rrot: r48195 | cotto++ | trunk (2 files):
keep packfile segment names from being GC'd until the segment has been destroyed
rrot: r48196 | cotto++ | trunk (2 files):
partially revert previous commit, unbreak the build
06:04 baest joined
cotto How nice that a GCable is created before the gc registry is initialized. 06:05
ttbot Parrot trunk/ r48196 MSWin32-x86-multi-thread make error tt.taptinder.org/file/cmdout/360025.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 06:14
06:16 uniejo joined 06:18 Casan joined
dukeleto (unbreaking the build)++ 06:21
06:21 slavorgn joined
cotto I'm starting to think that bacek's approach is actually the faster and more straightforward one. 06:23
chromatic I don't see why they need registering. 06:26
Does nothing else in packfiles need marking?
cotto The GC gets hungry and I guess they look sufficiently delicious after a while. 06:27
06:28 another_bacek joined
cotto I had a similar problem with the profiling runcore 06:28
06:29 AndyA joined
chromatic How about marking names in find_const_iter()? 06:31
06:32 dduncan left
chromatic Do you have example code which shows the segfault? 06:33
another_bacek thawed constant strings allocated from constant pool
cotto That test I added shows the strings getting collected. 06:35
dukeleto hopes that now that there is another bacek, stuff gets done twice as fast 06:36
dukeleto spoke too soon
06:37 GeJ joined
chromatic cotto, I think I have it. Testing now. 06:37
cotto I have a lta fix that doesn't break anything, but I'd love to see something better. 06:38
chromatic t/tools/pbc_dump.t (Wstat: 0 Tests: 7 Failed: 0)
TODO passed: 7
cotto good sign
that's the one I just added 06:39
chromatic Is there a TT for this?
cotto nope
It's related to 1127 indirectly.
I just decided to dig into it today. 06:41
chromatic Incoming.
purl duck!
cotto thanks! 06:42
chromatic That gave me a chance to remove some ugly from that file anyway -- not much, but some. 06:43
Now truly afk for the night. 06:47
cotto g'night 06:50
dalek rrot: r48197 | chromatic++ | trunk (2 files):
[pf] Added mark for packfile names.
 06:51
rrot: r48198 | chromatic++ | trunk/src/packfile.c:
[pf] Tidied code; no functional changes.
ttbot Parrot trunk/ r48198 MSWin32-x86-multi-thread make error tt.taptinder.org/file/cmdout/360100.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ )
06:52 fperrad joined
dukeleto i would like to check that Devel::Cover exists at the beginning of "make cover", anybody have an idea for the best way to do that? 06:56
i am thinking just do perl -MDevel::Cover and check $? 06:57
some tests fail under "make cover" but not "make test". This is disconcerning. 07:13
ah, i see. the profiler messes up some of the output occasionally 07:14
07:21 somebody__ joined
dalek rrot: r48199 | dukeleto++ | trunk/t/pmc/capture.t:
[cage] Fix a coding standard test
 07:24
ttbot Parrot trunk/ r48199 MSWin32-x86-multi-thread make error tt.taptinder.org/file/cmdout/360196.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 07:32
07:32 somebody_ joined
cotto partcl? 07:58
purl partcl is tcl on parrot or code.google.com/p/partcl or git://github.com/partcl/partcl.git
dukeleto i keep getting "PackFile_unpack: This Parrot cannot read bytecode files with version 8.0" with PL/Parrot and it is driving me insane 08:01
dukeleto deletes a bunch of stuff and tries again 08:03
moritz dukeleto: git clean -xdf # and rm -rf $parrot_install_location has helped me 08:05
beware, it deletes all untracked files
dukeleto moritz: i just realized that my parrot_install_location had a bunch of crud I didn't expect, and blew it away 08:06
moritz++ for the support
moritz: yep, removing the install directory worked! good to know. 08:07
08:08 theory joined
