Parrot 0.9.0 | parrot.org/ | 566 RTs left.
Set by moderator on 31 January 2009.
00:08 TiMBuS joined 00:09 AndyA joined 00:46 Fayland joined 01:14 kid51 joined
s1n was PARROT_FLOATVAL_INF_NEGATIVE (and POSITIVE) replaced by something? 02:20
02:30 arbingersys joined
s1n nevermind, my mistake 02:36
dalek rrot: r36268 | petdance++ | trunk:
bunch of Critic changes
 02:52
03:12 bacek_ joined 03:19 galf joined 03:37 Theory joined 03:38 Theory joined 03:53 mberends joined 04:00 Fayland joined
TiMBuS make[1]: Entering directory `/home/timbus/parrot/docs' 04:36
/usr/bin/perl -I../lib -MParrot::Distribution -e "Parrot::Distribution->new()->delete_html_docs()"
syntax error at ../lib/Parrot/Distribution.pm line 814, near "}"
Compilation failed in require.
i get this running make clean, is it a known issue?
04:48 Andy joined
GeJ I get the same one during "make html" 04:54
looks like the trailing semicolon is at fault 04:56
04:57 Fayland joined
s1n Coke: i'm getting the merry-go-round with this rt access thing 05:01
05:16 Tene joined 05:27 Theory joined
dalek rrot: r36270 | allison++ | trunk/lib/Parrot/Pmc2c/PMCEmitter.pm:
[cage] Ripping out a chunk of code from the old MMD system that isn't used
 05:32
rrot: r36271 | allison++ | branches/pdd28str_part2:
[pdd28str] Removing old 'string_repeat' function, since everything is already
 05:46
05:56 jrockway joined
dalek rrot: r36272 | allison++ | branches/pdd28str_part2:
[pdd28str] Renaming 'string_substr' to 'Parrot_str_substr'.
 06:04
06:24 rafl joined
dalek rrot: r36273 | allison++ | branches/pdd28str_part2:
[pdd28str] Renaming 'string_replace' to 'Parrot_str_replace'.
 06:31
rrot: r36274 | allison++ | branches/pdd28str_part2:
[pdd28str] Renaming 'Parrot_str_chopn' and 'Parrot_str_chopn_inplace'.
 06:49
07:10 uniejo joined 07:32 iblechbot joined
dalek rrot: r36275 | allison++ | branches/pdd28str_part2:
[pdd28str] Renaming 'string_compare' to 'Parrot_str_compare'.
 07:33
rrot: r36276 | allison++ | branches/pdd28str_part2:
[pdd28str] Renaming 'string_equal' to 'Parrot_str_equal'.
 07:51
08:49 skv___ joined 08:58 skv_ joined 09:02 Gerd joined 09:25 alvar joined 09:37 skv_____ joined 09:42 bacek joined 09:47 tomyan joined 10:13 masak joined 10:14 skv joined 10:42 kj joined 10:52 bacek joined
dalek rrot: r36277 | rurban++ | trunk/lib/Parrot/Distribution.pm:
revert r36276, illegal syntax keys %{$generated}
 10:58
11:00 skv joined 11:05 register joined
register I can't find clear documentation about what this command is supposed to do: parrot.exe Perl6Grammar.pir --output=PGE\\builtins_gen.pir PGE\\builtins.pg 11:06
kj register: it runs the program Perl6Grammar, giving it an option --output, and a value for that option (the value after =), and then the argument for the program Perl6Grammar 11:08
TiMBuS beaten while typing it
:< 11:09
kj perl6grammar is a compiler that translates the .pg file (a grammar file) into PIR
register ok
but the output option should contain only .pbc or .pasm extensions right? 11:10
at least this is what is writtend in the docs
TiMBuS its the output of Perl6Grammar.pir
kj no 11:11
a .pg file is a subset of perl6, basically
so you need to translate it into PIR
output extensions should only be .pbc (and pasm, in theory ,but that doesn't work properly), if you run the file directly into parrot 11:14
register ah gotcha. I mixed it up the parrot --output option with the Perl6grammar option. Now i got it. thanks
kj that means, you only want to use the parrot assembler; in this case, you run a program on parrot, that does the main thing, namely compiling a .pg file into PIR. That, can then be compiled to .pbc, if you want 11:15
oki
11:19 iblechbot joined 11:22 tomyan left 11:23 tomyan joined
dalek rrot: r36278 | jkeenan++ | trunk:
Eliminate some quotes in error messages to quiet perlcritic and allow tests to once again pass.
 11:42
11:43 skv joined
s1n Coke: rbrt said: "Coke can add you. Patrick can too.  They may not remember that the Perl6 and Parrot queues use the same access controls.  While I administer the system, I try not to be involved in the day to day management of who has access to Perl6/Parrot bugs." 11:51
11:53 jimmy joined 12:16 donaldh joined 12:22 kid51 joined
nopaste "kid51" at 24.188.182.149 pasted "Syntax errors in Parrot::Distribution" (7 lines) at nopaste.snit.ch/15480 12:23
kid51 I think this should do it. 12:29
dalek rrot: r36279 | jkeenan++ | trunk/lib/Parrot/Distribution.pm:
Correct syntax error; eliminate extra semicolon.
