Parrot 1.0 Released | parrot.org | 385 RTs left!
Set by moderator on 26 March 2009.
00:09 AndyA joined 01:05 crythias joined 01:13 msmatsko joined 01:22 pinkyandthebrain left
dalek rrot: r37752 | pmichaud++ | trunk/compilers/pge/PGE/P5Regex.pir:
[pge]: Update P5Regex so that its parser recognizes specific stop tokens.
 01:55
02:06 cotto joined 02:09 Theory joined
dalek kudo: 353aca7 | pmichaud++ | src/parser/ (3 files):
Enable :P5 on m/.../ and rx/.../ patterns.
 02:10
shorten dalek's url is at xrl.us/bemrgc
davidfetter HAI 02:23
where do i find refs on the numbering conventions? 02:24
for parrot releases
rg there is trac.parrot.org/parrot/wiki/ParrotRoadmap and www.parrot.org/news/vision-for-1_0 probably still applies. what are you looking for? 02:31
davidfetter rg, that latter one helps a bunch. thanks :) 02:32
chromatic The deprecation points are 1.4, 2.0, 2.6, 3.0, 3.6, etc. 02:42
davidfetter i was thinking more in terms of bug fix releases
<-- playing with some of the packaging bits 02:43
chromatic There may not be bugfix releases. 02:44
davidfetter "Monthly releases will not be packaged, though critical bug and security fixes will be maintained in updated packages (1.0.1, 1.0.2, etc)" 02:49
is that no longer true? www.parrot.org/news/vision-for-1_0
rg that is correct, but we hope there won't be any critical or security bugs 02:51
03:01 Theory joined
dalek kudo: 7c69c7b | pmichaud++ | (2 files):
Add some more t/spec/S05-modifier files to spectest.data .
 03:13
shorten dalek's url is at xrl.us/bemrk9
03:25 tetragon joined 03:42 janus joined
dalek kudo: 4929856 | pmichaud++ | t/spectest.data:
Add unspace.t to spectest.data .
 03:48
shorten dalek's url is at xrl.us/bemrpp
dalek rrot: r37753 | allison++ | trunk/DEPRECATED.pod:
[cage] Add a deprecation notice for paths prefixed with 'library/' or
 03:50
04:36 TiMBuS joined 05:05 Chile joined 05:07 Khisanth joined 05:13 Woody4286 joined
cotto NotFound, ping 05:47
NotFound, nm 05:57
05:59 masak joined
Chile are there any compilers for parrot right now that would be considered stable enough to be useful for (non-critical) development? 06:12
the site and lists seem to be on the cautious end about stability & usability 06:13
cotto Chile, Rakudo is probably your best bet. It's in very active development, but once something works it should stay working. 06:14
Chile good to know. is there an ETA for dust starting to settle, or is it just finished when its finished? 06:15
cotto You mean Christmas? 06:16
Chile guess so? 06:17
cotto I don't know when a Rakudo release will pass all of Perl6's spectest. pmichaud would be a good person to ask that.
Christmas?
purl Christmas is all jolly goofy shit. weee! or o/~ we wish you a spendy christmas o/~ x3; o/~ and a 'spensive new year o/~ or not yet or next year
cotto Christmas is also when Perl6 is released. 06:18
purl okay, cotto.
Chile nice.
I'm less concerned with perl6 specifically and just Some Language Targeting Parrot, but Perl 6 is as good as any
cotto What's your interest in a Parrot-hosted language? 06:19
research project or something similar? 06:20
Chile I'm making a choice of integrating a scripting language into a project.. parrot would be nice, since it'd be flexible for end-users in terms of language pref, but I don't want to be fighting the implementation of my scripting language while I'm writing code for the project 06:21
NotFound cotto: pong 06:36
07:07 uniejo joined
cotto NotFound, I was just concerned about the commit where you added an #include for dlfcn.h instead of adding it to config/gen/config_h/feature_h.in . 07:21
s/adding it to/adding it in/ 07:22
Would something like this work?: 07:26
nopaste "cotto" at 96.26.202.243 pasted "alternate solution to dlfcn header include" (31 lines) at nopaste.snit.ch/15971
NotFound cotto: I think is best that includes are located near the point of usage 07:34
If it were to avoid the #ifdef HAS_HEADER, it may be useful to do it another way. But the file is alreadey full of that #ifdef 07:36
cotto NotFound, I disagree, but not very strongly. If your commit fixes a problem you saw, I'm content to leave it like it is. 07:38
NotFound And one more point: I prefer #includes located at the .c files that need it, rather than at .h files 07:39
I dindn't saw the problem, It was commented here yesterday that it was failing on... some bsd, don't remember well 07:40
Maybe we must have opened a ticket about that 07:41
Must go to dayjob. SYL 07:42
07:46 riffraff joined
cotto bye 07:47
opening a ticket (or starting a thread) sounds like a good idea
07:51 Santtu joined 08:03 bacek joined 08:17 amoc joined 08:26 alvar joined 08:41 isop joined 09:33 dduncan joined 09:40 contingencyplan joined 09:52 bacek joined 10:10 mikehh joined 10:17 skv joined 10:19 dduncan left 10:36 uniejo_ joined 10:45 alvar joined 11:17 jrockway joined 11:42 rob joined 11:57 ruoso joined 12:24 mikehh joined 12:35 rg joined 12:54 bacek joined, korshak joined
korshak hi! has anyone built parrot using mingw? 12:58
under win32, yes
jonathan hi 13:00
I believe it to work.