moritz I just wonder why that happens. I thought the .pbc files were all in directories that contained the version 08:08
dukeleto moritz: the odd thing was, i would get that error on the first call to Parrot_new(), i wasn't even loading bytecode yet 08:09
moritz: somehow, Parrot_new does something wrong and triggers that unpack error 08:10
moritz dukeleto: funny thing
purl funny thing is probably that the spam doesn't even link to real porn
dukeleto moritz: it may be the whole "fake a packfile" thing, but I barely understand how that works 08:11
moritz: do you know of a way to programmatically figure out the location of perl6.pbc ? 08:16
moritz: currently i have to specify it as an environment variable to PL/Parrot, but it would be nice if that didn't need to be the case 08:17
moritz dukeleto: it has a fixed relative path to the installed parrot... lemme check
dukeleto: it's in $parrot_prefix/lib/$parrot_version/languages/perl6/perl6.pbc 08:18
if it's installed (we need to educate our users to install it anyway
)
dukeleto moritz: yes, i guess that is programmatic enough 08:20
dukeleto runs the PL/Perl6 test suite for the first time in a few weeks and finds a new error message 08:21
==SORRY==
Unable to find Perl 6 low-level runtime files
moritz dukeleto: did you run 'make install' in rakudo? 08:22
you should :-)
dukeleto hmmml. rakudo tells me to install it first, but I am using the installed perl6.pbc
moritz: yes, i am definitely using the installed perl6.pbc
moritz huh
dukeleto learned that lesson enough times 08:23
moritz: i think i have some LD_LIBRARY_PATH monkey business 08:26
moritz ugh 08:27
dukeleto needs to remember to restart Postgres whenever he changes LD_LIBRARY_PATH but fails at doing that 08:29
moritz: i fixed it. i was compiling against a different Parrot than --gen-parrot created, because my $PATH was incorrect 08:34
dukeleto is now grateful for that error, instead of a coredump or having his hair burst into flames
moritz IMHO rakudo should always use the generated parrot (when configured with --gen-parrot), and not the one from PATH 08:35
dukeleto moritz: yes, but my PL/Parrot runs parrot_config to figure stuff out, and I didn't have the (correct) path to Rakudo's parrot_config first
moritz ah 08:36
so not rakudo's fault
dukeleto moritz: so I was compiling with the wrong parrot_config info
yay! PL/Perl6 still has 6 passing tests. I am excited to hack on PL/Parrot against Rakudo* 08:37
08:37 lucian joined
moritz dukeleto: Rakudo compiler freeze for R* is in about 9+1/2 hours 08:37
wait, two hours more 08:38
moritz is time-zone confused
dukeleto is excited. After I go to sleep in a few minutes and then wake up, R* will almost be out the door ;) 08:41
moritz a well, compiler freeze != relase
*release
but still, release is tomorrow, fsvo tomorrow 08:42
moritz is pretty excited about it
dukeleto moritz: yes, i think many people are very excited, especially Parrot people 08:43
moritz especially Rakudo people :-)
dukeleto moritz: how do I turn a ResizablePMCArray into a Perl 6 Array, in PIR ? 08:49
moritz: is that possible, or a fools errand?
moritz dukeleto: $P0 = '&infix:<,>'($P1, :flat) 08:50
dukeleto: (from within the perl6 HLL - you might to obtain with with get_root or get_hll_global or so first if your namespace is something else) 08:51
dukeleto moritz: ooooh, shiny!
moritz: i have the general issue of calling Perl 6 from PIR, and wanting to pass arguments to Perl 6 functions, but I have Parrot datatypes 08:52
moritz dukeleto: and for hashes tehre is '&CREATE_HASH_FROM_LOW_LEVEL"'(parrot_hash)
that's nasty, yes
we need to improve our HLL Interop somehow
dukeleto moritz: what about the basic data types like int, float and string? 08:53
moritz: do those need any data marshalling?