 12:30
12:35 AndyA joined 13:31 davidfetter joined
dalek rrot: r36280 | rurban++ | trunk/docs/dev/byteorder.pod:
[docs] byteorder.pod

  - add Low level FLOATVAL fetch and convert functions
  - add reference to src/packfile/pf_items.c
  - add pod coda		 13:44
Coke s1n: (same access controls) no, I certainly did not realize that. 13:50
s1n: what is your rt id?
13:51 skv joined
Coke (follow the steps in docs/project/meta* and let me know when you're set.) 13:51
dalek rrot: r36281 | NotFound++ | trunk/src/pmc/class.pmc:
fix Class pmc vtable remove_attribute declaration
 13:57
rrot: r36282 | cotto++ | trunk/tools/build/nativecall.pl:
[pmc] PMC_int_val -> VTABLE in src/nci.c
 13:59
cotto I like playing with code generators. 14:01
Then again Parrot itself is one giant code generator generator. Maybe that why I'm here. 14:02
Coke (code generators) Really? I have an RT for you, then. 14:03
cotto woohoo! 14:04
Coke rt.perl.org/rt3/Ticket/Display.html?id=39313 14:05
cotto ooh. an oldie
Coke the pmc2c generator drops things on the floor.
cotto nice feature, that
Coke so you might be trying to override something, and have your definition not show up in the resulting .c
14:05 Debolaz joined
Coke /and/ not error out. 14:05
I think if we just do a passthrough of everything, and have the C compiler fail on things it doesn't understand, that's fine.
er, of everything that isn't transformed. 14:06
cotto that could help make PMCs less confusing
Coke that ticket was only opened 972 days ago. =-) 14:07
if you hurry, you can close it before it hits 1000. =-)
moritz 365*3 14:08
purl 1095
cotto It's kinda sad that there are a couple dozen tickets older than that. 14:10
14:12 gryphon joined 14:14 Whiteknight joined
dalek rrot: r36283 | cotto++ | trunk/src/dynpmc/rational.pmc:
[pmc] use VTABLE functions instead of PMC_int_val
 14:14
Coke msg particle rt.perl.org/rt3/Ticket/Display.html?id=44393 has been waiting for your input since for about a year and a half. Can you please comment on the ticket? 14:15
purl Message for particle stored.
Whiteknight Coke, I think that's a duplicate of RT#46405. I'm on it 14:24
I've got a fix in mind, I'm just looking for a test case to prove once I've fixed it
Coke Whiteknight: you mean the one I pinged particle on? 14:38
Whiteknight yeah
Coke have the failing tests showing up on linux been addressed? 14:43
What do folks think about the parrotcode blog vs. the parrot.org site? 14:46
Whiteknight Coke: which failing tests? the t/native_pbc/ ones? 14:47
kj Coke: the parrotblog.org site you mean?
Coke Whiteknight: smolder.plusthree.com/app/public_pr...port/17732
shorten Coke's url is at xrl.us/beenkw
Coke kj: yes.
kj I wouldn't mind if it would be integrated
to have a more unified communication channel
parrotblog.org could still link to it
Coke mmm. I know unified isn't all web2.5 ...
Coke is pondering an article, and will probably drop it on parrot.org 14:50
kj but parrot.org is more of a official news thingy 14:51
while people like me, with no official status, can just drop stuff on parrotblog...
Coke that's not a terrible distinction. 14:52
14:52 PacoLinux joined
Coke (no official status) you're a committer. =-) 14:52
kj well, dunno. If i'm writing on gory details of how to this or that while implementing, say, a PIR compiler :-) 14:53
Coke yay, PGE build is segfaulting for me on feather. 14:56
Infinoid the test_pbc_annotation_persistence rule in the Makefile isn't -j safe. During make test -j, it's executed *way* before parrot is actually built. 15:02
moritz yes, I opened a ticket for that 15:03
Infinoid I think I can fix it. What's the ticket number?
kj Coke: (parrotblog) for a short-term solution, making parrotblog the same style (same logo etc.) would at least communicate a closer link relationship between parrot.org and parrotblog.org 15:05
moritz Infinoid: TT#231
Infinoid moritz: Thanks, fix is on the way. 15:12
moritz Infinoid++ 15:13
Infinoid (just fixing up some codingstd stuff too) 15:14
dalek rrot: r36284 | Infinoid++ | trunk/config/gen/makefiles/root.in:
[makefile] The t/op/annotate.t test can't run before parrot is built.

This fixes TT #231.		 15:21
rrot: r36285 | Infinoid++ | trunk:
[cage] Fix codetest failures from tabs.t, c_code_coda.t and c_header_guards.t.
