Parrot 1.1.0 Released | parrot.org/ | 332 RTs left
Set by moderator on 21 April 2009.
00:00 Whiteknight joined
cotto it's a bit wordy. I'd call it decnum-dynpmcs, since I don't think there are any other projects that have "dynpmcs". 00:01
your call
wayland_ how about darbelo-and-cottos-groovy-parrot-project-doing-stuff-with-decnumber-and-dynpmcs ? :) 00:03
cotto Yeah. Go with that. 00:04
darbelo I dunno, lacks punch...
cotto although whiteknight should be in there too
and more adjectives 00:05
darbelo darbelo-cotto-and-whiteknights-awe-inducing-project-that-does-amazing-fabulous-stuff-with-decnumber-and-dynpmcs-for-the-parrot-vm 00:06
cotto winner
darbelo I was going to use virtula-machine instead of vm, but it seemed too wordy. 00:07
Whiteknight where am I supposed to be?> 00:08
darbelo, by the way I don't think we have really met before
cotto you're in there
Whiteknight but your proposal was very impressive
cotto oh yeah. we need to do introductions 00:09
wayland_ introduces self. Hi everyone. I'm the channel clown who makes stupid jokes. I'm visiting from #perl6 until parrot and rakudo work as RPMs 00:10
Whiteknight hello, my name is whiteknight, and I'm a codeaholic 00:12
darbelo hello whiteknight. 00:14
wayland_ Did I pass the Turing test? :) 00:15
darbelo wayland_: Well, you forgot that you are a machine, that counts for something. 00:16
rg cotto/whitenight: since you're both here, do you have a minute to revisit tt#551?
darbelo cotto/whitenight: code.google.com/p/decnum-dynpmcs/ 00:18
rg whiteknight of course
darbelo: watch for my mistakes at least ;) 00:19
Whiteknight rg: sure. What do you need? 00:20
cotto darbelo, add christoph_googleacct at mksig dot org 00:21
rg direction about where we want to go with that ticket. either pick a patch and apply it or add a note saying what you want done.
cotto darbelo, I've been meaning to ask, do you see any compelling reason for separate BigInt and BigNum PMCs? 00:22
Alright! Time to vandalize the wiki. Sucker!
darbelo is "The PDD told me to do it!" compelling enough? 00:25
Whiteknight rg: Any solution we find will be only a temporary fix until we un-bad the JIt system 00:26
so I say you pick whichever bandage that sucks the least for the foreseeable future
cotto Whiteknight, that could be a very long-term temporary fix, depending on how much more awesome needs to go into the JIT system. 00:27
darbelo is easy to compel
cotto Infinoid, can you add decnum-dynpmcs to dalek? 00:28
Whiteknight darbelo: +1 from me
Infinoid parrot: r316512 | infinoid++ | trunk/src/jit (14 files):
parrot: [jit] Add 85% more awesoem
rg personally, i'd go for the simple "don't handle" workaround.
Whiteknight darbelo: Can you add wknight8111 at gmail dot com?
00:28 ruoso joined
Infinoid parrot: This patch makes our robot overlords happier, so maybe they'll turn the heat on in our oppressive space prison. 00:29
cotto: Where's it hosted?
darbelo done.
Infinoid: code.google.com/p/decnum-dynpmcs/ 00:30
Whiteknight yay! Thanks darbelo++
Infinoid ah, googlecode. Ok, I've been meaning to genericise the googlecode rss handling anyway, and add a few other projects there. I'll add it to my list.
rg and either leave the ticket open or find a place for "unlikely to fix" bugs.
cotto Infinoid++ (also, iwbni dalek could map my google account name to my irc handle) 00:31
Infinoid "well, why'd you use a different username then?"
cotto because Google required it?
Infinoid "this parking space reserved for Senior VP of Cottoness" 00:32
some of the github people would benefit from karma lookup tables too. I'll keep it in mind 00:33
cotto I was thinking that too
There's some significant karma fragmentation.