moritz dukeleto: yes, otherwise some built-in methods fail to work
&prefix:<~> for strings
don't know about int/float
dukeleto moritz: the info you just told me about gives me enought to be dangerous :) thanks! 08:55
moritz :-) 08:56
dukeleto moritz: i am getting error:imcc:syntax error, unexpected ADV_FLAT (':flat') 08:59
08:59 AndyA joined
dukeleto moritz: i will try without it 08:59
moritz dukeleto: no, just omit the comma before it
dukeleto moritz: ah, after doing that, it finally tells me it can't find &infix:<,>, because I didn't get_* it 09:01
moritz dukeleto: from HLL then
dukeleto moritz: something like get_hll_global '&infix:<,>' ? 09:03
moritz: that doesn't seem to work. i am not grabbing it from the correct namespace 09:05
moritz dukeleto: $P0 = get_root_global ['perl6'], '&infix:<,>' or so 09:06
dukeleto moritz: that did something! but now I am getting Parcel objects, does that sound about right ? 09:08
moritz yes 09:09
parcel is a bit like an array, but read-only
rakudo: say Parcel ~~ Positional
p6eval rakudo 2808a5: OUTPUTĀ«1ā¤Ā»
dukeleto moritz: i am closer, but @_ is still not being populated 09:14
moritz: thanks very much for your guidance, i now have new shiny things to hit my head on, instead of old boring things :) 09:15
moritz try
sub f(@a) { say @a.perl };
and then call '&f'($your_Parcel) 09:16
dukeleto moritz: i will try something like that. my issue is that I am building up a string of Perl 6 from PIR, then executing it, but I am sure my wrapper code is wrong 09:19
moritz: it was just the first thing that kinda worked
09:19 rurban_ joined
dukeleto moritz: i am doing : my $r = eval q<<< sub { ... } >>>; $r.() 09:20
moritz: where ... is the perl 6 code that gets inserted
moritz: my issue is that i am not passing anything into $r.()
moritz: github.com/leto/plparrot/commit/d80...28d3c98864 09:24
moritz dukeleto: you should really obtain the return value from eval, and invoke that with arguments 09:27
dukeleto: or look into Perl6::Compiler, maybe that's easier to use from the outside 09:28
dukeleto moritz: i am getting the return value from eval, but i guess i don't know how to pass the arguments in, yet 09:31
moritz: Perl6::Compiler may be a nicer solution to this
moritz if you just put the sub { ... } into the eval, the return value is the sub
which you can invoke
dukeleto moritz: yes, that is what I am doing, but my parcel is a PIR variable. how do I reference it in a bit of Perl 6 code that is a PIR string? 09:32
moritz in the perl 6 code, you write a parameter into the sub 09:33
p6code - 'sub f($x) { say $x.perl } ' 09:34
s/-/=/
$P0 = eval(p6code)
$P0($parcel)
dukeleto moritz: i think i am doing that, but obviously not correctly. $P0($parcel) does not seem to work 09:37
moritz: i have to invoke the sub in the p6code, or else it just defines the sub but never runs it
moritz unless you return the sub from the eval, and invoke it from the outside 09:38
dukeleto moritz: and that is where i run into the issue of not having access to the PIR variable in the p6 codestring
moritz: yes, that is what I should do
moritz: i am doing $P1 = $P0.'compile'(code), and then $P2 = $P1(parcel), is that what you mean? 09:40
moritz yes
oh wait 09:41
maybe you need
$P2 = $P1()
$P2(parcel)
dukeleto moritz: i think that may be right, because I was getting a Perl6Sub returned 09:42
moritz: WOOOHOO! It worked. I just passed variable from Postgres -> Parrot -> Perl 6 -> Parrot -> Postgres !!!!! 09:44
moritz \\o/ 09:46
dukeleto moritz: very very exciting. Just in time for R* ! 09:48
moritz: i just gave you a commit bit to PL/Parrot, just in case :) 09:50
moritz "he was helpful, let's suck him into our code base" :-) 09:51
dukeleto is so excited that he goes to bed
moritz: indeed!