15:25 tomyan left 15:26 tomyan joined 15:32 tomyan left
cotto Is there a way to get an RSS feed for changes to the trac wiki? 15:43
15:44 skv joined
dalek rrot: r36286 | cotto++ | trunk:
[pmc] unionval -> VTABLE for in-tree users of the Integer PMC
 15:46
15:47 khisanth_ joined
Whiteknight cotto: I can't find an RSS feed for it, no 15:51
lathos Here's a sample was to do it: dev.simon-cozens.org/songbee/timeli...format=rss 15:52
shorten lathos's url is at xrl.us/beenwe
lathos Replace for whichever trac you're using. 15:53
15:54 Theory joined
cotto it looks like trac needs an extension to support that feature 15:58
Infinoid Now t/native_pbc/number.t test 3 is failing on x86-64. (Previously, #1 and #2 were failing, but they've since been TODOed.) Looking at the test, it seems like it's supposed to be x86-64 specific. 16:00
particle i've got 1-3 failing msvc 16:01
Infinoid That's on x86-32, right?
You might want to mention that on TT #254, it sounds like they've been only paying attention to linux and darwin. 16:02
particle yep
rg then at least i'm not the only one who didn't find that rss feed ;). But someone told me where to find it: trac.parrot.org/parrot/timeline 16:07
Infinoid Right, that's the global rss feed. I was looking for a wiki-specific one a while ago, but couldn't find one. 16:08
rg you can set the options to only show wiki changes
lathos And if you add ?wiki=on&format=rss you get the address that I said a few moments ago.
Infinoid lathos++ rg++ 16:09
particle Infinoid: is this for a bot?
Infinoid Yeah, I've had a couple of requests for dalek posting wiki updates in here
Coke particle: can you run the command " make pdump; ./pdump t/native_pbc/integer_1.pbc" for rurban?
(he's asked everyone else who is reporting problems to do that.) 16:10
particle ah, wilco
Infinoid++
Infinoid (I haven't actually *done* anything about those requests, yet.)
16:14 register joined
register how do I modify a trac ticket? 16:14
Coke login?
purl hmmm... login is still checked in the database
particle do you have a trac account at parrot.org?
register I don't think so 16:16
should I ask someone?
rg you can simply register yourself 16:17
Coke (upper right hand corner, click on "register"). be sure to confirm your email address. 16:20
Coke wonders if rurban's recent change broke the PGE build. 16:21
16:21 Andy joined
basic Infinoid (or anyone else): want to test email2trac again? 16:30
Infinoid basic: sure, one moment.
basic thanks 16:32
Infinoid Ok, I've sent another reply to tickets@parrot.org. No indication of it on the ticket page yet, though.
16:32 Tene_ joined, Zaba joined
basic doh, okay, one more combination of settings to try 16:33
Infinoid No problem, lemme know.
16:36 rurban joined
cotto Coke, does that error you just reported persist after a make reconfig? 16:37
#261
Infinoid rurban: Hi! t/native_pbc/number.t started failing for me again on linux/x86-64... but this time it's test #3. Looking at the test, I guess it's supposed to be native for me. What can I do to help sort this out?
particle after a realclean, you mean. i was just going to ask that 16:38
cotto or that
particle the ticket has 'perl Configure.pl' in it
i'm not sure that's enough to regenerate the pbc, though 16:39
16:39 hercynium joined
rurban Infinoid: strange x86-64 native works fine for me. 16:39
can you paste the result
Coke cotto "what is "make reconfig" ?
Infinoid Sure, one moment.
Coke cotto, yes.
looking again at #260, those might be the same issue.
cotto gah 16:40
rurban I just fixed the 64-bt alignment bug in directories and ound another one with string alignments...
Infinoid rurban++
cotto make reconfig is the same as make clean && perl Configure.pl, afaict 16:41
Infinoid basic: If you haven't already seen this, I got another "User unknown in virtual alias table" error from your postfix in reply to my most recent test.
particle is ajr here? 16:43
basic yeah, i just sent a config change to the relays, they should update within the hour. Hopefully that fixes it :)
nopaste "Infinoid" at 96.238.213.50 pasted "t/native_pbc/number.t failure on gentoo/amd64 for rurban" (67 lines) at nopaste.snit.ch/15481
cotto Coke, if you want to reproduce the segfault, you also need to run the command to generate PGE.pbc before running gdb
dalek rrot: r36287 | whiteknight++ | trunk/src/tsq.c:
[Core] Add stub file description to tsq.c
16:43 skv joined
Infinoid rurban: There's your nopaste. It doesn't crash, it just has different output... 16:43
cotto Coke, is the problem caused by r36286 or does it appear before then? I saw similar problems trying to convert the Integer PMC to use ATTRs, but I made sure not to commit that code. 16:44
dalek rrot: r36288 | whiteknight++ | trunk/src/gc/resources.c:
[Core] Add some basic file documentation to src/gc/resources.c
Infinoid basic: Ok, I can try again in an hour.
rurban gentoo/amd64 hmm. I saw this error before
cotto (active day today, btw) 16:45
Infinoid cotto: Happy Monday!
basic thanks, i've gotta run to class
dalek rrot: r36289 | whiteknight++ | trunk/src/packout.c:
[Core] add some basic file-level documentation to packout.c
Infinoid basic++, thanks!
rurban looks like another double float issue. maybe I should create for each failre a new ticket? we had a wrong 64bit advance, a wrong 64bit ALIGN_16, and a wrong 64bit string ROUND_UP. 16:46
and now a 64bit native double float issue. should not happen
Infinoid Our long-term goal is to be able to support every format from every architecture, right? Maybe the tests should be laid out with that in mind. 16:47
(That said, I'm not very familiar with the current state of things.)
rurban sure. what is missing is 64bit big-endian
dalek rrot: r36290 | whiteknight++ | trunk/src/dynext.c:
[Core] Add basic file-level documentation to dynext.c
 16:48
rurban that woudl be test _4.pbc
Infinoid Hmm. What's an example of that? Tru64?