Infinoid yeah. and purl doesn't have account aliases like ohloh does
cotto nope 00:34
btw, where's dalek's source live?
feather?
purl i guess feather is feather.perl6.nl/ or a perl6 community development server or run by juerd.
Infinoid I don't know. I'm told some guy named Masque runs it
looks like he runs it directly on one of the magnet irc server machines 00:35
cotto Wait a sec. If you don't know, how do you modify it?
Infinoid oh, dalek! sorry, was still thinking of purl
yeah, that's on feather, it's diakopter's bot
runs in /home/dalek, managed by daemontools.
I think someone put the rss plugins into pugs svn somewhere, but I've been managing them locally and rsyncing them in 00:37
cotto juerd?
purl well, juerd is root or at juerd.nl/ or mailto:juerd@juerd.nl
cotto Does he check that address? 00:38
00:38 bacek_ joined
Infinoid I don't know. looks worth a try 00:38
cotto we'll see 00:45
darbelo, let me know if you have any issues getting the new svn repo into a useful state 00:51
darbelo Okay, I'm going to hunt some food now. So I'll be offline for about an hour or so. 00:52
cotto no rush. 00:53
darbelo I'll start playing with the repo when I get back. Feel free to vandalize it in the mean time :) 00:54
cotto wonders what's happening with tewk and allison 00:55
00:56 tetragon joined 01:22 s1n joined 01:37 AndyA joined
cotto afk 02:01
02:15 Fayland_logger joined 02:17 Ademan joined 02:20 ruoso joined 02:22 darbelo joined 02:26 Fayland_logger joined 02:32 Fayland_logger joined 02:35 Fayland_logger joined 02:42 Fayland_logger joined, janus joined 02:43 darbelo joined
Infinoid pmichaud: ping 02:44
msg pmichaud Should we worry about TT #550? I don't know ICU at all, or how this code is used. Is this something that will be subjected to direct user input? If so, we should probably apply filters to keep parrot from crashing. If it's purely internal, I think we can probably close the ticket as kid51 suggests. 02:49
purl Message for pmichaud stored.
03:14 darbelo left 03:52 tetragon joined, amoc joined 04:18 davidfetter joined 05:01 iblechbot joined 05:08 Andy joined
cotto allison, ping 05:23
msg allison It's a bit of a long shot, but if IBM were willing to let us use the decnumber code under the Artistic License 2.0, would that be enough to allow us to distribute these PMCs (and decnumber+tests) as part of Parrot? 05:36
purl Message for allison stored.
06:12 flh joined 07:25 allison joined
cotto allison, ping 07:56
07:58 gaurav joined 08:25 masak joined 08:44 HG` joined 09:51 riffraff joined
riffraff hi everyone 09:52
purl Howdy, riffraff, you fantastic person you.
riffraff as the release 1.1.0 is out, shouldn't some packages appear under developers or stable downloads? 09:53
ah it's on devel now
ignore me pleae :)
cotto riffraff, I'm one step ahead of you. 09:56
riffraff :) 09:59
10:03 bacek joined
bacek good evening 10:03
no coo^W messages... sigh... 10:04
What is current policy for highly experimental, throw-away branches?
10:14 he joined, wayland76 joined
nopaste "he" at 158.38.152.119 pasted "Another tested combination for PLATFORMS, ref. smolder.plusthree.com/app/public_pr...ils/20420" (12 lines) at nopaste.snit.ch/16392 10:17
shorten nopaste's url is at xrl.us/bep79d
wayland76 Do we have any patch-reviewing types around? Can I remind any of those wonderful people of trac.parrot.org/parrot/ticket/426 ?
cotto That's a big honking patch. 10:20
wayland76 It is rather 10:21
cotto It's going to be tricky to get that reviewed.
wayland76 But a lot of it is moving code from two places to one place
Why is that?
cotto Can it be broken down?