korshak i wish so.. 13:01
jonathan Or at least, I broke the build for that a few days back, and somebody complained, and I patched it and it work again...so I think someone has been building it successfully of late.
korshak i got parrot-1.0.0, mingw gcc version 3.4.5 (mingw-vista special r3), but it dies with message like this nopaste.org/p/apWaVFII7 13:04
13:06 gryphon joined
jonathan That's an...interesting way to fail... 13:07
jonathan hasn't seen that one before
korshak hmm it looks like wrong LD detected
jonathan Configure.pl should have generated a my_config or similar file that may give some hints on that 13:08
korshak maybe problem in me environment
Linker and Libraries: ld='g++' <-- it's wrong
simontwo and it's not the backslashes causing command-line havoc? 13:10
korshak but if i try Configure.pl --link=ld it fails "Compilation fails with 'gcc'" 13:11
simontwo: did not understand you, but actually no 13:13
Coke did you specify any options on the command line to configure.pl ? 13:15
korshak first time i just typed Configure.pl 13:18
13:18 bkuhn joined
korshak it wrote in makefile: LD := g++ 13:19
and in 'myconfig': ld='g++', ldflags=''
then i try "Configure.pl --link=ld": i dies on inter::progs: inter::progs - Determine what C compiler and linker to use...Compilation failed with 'gcc' 13:20
Coke which make are you using to build?
korshak mingw32-make, as Configure.pl told me
Coke ok.
is gcc in you rpath? 13:21
(and/or g++) ?
korshak yes
Coke both?
purl i heard both was country *and* western or salt and MSG
korshak there's no g++
Coke checks his strawberry perl to see what it does. moment.
korshak there is only ld in MinGW
13:22 donaldh joined
Coke I think there's two different things you'd need to override there. 13:23
(man is svn slow thru this proxy)
korshak =)
13:24 mj41 joined
korshak i edited makefiles LD and set it to 'ld' 13:24
but i dont know correct LDFLAGS for ld
lol 13:25
i just need to install g++ in MinGW
Coke it should be possible to build without g++. 13:26
korshak sorry, guys, false alarm
hmm...
Coke ... but I don't know how. =-)
korshak at this moment - no :)
Coke looks like strawberry perl (my mingw32 distro) comes with g++
korshak i use activeperl 13:27
under win32
Coke huh. ISTR when I used to use that, it preferred nmake.
that was many versions ago, though.
sorry. 13:28
if you can install g++, that's probably the quickest way out atm. =-)
korshak np, i downloading it now =)
Coke on the plus side, this has inspired me to actually try to build on windows for the first time in months. 13:29
korshak :) 13:30
third try..
hmm, looks good =) 13:31
Coke yay. 13:32
Coke tries a smoke run from this box.
13:34 NordQ joined
dalek kudo: 9fa0fca | pmichaud++ | docs/spectest-progress.csv:
spectest-progress.csv update: 339 files, 7795 passing, 0 failing
 13:51
shorten dalek's url is at xrl.us/bemsox
13:52 korshak left 13:56 korshak joined 14:07 korshak left, korshak joined
Infinoid if I want to call a pmc method from a vtable in the same PMC, what's the best way? I dunno the calling conventions that well. 14:16
PDD03 mentions a call_methodcc(), but that doesn't seem to actually exist.
14:19 Theory joined 14:20 Andy joined
pmichaud Infinoid: perhaps PCCINVOKE ? 14:25
see src/pmc/complex.pmc:1423, in which "sin" and "cos" are called from within "tan" 14:26
14:26 korshak left
jonathan Well ain't *that* going to perform well! 14:26
Infinoid Yeah, that's what I was wondering. But at least it should allow subclasses to override the method 14:27
thanks!