moritz good night 09:52
dukeleto moritz: a great night indeed. see ya on the flip side
dalek parrot: 704050d | dukeleto++ | t/sql/plperl6.sql:
Add a test for returning a function argument in PL/Perl6
parrot: e018636 | dukeleto++ | plparrot.c:
Create a PMC_ISA macro
parrot: 2985ab3 | dukeleto++ | plparrot.c:
Return 0 for Proxy or Any objects
parrot: 731c74f | dukeleto++ | t/sql/plperl6.sql:
Refactor t/sql/plperl6.sql and add a test for multiple arguments
parrot: 690db53 | dukeleto++ | t/sql/plperl6.sql:
Add another test to PL/Perl6 about argument handling
parrot: 44a985f | dukeleto++ | plp (2 files):
Attempt to pass arguments in Rakudo
parrot: 29a86df | (Jonathan "Duke" Leto)++ | (2 files):
Merge branch 'master' into perl6_args

   plparrot.c
parrot: d80522c | (Jonathan "Duke" Leto)++ | (2 files):
Convert Parrot arrays to Perl 6 arrays, moritz++
parrot: bc5452c | (Jonathan "Duke" Leto)++ | (2 files):
Marshall data into Perl 6 from Parrot, moritz++

defining a subroutine in Perl 6, and then executing and passing arguments into it from PIR, instead of attempting to do that from Perl 6.
11:08 mj41_ joined 11:23 mj41_ joined, ttbot joined 11:32 preflex joined 11:33 mj41_ joined 11:53 Austin_Hastings joined 11:59 whiteknight joined 12:19 ruoso joined 12:28 bluescreen joined
Coke hey, IRC spam in parrotsketch. 12:46
Can someone who understands IRC permissions whitelist people for that channel ?
msg chromatic - not only do I occasionally complain about -tn, I occasionally fix it, too!
purl Message for chromatic stored.
Coke partcl is also github.com/partcl/partcl-nqp 12:48
purl okay, Coke.
Essobi Morning. 12:59
Coke guten morgen.
13:06 bacek joined 13:08 aloha joined 13:11 whiteknight_ joined
atrodo cotto> (about no one else editing the lorito document) I almost edited it, but then I realized I'd only be able to answer on how I would implement it 13:24
13:55 jjore_ joined 14:01 NOTevil joined 14:03 bubaflub joined, bubaflub left 14:06 jan joined 14:12 plobsing joined
dalek rrot: r48200 | NotFound++ | trunk/src/pmc/capture.pmc:
mininal refactor of Capture.mark, no functional changes
 14:17
14:20 patspam joined, bubaflub joined 14:27 NotFound left, NotFound joined
ttbot Parrot trunk/ r48200 MSWin32-x86-multi-thread make error tt.taptinder.org/file/cmdout/360257.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 14:27
NotFound Coverity of string/api.c 14:29
/* ZZZZZ workaround for something not setting up encodings right */
6954
A lot of a times for such workaround.
dalek rrot: r48201 | NotFound++ | trunk/src/string/api.c:
use charset compute hash function when encoding has no its own, instead of blindly working around
 15:08
ttbot Parrot trunk/ r48201 MSWin32-x86-multi-thread make error tt.taptinder.org/file/cmdout/360314.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 15:10
dalek kudo: 8049d92 | Coke++ | README:
Apply spirit of RT #70642, Courtesy Scott Penrose
 15:21
kudo: 3fbd628 | (Tadeusz Sośnierz)++ | (3 files):
Added IO::Stat, and a .stat method for IO
 15:27
kudo: 7e8134c | moritz++ | tools/contributors.pl:
[tools/contributors.pl] avoid infinite recursion, general cleanup
15:32 Austin_Hastings joined 15:34 Austin_Hastings_ joined
dalek nxed: r583 | NotFound++ | trunk/examples/fly.winxed:
make sure point of view is set before first draw in example fly
 15:36
16:03 chromatic joined
cotto ~~ 16:07
16:08 JaphMalik joined
Coke ~~ 16:10
dalek rrot: r48202 | NotFound++ | trunk/t/pmc/exceptionhandler.t:
test ExceptionHandler get severity values
 16:15
16:19 davidfetter joined
ttbot Parrot trunk/ r48202 MSWin32-x86-multi-thread make error tt.taptinder.org/file/cmdout/360365.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 16:24
16:27 theory joined
dalek parrot: 66d9f9c | (Jonathan "Duke" Leto)++ | t/sql/plperl6.sql:
Make the fibonacci test take an input argument
 16:35
16:37 darbelo joined
dalek rrot: r48203 | darbelo++ | branches/gsoc_nfg (24 files):
Sync with trunk
 16:50
16:55 nopaste joined, robin-gvx joined
cotto_work ~~ 16:58
atrodo hey cotto_work 17:06
purl cotto_work is always at work
cotto_work you might think that.