because it'll take a long time to do
10:22 bacek joined
wayland76 Well, let me tell you about the process, jkeenan and I went through, and then we'll see 10:22
First, we moved code into the Parrot::Install module
Then we wrote tests for it
dalek rrot: r38287 | bacek++ | branches/pmc_pct:
Experimental branch for PCT based PMC parser
 10:23
wayland76 Then we refactored it
So, in theory, if all the tests pass, then noting should change, except that we've gotten rid of lots of duplicate code, and have moved some code into libraries, so that it can be more easily shared 10:24
riffraff anyone could check if this works: perl tools/dev/create_language.pl foo && cd foo && perl Configure.pl && make 10:26
10:27 jrockway joined
riffraff on my box I get a referene to parrot-1.0.0 while runnign this from the 1.1.0 distribution 10:27
cotto wayland76, I'll try to assess it.
wayland76 cotto: Basically, if the code doesn't work, then "make install" (and/or "make install-dev") will fail
The only way I can think of that it could be broken down is to move one function at a time from duplicate-code to Parrot::Install 10:28
Tip: Reviewing the "create_directories" function will be easiest, and lines_to_files will be hardest 10:29
riffraff mh maybe related to an old version parrot laying around 10:30
wayland76 tools/dev/install_dev_files.pl is simpler than install_files.pl and may help you to get an overview of what we're trying to achieve codewise 10:31
lines_to_files should maybe have been called "read_manifests_and_create_file_hashes" 10:32
Any more questions, I'm happy to answer; if you put my handle in them, that will alert me to IRC :) 10:39
cotto takes deep breath and dives in
bacek Is 10 minutes of battery life enough for SVN commit? 10:46
cotto for whom?
purl for thee!
bacek for me.
cotto It usually takes me a matter of seconds
bacek here we go.
I'm in train on 3G internet. 10:47
And in Australia...
O! r38288
dalek rrot: r38288 | bacek++ | branches/pmc_pct (29 files):
Initial drop for PMC parsing.
bacek squashed ~30 commits in one :) 10:48
battery dying... See you.
wayland76 Now that's committed! (sorry :) )
o/
cotto I'd kinda prefer a bunch of small commits. 10:49
Sorry. I don't feel like I can give that a competent review, not fully understanding the install process. 10:55
10:56 wayland76 joined
wayland76 Sorry -- killed my xserver by accident. If I've missed any messages for me, let me know 10:56
cotto Sorry. I don't feel like I can give that a competent review, not fully understanding the install process. 10:57
moritz wayland76: that's what IRC logs are for :-)
wayland76 cotto: Is this another one for Allison? Or are there others to ask? 10:58
(Not that I've got anything against Allison or anything, I just don't want to overload her :) )
cotto me neither 10:59
hmmm.
I'm not sure who else you'd ask, though. 11:00
wayland76 That's what I thought :)
Ok, allison it is then. Ping allison? 11:01
cotto It's 0400 in Washington. She's probably asleep.
wayland76 Figures. Thanks. Fortunately I'll be staying up all night (it's 9pm here, and I've just had breakfast), so hopefully we should be on at the same time 11:03
cotto While you're waiting, I'm sure it'd help her if you tried to break it into smaller patches. 11:06
wayland76 Actually, it'd help everyone if I'd read kid51's comments more closely :) 11:18
Anyway, apologies to all. What kid51 wants with this one is road testing, not reviewing. Apologies especially to cotto for wasting his time. 11:19
(although no doubt reviewing would not be objected to)
11:40 bacek joined
bacek hi again 11:41
purl oh, you're back!
cotto bacek, how's that branch going to bootstrap? 11:47
bacek cotto: from current pmcs
cotto I mean if/when it's merged, will it require an installed parrot to build?
bacek cotto: and then we can put "compiled" .c files into repo
similar to yacc/lex 11:48
cotto I see, kinda like we do with imcc currently.
bacek indeed
cotto That could make for some really noisy commits
bacek I can imagine... 11:49
Step 2: resurrect PIR_PCT
Step 4: Profit!
cotto this is getting really meta
bacek, do you think it'd be useful to have a set of PMCs required for bootstrapping and to only have the "precompiled" versions of those? 11:53
cotto realizes that this is still a ways off
bacek cotto: It was actually chromatic++ idea about L1 ops.