(should PDD03 be updated?) 14:28
pmichaud Yes. But so should the calling conventions generally. :-)
14:39 particle joined 14:40 uniejo joined 15:13 rdice joined 15:20 donaldh joined 15:32 davidfetter joined 15:34 alvar joined 15:53 davidfetter joined, Tene joined 15:54 darbelo joined
Coke parrot_config --perl = "perl" - seems like this should be a full path. 16:07
particle is that on mac? 16:08
C:\\dev\\parrot\\trunk>parrot_config perl 16:09
C:\\usr\\perl-5.10.0\\bin\\perl.exe
Coke on the one I installed for testing. 16:14
wonder if it's just a $^X issue.
16:15 contingencyplan joined 16:29 mattp_ joined
dalek rrot: r37754 | allison++ | trunk/docs/project/debian_packaging_guide.pod:
[debian] A few updates to the packaging guide while building packages.
 17:00
rrot: r37755 | allison++ | trunk/docs/book/ch09_pct.pod:
[book] A few cleanups while reading.
 17:03
17:21 korshak joined 17:24 cognominal joined 17:30 korshak left
Tene [sweeks@kweh ~]$ parrot_config perl 17:38
/usr/bin/perl
purl /usr/bin/perl is canonical
particle no, ubuntu is canonical 17:39
purl okay, particle.
particle urk.
Coke <nelson/> 17:41
tewk www.parrotcode.org/docs/roadmap.html is blank and should auto redirect to trac.parrot.org/parrot/wiki/ParrotRoadmap 17:49
or trac.parrot.org/parrot/wiki/Parrot...ermRoadmap 17:50
Coke tewk: fixed. 17:54
tewk Coke++ 17:55
Infinoid adds the same redirect to parrot.org (trying to keep them in sync) 17:57
18:01 barney joined, korshak joined
Tene I should have talked tewk into coming to my Parrot presentation I gave at the LUG at his university. 18:05
Infinoid if I want to volunteer for roadmap entries, I just stick my name on the wiki, right? or is there someone I should inform? 18:18
Coke 1st, create a ticket if one doesn't already exist, then claim it and update the roadmap. 18:22
(we want to drive this by ticket eventually)
Infinoid cool. 18:23
Infinoid likes the embedded nanoparrot but supposes it's still a bit early for that 18:24
korshak as i understood, Parrot Compiler Tools did not comes with win32 package?
Infinoid The parrot-win32 package should contain PCT 18:25
korshak hmm
at least gen_class.pl is not there 18:26
dalek rrot: r37756 | allison++ | trunk/tools/dev/mk_native_pbc:
[cage] Use more portable comparison operator. Thanks to Havard Eidnes
korshak actually, theres only 7 *.pl files provided by setup-parrot-1.0.0.exe 18:27
or i miss something?
particle what are you trying to do? 18:28
korshak hard question =) 18:30
i try to follow "Create a template PMC file" chapter from Parrot-1.0.0.chm
particle excellent. however, i haven't read it recently :( 18:31
rg is there maybe a win32-dev package? like you need to make install-dev when you want the development tools installed
particle yeah, i'm wondering about that. fperrad isn't around. 18:32
he's the win32 packager
Infinoid there are a bunch of language addons and a docs package but I don't see a dev one. sourceforge.net/project/showfiles.p..._id=211372
shorten Infinoid's url is at xrl.us/bemttx
korshak well, i can build parrot from source, shure 18:34
at least, i can try =)
18:37 TonyC joined
Tene korshak: what are you wanting to do such that you're trying to make custom PMCs? 18:43
korshak i wont to explore parrot and looking where to start 18:46
s/wont/want/
18:49 eternaleye joined
mikehh Coke: you around 18:51
make fulltest fails distro_tests - (the last test) test - t/distro/test_file_coverage.t - no test file for socket 18:58
cotto seen wknight8111 19:00
purl wknight8111 was last seen on #parrot 9 days, 23 hours, 47 minutes and 26 seconds ago, saying: is too much of a wuss to quit his job [Mar 17 19:11:41 2009]
Tene korshak: PMCs are fairly low-level, and IMO not a very good starting point for someone new to Parrot. if they're interesting to you, though, don't let me stop you! :)
mikehh I presume that means that there is no corresponding test for src/socket.pmc 19:01
korshak no, i already found parrot/examples in source
Tene, thanx =)
actually im interesting in tools, which can help me to compile parse and understand xml, in future - xsl and xpath 19:02
bacek good morning 19:03
korshak not a real project, just a thing to play with
dalek rrot: r37757 | pmichaud++ | branches/pge1:
New branch for numerous PGE deprecations and refactors.