You'd be correct.
kthakore hehe
atrodo i had always wondered
kthakore cotto_work: what do you do?
atrodo "is he really at work, or is he just messing with everyone?"
NotFound Maybe his job is mess everyone. 17:07
kthakore atrodo: messing while at work
NotFound++ beat me to it
dalek kudo: fe29cde | Coke++ | README:
README should be in text, not POD.
 17:09
17:18 rurban_ joined 17:19 hercynium joined 17:27 macroron joined
dalek parrot: d70f086 | (David Fetter)++ | (4 files):
s/CREATE FUNCTION/CREATE OR REPLACE FUNCTION/
 17:28
17:31 bluescreen joined 17:36 tcurtis joined
dalek rrot: r48204 | fperrad++ | trunk/config/gen/config_pm.pm:
[win32] fix build (after r48179), handle backslash in string
 17:40
17:41 ash_ joined 17:45 [1]Casan joined
dalek parrot: 73373be | (David Fetter)++ | (3 files):
Drop languages before creating them.
 17:46
jnthn
.oO( And throw mugs before creating languages. )
 17:47
davidfetter heh
Coke I have a cache of nice mugs in my laundry room I could donate to the cause. 17:48
davidfetter what's shakin', jnthn ?
Coke s/nice/lame/
jnthn davidfetter: Earthquakes? :-)
purl UTC. , (). Magnitude , Depth km.
davidfetter in earthquake country
chromatic NotFound, where are you seeing new Coverity results? 18:03
NotFound chromatic: tapir2.ro.vutbr.cz/cover/cover-results/ 18:04
chromatic Oh, coverage. I thought you meant Coverity Scan. 18:05
NotFound Haven't gone that far,Ƨ
coverage gives enough places to work 18:06
chromatic Their scans haven't resumed yet sadly.
PerlJam chromatic: Are you going to make your one article a week on perl.com? 18:12
chromatic I have some Rakudo interviews to publish 18:13
jnthn \\o/
chromatic: I did actually get around to sending you mind, didn't I?
chromatic I believe so.
confirmed, jnthn.
jnthn OK, great.
18:14 jjore joined
jnthn I remember starting it clearly, then leaving it for a while, and couldn't quite remember if I'd got back to finishing it up and sending it. :-) 18:14
PerlJam chromatic: Will there be an article conincident-ish with the release of R* tomorrow? 18:16
chromatic I don't have anything ready for it, but I'd take something if someone or someones could make one. 18:17
PerlJam Whenever pmichaud finishes with the tarball, I may write something about obtaining and installing R* with an overview of what it comes with. 18:18
chromatic I'll work up a few paragraphs on what to expect and why it matters.
PerlJam chromatic++ 18:19
cotto_work I suspect that TT #1127 is due to compilers/imcc/pbc.c +2266 18:47
dalek rrot: r48205 | NotFound++ | trunk/config/gen/config_pm.pm:
set binary charset when build directory contains non ascii chars, first step to TT #1717
18:50 hercynium joined
NotFound I wish this game for chistmas: www.nds9.com/nds_games/TCG/4238___M..._4238.html 19:07
cotto_work chromatic: does replacing IMCC_INFO(interp)->ins_line with oldsize on that line look like a legitimate fix? 19:09
chromatic cotto_work, I don't understand the function well enough to say. 19:10
I can convince myself it's plausible.
cotto_work me too, but I'm gullible.
chromatic If all tests pass.... 19:14
cotto_work I'll play with it some when I get home tonight.
if nothing else, the debug segments look significantly less wrong
chromatic Less wrong is good. 19:15
We should rethink the debug mapping at some point; that'll speed up the profiling core substantially.