No, that's alpha (LE).
rurban really don't know, could be
Infinoid MIPS64 could be run in big endian mode, I suppose. I don't know anyone who has one, though.
rurban and there's another weird arch for floattype 1 16:49
there's even transformer code missing for that.
it's "x86 little endian 12 byte long double" 16:50
that would be test _5.pbc
Infinoid We need to bribe someone with a sparc64 box. 16:51
rurban maybe I can produce a vmware image for this arch. I already have the boot dvd as iso around 16:52
my last image ran out of discspace though. 3gig was not enough and zfs is weird
I had to destroy a interim snapshot to be able to rm some files 16:53
Infinoid Last time I tried to emulate sparc, I ran into the same problem, and it took days to get to that point.
Coke cotto; I have been bitten before by the 2 step juggle there in PGE; I'd much rather see a true intermediate file.
Infinoid I think I was emulating sparc32 and trying to install some opensolaris distro, though. 16:54
rurban I'll try on windows xp with sparc64 opensolaris 10 2008.05
Coke cotto: Revision: 36204 16:55
rurban but I need the gentoo amd64 cross-compiler first
Coke (the problem seems to exist at least back that far.)
rurban what's the name for the gcc-amd64 on gentoo?
Infinoid uh. "gcc" 16:56
x86_64-pc-linux-gnu-4.3.3
rurban thanks. will buiild now. in the meantime I'll disable the failing tests. we've got enough info now. 16:57
Infinoid Ok, lemme know.
Coke rant: the language/ blocks of languages makes it a PITA to svn up to older revisions.
"hey, you already have languages/WMLscript. *vomit*" 16:58
16:58 flh joined
cotto I'm glad I didn't break it, then. 16:59
I've broken enough already.
Infinoid I haven't broken nearly enough yet. I'll get right on that. 17:00
Coke wonders who installed parrot on feather. 17:02
cotto Infinoid++ 17:06
Coke cotto: doing an svn-bisect now to see when PGE started barfing.
Infinoid What's everyone's favorite date-handling CPAN module these days? I need to convert "Sun, 01 Feb 2009 23:06:36 GMT" into something I can sort() by. 17:08
(this is for dalek's tracwiki logging)
particle the DateTime folks will tell you theirs is best 17:09
rg likes DateTime, too
dalek rrot: r36291 | rurban++ | trunk:
More 64-bit pbc fixes for TT #254:

  - add lots of debugging code, enable with TRACE_PACKFILE 2
  - mark more failing tests for 64-bit.
   interestingly t/native_pbc/number_3 fails native on gentoo-amd64,
   but not on solaris-64int.		 17:10
17:12 rblackwe joined
rurban Infinoid: and what is the gentoo packager command ? 17:14
moritz "ebuild"? 17:15
rurban thanks, portage and ebuild
Infinoid emerge
moritz e-xecute 17:19
rurban I think I have to emerge system and emerge world also, this will need a while... :) 17:20
Infinoid Oh, cool, XML::RAI converts the rss date into a sane format automatically. 17:22
(XML::RAI)++
Coke 11 revisions left to check... 17:30
Infinoid Git and Subversion both have the nice advantage of global revisioning; the wiki has per-file revisioning. So I'm not sure dalek's current format is appropriate for wiki changes. 17:32
So what would you guys prefer?
tracwiki: r2 | coke++ | AbandonedLanguages:
tracwiki: trac.parrot.org/parrot/wiki/Abando...?version=2
shorten Infinoid's url is at xrl.us/beeoea
Infinoid or
tracwiki: coke++ | AbandonedLanguages version 2
tracwiki: trac.parrot.org/parrot/wiki/Abando...?version=2
shorten Infinoid's url is at xrl.us/beeoea
Infinoid or something else entirely?
moritz I like the r2 in front 17:34
s/in /up/
Infinoid Yeah, the problem is that it varies totally depending on the page being edited.
For instance, right below that would appear:
tracwiki: r45 | coke++ | Languages:
tracwiki: trac.parrot.org/parrot/wiki/Langua...version=45
So it seems a little confusing to me. It does look like the rest of dalek's output though, so I'm not *too* bothered by it 17:35
Coke given the page name is in the url, and you can only edit one at a time, the first line could just be for karma. 17:37
Coke doesn't care too much, though; consistency with other feeds good.
kj what's the difference between a dormant and an inactive language?
Infinoid So how about: tracwiki: coke++ | trac.parrot.org/parrot/wiki/Abando...?version=2
shorten Infinoid's url is at xrl.us/beeoea
particle Infinoid: how about v2 | coke++ | Languages: 17:39
Infinoid What, s/r/v/?
particle yep
Infinoid Ok.
cotto "Beeoea" sounds like the name of a town.