korshak bacek, good night =)
mikehh good morning bacek - actually tending towards evening for me 19:04
cotto mikehh, yes. It means there's no t/pmc/socket.t, which presumably would test the new Socket PMC.
bacek wonders why he woke up so early...
cotto: my bad. I didn't added it. 19:05
Is parrot-dev available over NNTP somewhere? 19:06
cotto I can apply it if you have a patch.
bacek cotto: ok. Give me few minutes. I've got a few changes for sockets.
cotto no rush 19:07
nopaste "bacek" at 123.243.38.218 pasted "t/pmc/socket.t patch" (45 lines) at nopaste.snit.ch/15972 19:09
bacek oops.
I forgot to change one line :) 19:10
nopaste "bacek" at 123.243.38.218 pasted "t/pmc/socket.t patch again" (45 lines) at nopaste.snit.ch/15973 19:11
bacek btw, in current pdd22 no special network ops anymore. But I think that keeping "socket" and "sockaddr" ops better than remove them. 19:12
Tene why? 19:13
pmichaud I think "socket" should be a method on a Socket PMC
"sockaddr" can also be a method on a Socket PMC, for that matter.
Coke mikehh: here now. 19:14
please open a trac ticket for that.
unless cotto has already done it.
bacek pmichaud: "socket" creates new Socket.
Coke catches upa nd sees there's a patch. whee
pmichaud bacek: it doesn't have to.
Coke socket should just be new 'Socket'. :P 19:15
pmichaud any more than "open" doesn't have to create a FileHandle
cotto bacek, patch applied (plus manifest update)
bacek pmichaud: but "open" actually create "FileHandle" AFAIU
Coke (me is merely commenting ont he recently added socket opcode.)
(not the desired set of opcodes/methods)
pmichaud bacek: okay, you're correct that "open" opcode creates a FileHandle 19:16
dalek rrot: r37758 | cotto++ | trunk (2 files):
[PMC] add a stub test for the Socket PMC (bacek++ for the patch)
 19:17
pmichaud so in that sense I'd say it's okay for "socket" opcode to create a Socket PMC. But see Allison's mailing list comments about that.
bacek pmichaud: url?
mikehh coke: hi, I was wondering if there was any progress re the post config test failures TT#489
cotto Coke, I haven't done anything related to sockets other than apply bacek's patch just now.
Coke mikehh: no. 19:18
any progress would be on the ticket.
bacek pmichaud: found. About dynops?
pmichaud bacek: lists.parrot.org/pipermail/parrot-d...01898.html
shorten pmichaud's url is at xrl.us/bemt24
pmichaud still, the point is that since FileHandles have an 'open' method, it would be possible for Socket PMCs to have a 'socket' method 19:19
at which point the opcodes aren't really necessary.
bacek pmichaud: ok. 19:20
pmichaud however, IAN The Parrot Architect.
korshak one more question: mk_language_shell.pl should be run only from source dir or it possible run it from dir, where parrot installed by "make install-dev"?
Tene korshak: you can run it from anywhere 19:22
should be able to
if you can't, i think it's a bug.
korshak make install or make install-dev does not installs parrot/tools 19:23
pmichaud (I plan to write an alternate version of mk_language_shell.pl based on Rakudo's build environment.)
korshak it is in dir with source 19:24
but make install doesnt bring it to c:\\parrot (for example)
btw, how to change installation dir? =)
pmichaud korshak: --prefix= to Configure.pl, I think.
korshak oh! thx 19:25
Tene korshak: it *should* be part of 'make install-dev'
but I don't think it is right now
korshak ok, understood 19:26
19:29 cognominal_ joined
dalek rrot: r37759 | pmichaud++ | branches/pge1/compilers/pge/PGE/Match.pir:
[pge]: Small rename of 'result_object' to 'ast' (per S05).
 19:30
Coke renaming things makes me think of deprecation.
pmichaud result_object is already deprecated. 19:31
eligible in 1.1
Coke k. 19:33
it's not just being removed, it's being renamed?
(looking at the listing in DEPRECATED.pod)
pmichaud yes.
dalek rrot: r37760 | pmichaud++ | branches/pge1/compilers/pge/PGE (2 files):
[pge]: PGE::Exp::Literal nodes now _always_ get their text from the ast obj.
 19:34
pmichaud at the time I wrote the DEPRECATION notice I didn't know what it would be named, only that it would be something different.