19:17 mtk joined
cotto_work all tests pass 19:18
purl Time to write more tests!
cotto_work At the very least, I want to write a test case for this. 19:19
19:29 cognominal joined
dukeleto is there a ticket for many tests failing under "make cover" yet, or should I make one? 19:30
the coverage tools fiddle with the output and cause some test failures 19:31
19:33 AndyA joined
darbelo Ship it! 19:35
19:35 eternaleye joined
darbelo Gah. Wrong window. 19:35
whiteknight purl Ship it! is <reply>Ship this piece of garbage? Better idea is to go back and write a lot more tests 19:41
purl OK, whiteknight.
NotFound "Returns -1. It makes no sense to try and search for a substring in raw binary data" What? 19:44
darbelo NotFound: binary is the non-charset. It just holds data. 19:48
NotFound darbelo: for me it makes sense to search for data in data. 19:49
chromatic Certainly for magic (5)
darbelo Well, yeah. But AFAICT binary tries to discourage it's own usage. 19:50
"Hi, I'm the binary charset. Don't use me!" 19:51
NotFound darbelo: is near to reach that purpose with me, but I'm trying to resist.
I think I'm going to setup a stupid workaround.
dalek parrot: e8d44d9 | (David Fetter)++ | html/index.html:
Updates for Rakudo Perl 6
darbelo Or we could implement searching in binary strings.
NotFound Let me know when it's done ;) 19:54
darbelo I'll put it on my GSoC TODO.
NotFound In the meantime I will try some workaround for TT #1717 19:55
dalek parrot: 4b9ccb4 | (David Fetter)++ | html/template.html:
Added template.html
 19:57
19:58 snarkyboojum joined 20:01 rurban_ joined
dalek parrot: e9db30b | (David Fetter)++ | html/index.html:
Divided PLs out into their own pages
 20:03
pmichaud phone 20:05
20:05 bubaflub joined
Coke call. 20:05
mmheheh.
yay, someone submitted a patch to Modern::Perl instead of just complaining! 20:06
dalek parrot: 8c87e64 | (David Fetter)++ | html/ (2 files):
Broke out language-specific files
 20:09
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#35144), fulltest) at r48205 - Ubuntu 10.04 amd64 (g++ with --optimize) 20:18
darbelo NotFound: Amusingly enough, the ascii rindex (reused by iso-8859-1) looks suitable for use in binary :) 20:27
If the other functions in that class are the same you'll have a patch real soon.
NotFound darbelo: several ascii functions are suitable for moving to "common" or something alike.
darbelo It's a thin wrapper around Parrot_byte_[r]index() in src/utils.c really. 20:30
Easy enough to share for now.
Hah! tests fail if I make binary dtrt. 20:45
dukeleto PL/Parrot now has irc logs: irclog.perlgeek.de/plparrot/today moritz++
20:46 wagle joined
moritz btw if you want other channels logged too, just ask me 20:47
it's very little effort to add one (especially if it's on magnet or freenode)
nopaste "darbelo" at 192.168.1.3 pasted "Make binary string [r]indexable." (101 lines) at nopaste.snit.ch/22401 20:50
darbelo NotFound: Does nopaste.snit.ch/22401 do what you need? 20:51
NotFound darbelo: I think so, but I doubt such a change will be accepted without a deprecation cycle. 20:52
darbelo That's why I nopasted instead of committing. 20:53
This wa related to a ticket, right? 20:54
I can attach it there. 20:55
NotFound TT #1717 but is not trivial to see the relation
darbelo Oh, we started treating treat non-ascii $PATH data as binary for config purposes so we could fix building on widows with CJK paths. 20:58
cotto_work That seems to make partcls 21:00
NotFound darbelo: that was just the first step. Working on the second I found this problem.