Infinoid also removes the trailing colon, as there's no log following. 17:40
moritz that way the irc logs won't confuse them with svn revisions
Infinoid Ok, new plugin installed. 17:46
17:46 dalek joined
dalek tracwiki: v2 | cotto++ | SandBox 17:49
tracwiki: trac.parrot.org/parrot/wiki/SandBox?version=2
Coke cotto: it was apparently fperrad.
moritz the ticket count in the topic - is that RT + trac? or just RT? 17:51
the weekly summary says it's only 485
in RT
cotto are there any failing tests in t/pmc/class.t? 17:52
kj t/dynpmc/foo.t is failing on win32. *again* 17:54
moritz on linux only t/native_pbc/number.t fails (1-3) 17:56
rurban still, I just committed a todo
?
for libnci_test I have a failing linking on solaris 64-bit. $linkflags is not propagated (-m64 is needed there) 17:58
particle is getting sick of realclean
Coke it says "RTs". (so not trac)
rurban moritz: did you do a svn up for t/native_pbc/number.t ? 17:59
Coke particle: like failures on -j, it means we haven't spec'd our deps properly.
doesn't help that we're building things in Configure.pl that should be generated during make.
moritz rurban: I did, but I'll do again..
particle coke: the makefile templates keep changing, so i need to reconfig all the time now
profiling and optimizing pmc2c and c2str could improve my build time a bunch 18:01
Infinoid I'd like to implement a ccache equivalent for those two. I may have mentioned that before.
particle yeah, that'd be nice
cotto Go for it! I'd love to see build times go down some.
particle or have a daemon run processing requests for them
Coke or we could properly only update things if it was necessary. 18:02
Infinoid Ccache on linux removes the compilation overhead, for the most part. Those two are the biggest remaining processes.
Coke (daemon) overcomplicate much?
particle after realclean, it all needs to be rebuilt
Coke so avoid realclean more often instead.
particle coke: something to prevent the .pm files from being reloaded every invocation
coke: you can't avoid realclean when makefile templates change
they've been changing every day for weeks now, it seems 18:03
Infinoid Loading .pm files happens pretty fast, here. But I guess spawning processes has more overhead on win32, generally.
Coke no, but assuming that's the common case and something we should optimize for is wrong.
moritz ok, I svn up'ped, and have lots of new failures - or do I need another realclean? 18:04
particle try realclean, there shouldn't be lots of failures
Infinoid realcleans by default
Coke moritz: if you have ask, the answer is generally yes.
rurban I did no realclean for my updates, hmm
particle doesn't have ccache
rurban but maybe the string tables changed 18:05
Infinoid Almost always does... 18:06
basic Infinoid: any luck? 18:10
nopaste "moritz" at 91.10.140.202 pasted "Failing tests on r36291 (32 bit linux)" (36 lines) at nopaste.snit.ch/15482
moritz that's after reclean + Configure.pl 18:11
Infinoid basic: the server accepted the message, but I see no change to the ticket
basic: I have no idea whether the stuff I'm sending is formatted correctly, or not.
To: tickets@parrot.org, Subject: Re: [parrot-tickets] [Parrot] #49: install email2trac 18:12
rurban moritz: 64bit? 18:15
purl hmmm... 64bit is a little short. Go for 128bit signed.
moritz rurban: 32 bit 18:16
rurban strange: cannot reproduce the failing number test. can you paste perl t/native_pbc/number.t
moritz rurban: nopaste.snit.ch/15483 18:18
rurban ah! thanks
18:18 skv joined
moritz gotta go now, bbl 18:21
kj rurban: I get the same output as pasted, on win32 18:22
rurban wait a sec, I'll try a clean build 18:23
18:23 chromatic joined
dalek rrot: r36292 | fperrad++ | trunk/t/codingstd/filenames.t:
[t] localize a variable
 18:27
18:32 cognominal joined
szabgab It would be nice if someone with Rakudo and Test.pm foo could take a look at bug [perl #62992] I just filed 18:36
18:38 Ron joined 18:41 bacek joined, barney joined
rurban I see. I've broken the pbc reader, sorry. 18:41
particle ok, i've got ccache working with msvc, but is there anything i need to configure? environment vars, options, etc? 18:45
rurban take care of LIBS and INCLUDE. they must be defined 18:48
particle i mean like cache dir or size 18:49
rurban not LIBS, it is LIB
dalek tracwiki: v18 | cotto++ | PMCUnionDeprecationTasklist 18:52
tracwiki: trac.parrot.org/parrot/wiki/PMCUni...version=18
shorten dalek's url is at xrl.us/beeon7
18:59 rg joined
rurban All tests pass on gentoo-i868 and cygwin, waiting for solaris 64int... 19:06
dalek rrot: r36293 | cotto++ | trunk/src:
[pmc] switch Integer PMC to ATTRs
 19:07
19:20 timbunce joined
Whiteknight I like having the wiki updates posted here to the channel! 19:20
it's very helpful
dalek tracwiki: v19 | whiteknight++ | PMCUnionDeprecationTasklist 19:23
tracwiki: trac.parrot.org/parrot/wiki/PMCUni...version=19
shorten dalek's url is at xrl.us/beeorn
cotto agreed 19:29
plus it means more karma
Coke, Tcl may need some updating to work with that change to the Integer PMC. 19:30
dalek tracwiki: v4 | whiteknight++ | CallingConventionsTasklist 19:31
tracwiki: trac.parrot.org/parrot/wiki/Callin...?version=4
shorten dalek's url is at xrl.us/beeot7
Infinoid Who owns shorten? It would be nice if it could ignore those.
cotto shorten, owner
Whiteknight shorten: owner 19:32
dalek rrot: r36294 | rurban++ | trunk/src/packfile.c:
[core] Fix the premature alignment fix from r36291.