(there was also the possibility that it would be gone altogether)
Coke Ok. If you could update trac.parrot.org/parrot/wiki/Deprecation with the replacements when you're done, that would be very helpful. 19:35
pmichaud will do.
purl somebody said will do. was there a more appropriate channel for me to ask question regarding trycatch in?
rg forget will do. 19:36
purl rg: I forgot will do
korshak huh. can anybody force purl russian word "ŠæрŠøŠ²ŠµŃ‚"?
rg anyone can tell purl stuff ;) 19:37
dalek rrot: r37761 | pmichaud++ | branches/pge1/compilers/pge/PGE/Match.pir:
[pge]: Small rename of 'text' to 'Str' (per S05).
korshak no! look at that:
ŠæрŠøŠ²ŠµŃ‚
purl, ŠæрŠøŠ²ŠµŃ‚
purl i think ŠæрŠøŠ²ŠµŃ‚ is hello
it has been said that ŠæрŠøŠ²ŠµŃ‚ is hello
rg see ;)
korshak ok. now: 19:38
purl, forget ŠæрŠøŠ²ŠµŃ‚
forget ŠæрŠøŠ²ŠµŃ‚
rg hmm
purl korshak, I didn't have anything matching ļæ½ļæ½ļæ½ļæ½ļæ½ļæ½ļæ½ļæ½ļæ½ļæ½ļæ½ļæ½
korshak ŠæрŠøŠ²ŠµŃ‚
purl it has been said that ŠæрŠøŠ²ŠµŃ‚ is hello
korshak you see?
rg ok, i see 19:39
korshak actually, ŠæрŠøŠ²ŠµŃ‚ really means hallo. but often used as 'hi'
now imagine, that somebody forced purl to say something on each "hi" =)
Coke hi 19:40
purl privet, Coke.
korshak hmm
diakopter hi
rg hi 19:41
diakopter rg: you're not purl
rg thinks purl is partion
*partial
korshak hi has karma of -3 19:42
Infinoid karma hai
purl hai has karma of -1
Infinoid hai++
rg but i guess purl has an encoding bug in forget 19:43
korshak perhabs. better he has bug in matching words 19:44
19:53 korshak left 19:55 korshak joined
korshak hmm. it looks like PCT are designed now to use them only from source dir, not from installation 19:56
tewk Tene: didn't know you were speaking, I would have tried to make it.
20:03 pinkyandthebrain joined
pmichaud PCT shouldn't care, actually. 20:03
I think there are several PCT-based languages running from installation. 20:04
Tene I've seen failures of rakudo being unable to locate PCT from an installation directory.
I'll see if I can reproduce tonight.
korshak pmichaud: ofcourse. because make install does not installs *.pl =) 20:05
pmichaud well sure, rakudo still has difficulty with parrot install
but I think that's more because of the dynops and dynpmcs than because of PCT
particle might need make install-dev 20:06
korshak i did
after install, install dev there's aboud 4 *.pl files in installation dir
Coke wonders if he'll need to create a 'parrot-dev' macport. 20:07
pmichaud Coke: likely.
korshak and even if i put content of tools to instdir/tools manually
pl files can not find their libs
for now i found most usable way to use that *.pl utils from source dir, where parrot was maked 20:08
dalek rrot: r37762 | pmichaud++ | branches/pge1/compilers/pge/PGE (3 files):
[pge]: stringification of Match objects is now .Str instead of .ast (per S05)
 20:17
bacek What is current best practice for sending patches? trac? 20:18
pmichaud bacek: yes.
bacek pmichaud: thanks
dalek rrot: r37763 | pmichaud++ | branches/pge1/compilers/pge/PGE/Match.pir:
[pge]: Numification of Match objects is based on matched text instead of .ast.
 20:20
bacek pmichaud: and if I have big changes send them as big single patch? Or send individually?
pmichaud depends on how inter-linked the changes are. 20:21
in general smaller patches are preferred.
bacek pmichaud: a lot of changes for sockets :)
dalek rrot: r37764 | pmichaud++ | branches/pge1/compilers/pge/PGE/Match.pir:
[pge]: Change $!item to $!ast.
 20:23
bacek trac.parrot.org/parrot/ticket/514 with socket updates. 20:31
I have a question: how properly implement 'new "Socket", some, more, args'? 20:32
pmichaud doesn't exist
bacek or it can be only 'new "Socket", init_hash'?
pmichaud it's just new "Socket", arg
where arg is canonically a hash these days.
bacek ok. thanks 20:33
dalek rrot: r37765 | pmichaud++ | branches/pge1/compilers/pge/PGE (2 files):
[pge]: Separate get/set into separate .ast and !make methods on Match.