cotto_work 's profile less broken.
darbelo I can open a RFC ticket for the binary changes. 21:02
Other than that, I can't think of any real solutions. 21:03
NotFound darbelo: I think I've found a good workaround using Parrot_byte_rindex, thanks for the idea. 21:04
21:06 hercynium joined
NotFound This is a mess anyway, we need a better way to deal with file and directory names. 21:07
darbelo If Configure could find out what charset/encoding it's getting path data in half of this problem would go away. 21:11
chromatic Interpolating raw strings into a template for config_lib.pir is half of the problem. 21:12
NotFound darbelo: I can easily create a path than makes any guess fail.
darbelo NotFound: That's my point. Guessing isn't the solution. 21:13
NotFound And using binary isn't the solution, because people expect to able use filenames as strings, and parrot binary string have strong limitations. 21:14
darbelo Configure is the one writing that string down. It should also write down what the encoding/charset that string is in. 21:15
chromatic Should Configure look at Perl 5's locale settings?
NotFound darbelo: And what is? 21:16
I can write a path with some directories in latin-1 and some in utf8.
In fact, I have several like that at my job. 21:17
I have a workaround that builds parrot and pass tests except several from pct_complete_workflow 21:18
Checking now if it passes with ascii build path
The ones that fail gives "Malformed string" at some point, as one can expect. 21:19
All test pass with ascii path 21:25
pmichaud anyone who is interested in seeing (and commenting on) early drafts of the R* announcement: pmichaud.com/sandbox/2010.07.txt (and join freenode/#perl6 if you have comments or suggestions) 21:28
cotto_work I appreciate the double meaning of ... in the feature list 21:30
probably unintentional given that it's in both lists 21:32
NotFound nopaste? 21:33
purl nopaste is nopaste.snit.ch (works with the script in $_PARROT/tools/dev/nopaste.pl) or paste.scsys.co.uk or www.extpaste.com or gist.github.com or App::Nopaste or codepeek.com/paste/ or (: pastebot)
nopaste "NotFound" at 192.168.1.3 pasted "Workaround for TT #1717" (27 lines) at nopaste.snit.ch/22402 21:34
NotFound Opinions? 21:35
purl Opinions are like assholes. Everyone has one and it's usually full of shit.
chromatic I don't understand it. 21:36
NotFound CHARSET_RINDEX does not work with binary charset, so it short-circuits it. 21:37
chromatic It's just looking for a dot?
NotFound chromatic: yes 21:38
chromatic I was overthinking it then.
As a workaround it's not awful.
It'd be nice to avoid that strcmp if possible, but it's probably not.
NotFound A dot after the last slash
chromatic: no easy way unless accessing string internals. 21:39
And library.c already does that too much.
chromatic Can we make CHARSET_RINDEX work with binary charsets?
NotFound chromatic: that was last nopaste from darbelo 21:40
chromatic Fixing CHARSET_RINDEX seems cleanest.
NotFound chromatic: yes, but maybe pople want a deprecation notice for that. 21:41
21:42 bacek_ joined
chromatic For adding a feature? Crazy talk. 21:42
21:43 aloha joined
NotFound chromatic: look at charset/binary.c, it does not work on purpose. 21:43
chromatic Right, but that's not a deprecation event as I see it. 21:44
NotFound I fear the deprecation policy chainsaw X-) 21:45
cotto_work My chainsaw seems to have backfired on me. 21:47
ow
darbelo decides to risk amputation. 21:48
cotto_work Is it likely to break anything? 21:49
darbelo AFAICT, no.