  - this makes 10 testfiles pass again: nopaste.snit.ch/15482		 19:33
particle metamark.net/bot
rurban moritz: r36294 should have fixed it. 19:35
particle rurban: you have a comment in R36294 about it being broken on 64-bit
still true?
purl If it is a fact, I will believe it!
rurban yes, 2 failures on 64-bit 19:36
19:36 ask_ joined
rurban And a linkflags problem in the Makefile. I just create a new ticket 19:36
Coke purl, shorten? 19:37
purl shorten is, like, a bot interface to metamark.net/ or shortens long urls in the channel or shortens shorter urls in /msgs or made by ask or thefeed.no/marcus/archives/000327.html # osx service or deaf to repetition or metamark.net/bot
19:37 ask- joined
particle why are my cl and lib commands failing? 19:37
rurban how do they fail? 19:41
particle lib cr blib/lib/libparrot.lib src/string/api.obj src/ops/core_o...
LINK : fatal error LNK1181: cannot open input file 'cr' 19:42
19:42 timbunce joined
dalek rrot: r36295 | barney++ | trunk/src/packfile.c:
[codingstd] c_indent.t
 19:42
Infinoid That looks like it's assuming GNU ar.
particle also: 19:43
cl : Command line warning D9035 : option 'o' has been deprecated and will be removed in a future release
cl : Command line warning D9002 : ignoring unknown option '-g'
don't know when these showed up in root.in, checking
dalek rrot: r36296 | barney++ | trunk/src/gc/gc_private.h:
[codingstd] set svn properties
 19:44
Coke www.r-project.org/ 19:49
particle what about it?
i've been thinking about porting to parrot. 19:50
Coke I bet you it will be unbearably slow! =-)
particle not if i can use nci for the libs :)
19:51 cognominal joined
particle goes verbose with init::hints 19:53
dalek rrot: r36297 | barney++ | trunk/src/packfile.c:
[codingstd] c_indent.t
 19:59
rrot: r36298 | barney++ | trunk/src/packfile.c:
[codingstd] trailing_space.t
 20:00
rurban hmm: r36298 is empty 20:01
ah sorry, there are checkboxes at the right :) 20:02
barney it's full of stars ^^^^paces 20:04
rurban q: I want to add an argument to Parrot_readbc(interp, *fullname) to pass options down to pdump, which should not harm parrot reading pbc.
I want to attach a debug option to the pf struct.
Coke changing the packfile probably need to get bounced off the list. 20:05
rurban packfile won't be changed, just the signature.
particle hrmm, that func should probably be renamed Parrot_pbc_read
chromatic agreed 20:06
rurban It's hard to debug packfile.c without harming parrot. I want some private options for pdump, pbc_disassemble maybe also
Is embed.c set into stone yet?
I can paste my patch if you want to :) 20:07
nopaste "rurban" at 212.183.55.181 pasted "idea why Parrot_readbc needs another argument" (825 lines) at nopaste.snit.ch/15485 20:09
20:12 geof joined
Whiteknight particle: I would be interested in porting R to parrot 20:13
if you're looking for teammembers
chromatic No, don't take Andrew away from his other tasks!
Whiteknight particle: on second thought, I'll stay on the sidelines and offer my moral support :) 20:14
I really do need to get my nose back down to the grindstone, don't I? 20:15
particle YES. :P 20:16
20:16 register joined
Coke particle: hey, btw, can you close out that 3 year old ticket? thanks. =-) 20:16
particle Whiteknight said he's got it 20:17
:P
20:17 kj joined
Coke ah. Based on what I saw on the ticket, it involved you saying "yes, this looks fine." 20:17
Whiteknight which ticket is this now? (I have too many to keep track of)
Coke there wasn't a test case that failed or anything, just something that you said was dodgy.
20:20 NotFound joined
Coke Whiteknight: 44393 20:20
Whiteknight Coke: ok. I am going to take care of that ASAP 20:21
Coke awesome. I will have to find some other old ticket to bug particle about. =-)
rurban now we just need a trac ticket hook to irc 20:32
dalek rrot: r36299 | particle++ | trunk/config/init/hints/mswin32.pm:
[config] modify compiler-detection logic on win32, in order to enable ccache compatibility
 20:35
register the opcode reference at www.parrotcode.org/docs/ops/ is for pir or pasm opcodes?