 20:37
rrot: r37766 | pmichaud++ | branches/pge1/compilers/pct/src/PCT/Grammar.pir:
[pct]: Change access of (private) $!item to $!ast.
 20:40
20:40 korshak left
particle arg can be any pmc 20:44
just have to write init_pmc to deal with whatever type you want to pass 20:45
Infinoid bacek: hai 20:46
bacek: I've been working on this stuff too, please see groups.google.com/group/parrot-dev/...1ba4b0753? 20:47
shorten Infinoid's url is at xrl.us/bemug8
Infinoid my Socket.init_pmc tries to accept a hash or array, but I haven't tested the array stuff at all yet 20:48
20:49 eternaleye joined
bacek Infinoid: it shouldn't accept Array. 20:50
Infinoid it's easy to rip that out, I don't mind.
bacek Infinoid: "where arg is canonically a hash these days." 20:51
Tene It's not a standard, just the behavior of many other existing PMCs.
20:51 rblasch joined
Infinoid yeah. my thought on the Array was to try to make it look as much as possible like the C socket() function, but I don't care too much and it can be removed. 20:51
Looking at your patch, I think your allocation model fits into the existing system a bit better than mine did 20:56
20:58 Theory joined
particle bacek: unless the pdd says it must be a hash, it doesn't need to be 21:02
bacek particle: pdd doesn't specify socket constructor. 21:05
Infinoid Class and Complex are other PMCs which accept both arrays and hashes for the init_pmc() arg 21:06
particle Exporter probably does, as well 21:07
i can't remember atm
pmichaud most newer PMCs use a hash these days 21:09
Infinoid bacek: It seems I've duplicated some of your effort. Is there some part of this I can work on tonight which won't get in your way?
My goal has been to remove as many socket ops as possible from httpd.pir; they're all gone except for the "socket" and "sockaddr" ops 21:10
pmichaud (and allison specifically mentioned a hash in her email)
bacek Infinoid: see TT#516 and #517
So, if all socket ops gone. 21:11
Infinoid pmichaud: The patch in my email accepts either a hash or an array. But I'm ok with getting rid of the array part
bacek: heh, it seems I duplicated the #516 effort too (and asked the list thismorning to see if that was ok) 21:12
bacek Infinoid: 516 is small :)
pmichaud I'm fine with doing either hash or array as well. As you indicated, there are many places where array is far simpler to handle.
bacek vote for handling Arrays 21:13
# TODO provide sys/socket constants
How this can be implemented? 21:14
Infinoid probably the same way exception type constants are implemented
there's a .pasm file you can include for those
bacek Infinoid: o! thanks.
Infinoid 'except_types.pasm'
bacek: I want to rebase my work on your set of patches, there are a couple of extra fixes I want to add. Is there any reason why I can't start checking this stuff in? 21:17
bacek Infinoid: it's not pushed to github :)
Infinoid I'll just import them into my local stack
bacek Infinoid: no objections :) 21:18
21:20 donaldh joined
bacek ok, time to go. 21:25
afk # kids, family, etc
Infinoid bacek++
bacek back for few seconds. 21:32
Infinoid: don't forget to check result of Socket.socket in init_pmc. You probably have to return PMCNULL if socket failed. 21:33
bacek run away
Infinoid Yeah... haven't readded init_pmc yet 21:40
I adapted Socket.socket to the Parrot_io_socket return value change tho 21:41
dalek rrot: r37767 | pmichaud++ | branches/pge1/compilers/pge/PGE (4 files):
[pge]: Convert obsolete calls to 'result_object' to '!make'.
 21:42
21:43 Whiteknight joined
dalek rrot: r37768 | allison++ | trunk/DEPRECATED.pod:
[cage] Enter deprecation notice for old names of JSON libraries.
 21:46
rrot: r37769 | Infinoid++ | trunk (10 files):
Apply patch from bacek++ in TT #514:

  * Add a Socket.socket() method, use it in the httpd.pir example
  * Add a SockAddr.sockaddr() method, use it in the httpd.pir example
  * Remove a bunch of unused socket ops from io.ops and ops.num
  * Fix up the Socket PMC some, use more inheritance, add POD documentation		 22:09
particle yikes, i might have split that into a few patches 22:10
were the io ops deprecated?
if so, was deprecated.pod updated?
if not... don't remove those ops! 22:11
Infinoid I was under the impression the io ops were added after 1.0.0
I could be wrong, I'll check.
particle ah, you may be right. please check, though :)
Infinoid If I'm wrong, r37771 will have the same issue 22:12
22:12 donaldh left
dalek rrot: r37770 | Infinoid++ | trunk (7 files):
Apply patch from bacek++ in TT #516:
 22:12
Infinoid Yep, the ops were added in r37705, 1.0.0 was r37536 22:15
dalek rrot: r37771 | Infinoid++ | trunk/src/ops (2 files):
Apply patch from bacek++ in TT #517:
 22:16
Infinoid particle: If I understand the deprecation cycle stuff properly, I think we're safe.
particle were they deprecated before 1.0.0?