Except code like the last hunk of my nopaste. 21:50
GeJ Bonjour everyone. 21:51
cotto_work I'm inclined to go with chromatic. If we're making code work that didn't work before (as opposed to ripping something out), that doesn't sound like a deprecation.
chromatic ca va, GeJ?
cotto_work How do normal humans even keep up with #perl6?
darbelo The Chainsaw has spoken! 21:52
darbelo JFDIs
cotto_work I guess it's a big day for them though.
Tene cotto_work: I've never had a problem with it... just check in and catch up occasionally.
chromatic Laundry day. See you there. Rakudo Star... tum-b-ling. 21:53
darbelo There. r48206. Revert if anything breaks.
GeJ Hi chromatic. I've been better. Still in convalescence at home.
chromatic You can type, right? We saved a pile of tickets for you.... 21:54
NotFound darbelo: building it 21:55
purl building it is alright...if you do it only once a day
GeJ I could say that staying sitted for a while brings the pain back, but I'm so bored that I might eventually do something useful. :)
chromatic Boredom is the worst part of recovery. 21:56
GeJ I'd put it second, right after recurring pain.
chromatic For me it's boredom, killer bees, and pain. 21:57
NotFound There are drugs for the pain, but not for the boredom.... Legals drugs, ar least.
darbelo sings 'Lucy in the Sky with Diamonds' 21:58
GeJ I fail to find a logical link between recovery and killer bees. But that may be because of the fuzzy warmth induced by the pills I took earlier.
darbelo GeJ: It depends on the type of recovery. 21:59
NotFound With r48206 all test pass bulding in a utf8 path except t/compilers/pct/complete_workflow.t Failed tests: 9, 18, 27, 36, 45, 54 22:00
darbelo NotFound: Is that better, worse or the same as your workaround? 22:03
NotFound darbelo: same
dalek rrot: r48206 | darbelo++ | trunk (2 files):
Make index and rindex operations work for the 'binary' charset.
 22:07
darbelo Nothing strikes me as obviously wrong in that test. But there could be somethin subtly wrong I'm overlooking. 22:08
NotFound darbelo: probably concatenation of binary with other encoding strings. 22:09
darbelo I thought I had fixed the concatenation stuff already. 22:10
NotFound No, it comes from unescape 22:14
This is the problem: push args, "$TEST_DIR/complete_workflow_sample_input.txt" 22:17
my $TEST_DIR = "$BUILD_DIR/t/compilers/pct";
This fixes it: push args, binary:"$TEST_DIR/complete_workflow_sample_input.txt" 22:19
darbelo Yep. That does it. 22:21
NotFound I don't like to set that inconditionaly... working on a check. 22:22
darbelo NotFound++ 22:24
NotFound my $BD_CHARSET = $BUILD_DIR =~ /[^[:ascii:]]/ ? "binary:" : q{}; 22:26
Is that idiomatic perl?
darbelo NotFound++ 22:27
chromatic Idiomatic enough. 22:28
darbelo Say, how old a perl are we supporting nowadays?
5.8 ? 22:29
purl rumour has it 5.8 is fine. or www.catalystframework.org/calendar/2008/16
NotFound Done, r48207 22:30
22:33 lucian joined
dukeleto darbelo: i think it is 5.8.0, but i don't know for sure 22:34
chromatic I thought somewhere in the 5.8.6 myself.
dukeleto 5.8.0 introduced qr//, which Parrot uses, so it is definitely >= 5.8.0 22:35
dalek rrot: r48207 | NotFound++ | trunk/t/compilers/pct/complete_workflow.t:
fix pct workflow test for non ascii build path
 22:40
22:48 whiteknight joined
dalek rrot: r48208 | darbelo++ | branches/gsoc_nfg/src/string/api.c:
Remove a reference to charset hashing functions that got imported from trunk.
 22:57
rrot: r48209 | darbelo++ | branches/gsoc_nfg/src/string/encoding (3 files):
Add utf hashing. We can't fallback to the charset anymore.
rrot: r48210 | darbelo++ | branches/gsoc_nfg/t/native_pbc (4 files):
Regenerate native pbcs.
23:35 [1]Casan joined 23:58 Psyche^ joined

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!