Coke opcodes are both.
kj register: PASM or PIR opcodes is the same
ops can be used both in PIR and in PASM
only, in PIR you can use fancy stuff, like $I0 = length $P0 20:36
and in PASM, you'll have to write length I0, P0
Coke (those docs are auto-gen'd from the POD in src/ops/*.ops, btw.
particle well, in pir, you don't need to specify the full opcode name
kj particle: nor in PASM
you don't write print_sc "hi"
particle yeah, i guess that's true
kj that would be horrible :-S 20:37
20:37 jan joined
kj would increase compilation speed though :-) 20:37
chromatic barely
kj it's a wild guess. maybe not..
chromatic: you tried? :-)
chromatic I doubt we could measure it.
We'd probably want error checking in there as well, to prevent generating invalid bytecode. 20:38
kj well, if you consider that for each op the full signatured opname is calculated, that might take a while.
chromatic A few instructions, perhaps.
kj I don't think so. For each operand there's a _x thingy that needs to be written in a buffer 20:39
chromatic I'm sure we could make that more efficient. 20:40
kj any suggestions?
purl any suggestions are welcome. (including ripping it out entirely :))
20:40 register_ joined
register_ so I reformulate the question... 20:41
the opcodes there are expressed in pasm notation right?
kj ripping it wouldn't be a good idea :-)
particle no
kj register: if your file has .pasm extension, use PASM notation; and so for .pir/pir notation
particle the documentation doesn't specify the syntax to use, just the signature of the op 20:42
kj mm. I'm trying to use 2 different grammars in my little project, but that doesn't seem to work...
register_ i am still here...
basic Infinoid: okay, i think we've got the email2trac stuff working 20:43
particle you'll never call setattribute($P1, $P2, $S3, $P4)
Coke the base syntax is the same in both: <opcode> <comma separated list of registers or constants>
register_ ok
kj anybody experience in using 2 grammars in 1 program? 20:45
Infinoid basic: Ok, I've just sent another test message :)
Coke kj: yes. that's why tcl's 2 grammars are merged into one grammar.
kj Coke: so it's not possible to have 2 parsers?
that's a pity
Coke this is year old intel, and I was trying to share things between the grammars.
I think that, in general, it should work.
basic delivered to command: /usr/local/bin/run_email2trac --project=parrot 20:46
kj I create and register 2 different compilers... the grammars are completely disjoint
rurban particle: my msvc6 compilation still works fine for me. So you didn't break it. Good
Coke basic: will this take emails without tickets and turn them into new tickets?
Coke digs up the email2trac dos.
*docs
particle sweet. need to test icc, mingw, and bcc to be sure 20:47
basic i believe it does
Infinoid I've been testing by replying to an email sent about a TT #49 comment
particle rurban: i'm getting a unresolved external with dynoplibs, did you modify those flags?
dan_ops_switch.dll : fatal error LNK1120: 1 unresolved externals 20:48
myops_ops_switch.obj : error LNK2019: unresolved external symbol __imp__Parrot_cx_schedule_timer referenced in function _switch_myops
myops_ops_switch.obj : error LNK2019: unresolved external symbol __imp__Parrot_cx_handle_tasks referenced in function _switch_myops
Infinoid It kinda looks like my message got dropped on the floor within email2trac, though. No new ticket, no comment posted to TT #49.
rurban I have a patch pending which adds linkflags to those
particle ok, it'd be nice to compile parrot today :)
rurban nope, that's a make realclean thingy
Infinoid basic: I don't suppose it has a logfile? :) 20:49
rurban do you have an old libparrot.dll somewhere in your path? looks like so 20:50
particle i'm realcleaning now
rurban mingw also compiled fine 20:52
kj Coke: (2 different parsers) think it works, but it's just the invocation that I'm not sure about... 20:54
Whiteknight kj, I suspect you would have to create two different compiler objects and call .'compile'() on them both 20:56
kj what I want to do is to run 1 script, and then go into interactive mode to wait for commands
the script is in different language than the commands
dalek rrot: r36300 | rurban++ | trunk/config/gen/makefiles:
Fix TT #262, $linkflags missing for libnci_test.so and dynpmc

  --ccflags='-m64 -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_TS_ERRNO' --linkflags='-m64'
make
  - add @linkflags@ to Makefile, dynpmc.pl and dynoplibs.pl
  (tested on cygwin, mingw, msvc6, gentoo-i386, solaris-64int)		 20:57
particle kj: yes, so you run the interactive commands through the second compiler object 20:58
kj particle: exactly
so now i invoke 'interactive' for the commands stuff
instead 'command_line'
particle languages/foo/foo.pir should create two compiler objects 21:00
kj what do you mean?
particle the action for the command that starts interactive mode should invoke the second compiler
kj yes 21:01
so I pass a filename, which must be compiled by the first compiler
then, I'd like to get the prompt
so I can type commands
and I get to that
however, as soon as I type anyhing and press enter...