Infinoid They were *added* post-1.0.0
particle ok, then we're ok
Infinoid great, thanks 22:17
particle Infinoid++ bacek++
Infinoid dalek is going to take a while because it has a long polling interval, but r37774 tests successfully for me on linux and mingw
(using examples/io/httpd.pir to test it, I don't think there's anything in t/ yet)
particle wow, i just wrote a php api that actually looks good. 22:18
Infinoid Call Guinness
I didn't think that was possible :)
dalek rrot: r37772 | Infinoid++ | trunk/src/pmc/socket.pmc:
Fix the return value handling in the Socket PMC.
 22:19
Infinoid particle: Do you know anything about mingw? 22:20
particle yeah, i'd better call it a day
not much, really, just that it's make engine has a really long name
Infinoid I'm getting a warning about setsockopt()'s third argument changing an integer into a pointer without a cast, but MSDN says the function's third argument is an integer, just like on unix 22:21
Not a big deal, it still works, was just wondering if you had any ideas
I'll start grepping header files 22:22
particle not me, i only use mingw when there's a build or test error on strawberry
Infinoid Ok, thanks
particle good luck!
purl You'll need it.
dalek rrot: r37773 | Infinoid++ | trunk/src/io (2 files):
[io] Fix a bug calling setsockopt() with uninitialized data.
rrot: r37774 | Infinoid++ | trunk/src/io/socket_win32.c:
[io] Fix an error and a warning in socket_win32.c on mingw.
 22:26
cotto seen chromatic 22:27
purl chromatic was last seen on #parrot 19 hours, 43 minutes and 15 seconds ago, saying: There may not be bugfix releases.
dalek rrot: r37775 | coke++ | trunk/t/codingstd/c_function_docs.t:
[cage] remove unused variable; invert conditional to avoid code duplication.
 22:29
rrot: r37776 | Infinoid++ | trunk/t/pmc/sockaddr.t:
[t] Fix t/pmc/sockaddr.t to use methods instead of ops.
 22:42
22:44 Whiteknight joined
dalek rrot: r37777 | Infinoid++ | trunk/src/pmc/socket.pmc:
[cage] Fix a codingstd failure in my recent r37772 patch.
 22:46
cotto Whiteknight, have you had any luck converting the *ManagedStruct PMCs to ATTRs?
It looks like some jit code needs to be updated. 22:47
Whiteknight cotto: none at all. I've been busy with other things lately
cotto np
Whiteknight but tonight I'm focusing on Parrot, so I'll play with it
what kinds of JIT stuff need updatin'? 22:48
Tene I've seen a few failure reports lately of Parrot failing to build without --jitcapable=0
cotto Some code (src/jit/i386/jit_defs.c 2409) is putting stuff directly into PMC_data 22:49
dalek rrot: r37778 | whiteknight++ | trunk/src (2 files):
Fix TT #321 using a patch from ronaldws++
cotto I think I can get it if I can figure out how to call VTABLE_set_pointer 22:50
Infinoid Tene: that would be TT #18 again 22:51
Tene: Apparently selinux on fedora has recently gotten stricter.
Tene Infinoid: one of these reports was on a not-updated-in-years Fedora 7.
Infinoid awesome. In that case, I have no clue. 22:52
Tene I wonder why, what with all the information I've given you.
:)
Infinoid We had someone in here a couple days ago who started getting segfaults building PGE, who was able to compile it just fine the day before. Turns out it was segfaulting the first time a JIT function was called, and an selinux package update caused it.
(that was on FC9)
Whiteknight cotto: do you have that list of instances that need to be updated? 22:55
I lost mine
cotto Yes, but I may be able to finish the switch myself. 22:57
Whiteknight okay, cotto++
Now that I think of it, adding new JIT-supported architectures might make for a good GSOC project 23:00
I personally would be very happy if amd64 support were improved
cotto Is that a 3.5 month project?
Whiteknight could be 23:01
The i386 JIT would make a good template to follow
Chile is PIR to PASM compilation done in native code or interpreted code?