ResizableStringArray: Can't pop from an empty array! 21:02
particle it's very hard to debug from this far away :(
what if you use target=parse for the interactive mode?
kj mm. I can only pass --target=parse to the first compiler (when invoking the program) 21:03
mmm, maybe pass that as :named to interactive..
particle right.
kj that doesnt' help # $P1 = $P2.'interactive'( "parse" :named('target')) 21:05
i suspect it has something to do with the arguments that are passed to the first compiler
the .param pmc args is passed to that one, which shifts from that
but the second compiler doesn't even use this array.. 21:06
21:06 cognominal joined
kj ok, disabling the first compiler doesn't help. I'm doingsomething very wrong.. 21:09
particle so $P2 is your second compiler? where you create it, pass 'target'=>'parse' there
seems i'll miss both #parrotsketch and 6perl this week due to interviews 21:12
kj passing "target" when creating the compiler doesn't help 21:14
I mean, no parse tree is printed (just tested with 1st compiler, passing "target"=>"parse" to "new" method)
rurban BTW: languages/lisp has syntax errors 21:15
Shit! t/native_pbc/numbers.t now all fail on i386
moritz like it did a few hours before 21:16
rurban yes, but I thoght I fixed with the ALIGN_16 stuff. So it's the other change... 21:17
Whiteknight rurban++ # it's the thought that counts 21:18
particle kj: i think it's the 'command_line' method that parses it
i may be wrong, it's been a while...
rurban: i'm failing t/pmc/os, t/native_pbc/number, and t/src/compiler here
kj in "interactive" method, there's also a check for "target"
particle ah 21:19
rurban t/native_pbc/number is expected.
t/pmc/os probably for hardlink
I never saw t/src/compiler.t failing on MSVC, but I only test msvc6, and you have probably msvc8 or 9
particle OS.readdir is failing in os.t 21:20
# compiler_5.obj : error LNK2001: unresolved external symbol _PMCNULL 21:21
perhaps because of
# Failed to build 't\\src\\compiler_5.exe': LINK : warning LNK4044: unrecognized option '/Lblib\\lib'; ignored
kj oooooh I *think* i got it.... (trying) 21:22
rurban PMCNULL sounds familiar, a missing export setting for the dll 21:23
This error is described somewhere... 21:24
kj ahum. it helps if you actually .include the generated grammar... :-S
dalek rrot: r36301 | rurban++ | trunk/t/native_pbc/number.t:
skip t/native_pbc/number.t: giving up for today
 21:26
rrot: r36302 | fperrad++ | trunk/languages/lua/t/harness:
[Lua] smoke

See reports on smolder.plusthree.com/app/public_pr...reports/12		 22:04
shorten dalek's url is at xrl.us/beepku
22:05 rob joined
basic Infinoid: yep, found the log file, i think I got it working.. see trac.parrot.org/parrot/ticket/263 22:10
Infinoid Very nice. 22:14
The other common usage we'd like is to generate comments to existing tickets by replying to the ticket emails. Unfortunately, that generates an error message. 22:15
<tickets@parrotvm.osuosl.org>: Command died with status 251: "/usr/local/bin/run_email2trac --project=parrot"
basic Infinoid: did that happen just now? 22:20
Infinoid Yeah, 7 minutes ago.
I wonder if maybe it's concerned about the subject tagging done by mailman. By the time it receives a reply, the subject looks like "Re: [parrot-tickets] [Parrot] Some Ticket Title" 22:22
s/concerned/confused/
basic hmm
basic checks the docs
kj well, so much for my text adventure engine.. can't get 2 parser in the same program working :-(
basic Infinoid: try this for a subject line "[parrot] #263: Re: Test123" 22:24
22:28 Whiteknight joined
GeJ Good morning everyone 22:28
Infinoid basic: Got an error message from that. I also tried "[Parrot] #263: Test123" and "Re: [parrot-tickets] [Parrot] #263: Test123", errors there too. 22:31
(mailman is rewriting the Reply-to: header now, and I forgot to change it in one of my tests, so I also posted one test message to the list instead of trac.) 22:32
basic Infinoid: okay, thanks for looking into that, i'll troubleshoot some more on our end 22:34
it's very very close to working though :)
Infinoid Yep! And it is working well enough for us to point our parrotbug script to the right place. basic++ 22:35
I can link you to some message text if you want to run email2trac directly from the command line.
basic sure! 22:36
nopaste "Infinoid" at 96.238.213.50 pasted "email text for basic" (44 lines) at nopaste.snit.ch/15486 22:37
Infinoid That nopaste link should be a good starting point.
We're probably going to have to adjust our mailman config for both the reply-to header rewriting and for the tagging of the subject line. 22:38
dalek rrot: r36303 | fperrad++ | trunk/t/codingstd/trailing_space.t:
[codingstd] add Makefile source files
 22:40
23:24 bacek_ joined
dalek rrot: r36304 | allison++ | branches/pdd28str_part2:
[pdd28str] Renaming bitwise and boolean string functions.
 23:25
Whiteknight I'm getting a packfile-related failure in t/pmc/eval.t on x86_64 23:31
maybe more, tests are still running
dalek rrot: r36305 | NotFound++ | trunk/src/pmc/parrotinterpreter.pmc:
add experimental method 'stdhandle' to the Interpreter pmc
 23:36
23:37 kid51 joined 23:41 rurban_ joined
kid51 reviews today's posts and sees Whiteknight has got the cage cleaning bug :-) ! 23:51
23:57 jhorwitz 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!