Tene Chile: in a normal compile, there is no PIR to PASM step. 23:02
the PIR compiler goes straight to bytecode.
Chile ah
is that step native or interpreted?
Tene imcc (the PIR compiler part of parrot) is written in C.
Chile cool, thanks. 23:03
Tene np 23:05
Chile: there's a project (not yet completed) to replace imcc 23:06
If you're looking into the details of PIR compilation, you should look at pirc
Chile that's the PIRC directory?
alright
cotto it's in compilers/imcc
compilers/pirc is the newer pir compiler which will eventually replace imcc 23:07
Chile I noticed that pirc had stuff in t/ that looked like perl - is this runtime stuff or compiler-generator stuff?
Tene Chile: yes, pirc uses perl 5 to run its tests. 23:08
Chile alright 23:09
cotto Is PMC_DATA_IN_EXT still useful? I have a patch that will break it? 23:20
s/?$/./
dalek rkdown: ddc1372 | (Francois Perrad)++ | src/parser/grammar.pg:
fix EmphUI & StringUI
 23:22
shorten dalek's url is at xrl.us/bemuxw
dalek rkdown: 5124e85 | (Francois Perrad)++ | src/parser/ (2 files):
allow inline in BlockQuote & ListBlock
shorten dalek's url is at xrl.us/bemuxy
dalek rkdown: d92427e | (Francois Perrad)++ | src/parser/grammar.pg:
remove useless capture in Code
shorten dalek's url is at xrl.us/bemux2
cotto of course if Parrot doesn't explode when it's 1, testing won't be very hard.
nope. The build fails. 23:26
23:29 jan joined
Tene purl: seen kj? 23:29
purl kj was last seen on #parrot 26 days, 1 hours, 57 minutes and 55 seconds ago, saying: hello GeJ [Mar 1 21:30:45 2009]
Tene purl: seen kjs?
purl kjs was last seen on #perl 277 days, 7 hours, 55 minutes and 12 seconds ago, saying: yo [Jun 23 15:33:30 2008]
Tene Hmm... wonder where he's been... 23:30
Whiteknight I would like to add a get_integer method to the Eval PMC. Any objections to that? 23:32
I need a way to determine how many subroutines are contained in a compiled code file, and that seems the most logical way to do that 23:33
Infinoid Hmm. That makes sense, until we start wanting to get counts of things other than subs 23:36
Tene Whiteknight: can't it be a method instead?
I find the overriding of get_foo to be rather distasteful.
Whiteknight: perhaps a 'get_subroutine_count' method 23:38
Coke_afk Whiteknight: doesn't 'elements' work/ 23:40
?
Whiteknight Coke: elements does not work, that I am aware of 23:41
23:42 kid51 joined
Whiteknight Tene: Eval already has a get_pmc_keyed_int, so it makes good sense to me that get_integer returns the number of elements in the array like it would for any other array 23:42
so it's not an "override", just adding something that I think should have been there in the first place
Tene Whiteknight: Okay.
That's fine with me, then. 23:43
Whiteknight okay, thanks
Infinoid Funny, I thought "elements" returned the number of elements in the array like it would for any other array
If that's not what elements is for, I guess I've written PackfileDirectory wrongly... 23:44
Whiteknight Infinoid, when I do $I0 = $P0 to get the number of elements in an RPA, that calls get_integer
although elements() would maybe be a good alternative. I don't care what interface we use specifically, so long as I get the information I need
Coke Whiteknight: k 23:45
Infinoid Whiteknight: you're right, the Array base class returns the number of elements for both .get_integer() and .elements() 23:46
(the former wraps around the latter)
Whiteknight Infinoid: From what I have seen, for array classes get_integer and elements both return the same value
(which begs the question why we have a separate elements() VTABLE anyway, if it's always shadowed by get_integer)
Infinoid PackfileDirectory doesn't have a get_integer, only an elements. That's the extent of my knowledge of the pir array interface 23:47
Whiteknight Oh, i didn't realize that PackfileDirectory was an array-like PMC
Infinoid it's an array of segments 23:48
23:48 TiMBuS joined
Infinoid I wrote it according to the PDD13 spec, but I still don't know what the heck I'm doing in PIR land, so maybe I should add an alias for get_integer 23:48
23:49 tetragon joined
Whiteknight I'll add an elements method to Eval first, we can figure out something else to do with get_integer later 23:49
it is done. Fin. 23:55
dalek rrot: r37779 | whiteknight++ | trunk/src/pmc/eval.pmc:
add an elements() VTABLE to Eval PMC, to retrieve the number of Subs that the Eval contains
 23:56

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!