Implement very small subsets of ops. 11:54
cotto right; barely enough to implement everything else
bacek All PMCs will be implemented in this "L1" language instead of "pure C"
wayland76 L1? 11:55
purl L1 is the "stop" button on a sun keyboard
wayland76 That would be difficult to write programs with :)
11:56 ruoso joined
cotto no, L1 is a hypothetical language that would be used to implement PMCs and PIR-visible ops so that they could all be easily jitted. 11:56
purl okay, cotto.
cotto L1?
purl L1 is a hypothetical language that would be used to implement PMCs and PIR-visible ops so that they could all be easily jitted.
bacek irclog.perlgeek.de/parrot/2009-04-21#i_1083550
purl: L1 is also irclog.perlgeek.de/parrot/2009-04-21#i_1083550 11:57
purl okay, bacek.
11:57 aff joined
bacek So, half of "Step 1" done. 11:57
cotto I see. 11:58
bacek++ #bringing the future into the present
aff Hi I'm trying to build rakudo but it chokes on parrot (r38276). Known issue? 11:59
error:imcc:syntax error, unexpected VAR, expecting '(' ('sub') in file 'src/builtins/guts.pir' line 1362
bacek unfortunately, it can't parse _all_ real PMCs ...
cotto So you're done with the first 90% and just have to get the other 90% finished? 12:00
bacek cotto: as usual :) 12:02
aff A 'make clean' fixed the issue. Thanks to moritz at #perl6.
cotto bacek, how do you run the tests in your branch? 12:06
bacek prove -v compilers/pmc/t/01-parse.t
cotto: if can add 02-parse-all-pmcs.t it will be very helpful 12:07
(to test parse all src/pmc/*.pmc)
s/if can/if you can/
cotto So, the only thing that's missing is "you". 12:08
iwbn if make test worked too
cotto finds himself strangely drawn to the new branch 12:10
bacek :) 12:11
12:11 aff left 12:13 kid51 joined
cotto Someone who knows win32 needs to fill in os.pmc's readdir method 12:17
.
dalek tpfwiki: Jim Keenan | Parrot 12:18
tpfwiki: www.perlfoundation.org/parrot/index.cgi?parrot
rrot: r38289 | bacek++ | branches/pmc_pct/compilers/pmc (3 files):
Add "need_ext" as valid pmclass pragma.
rrot: r38290 | bacek++ | branches/pmc_pct/compilers/pmc (3 files):
Add "class_init" as valid method.
bacek cotto: POSIX's readdir should works on win32 12:19
Coke_afk msg allison - I gave you commit privs to partcl.
purl Message for allison stored.
dalek tpfwiki: Jim Keenan | RFP: Parrot Needs Better Smoke Reports 12:21
tpfwiki: www.perlfoundation.org/parrot/index...ke_reports
shorten dalek's url is at xrl.us/bfeof
Coke wishes irssi wouldn't let me send if my nick wasn't "Coke". =-)
msg kid51 thanks. deleted the page. 12:25
purl Message for kid51 stored.
bacek heh. pmc_pct branch already useful :)
12:26 particle2 joined
dalek rrot: r38291 | bacek++ | trunk/src/pmc/array.pmc:
[cage] Mark vtable methods as VTABLE in array.pmc
 12:28
12:33 fperrad joined
dalek rrot: r38292 | bacek++ | branches/pmc_pct/src/pmc/array.pmc:
[cage] Mark vtable methods as VTABLE in array.pmc
 12:34
rrot: r38293 | bacek++ | branches/pmc_pct/compilers/pmc (3 files):
Extend c_type with optional 'struct'
bacek (strict parsing)++ 12:37
Coke has a thought - let's make PMCs and OPs not require c style comment wrappers for pod. 12:38
12:38 rg joined
dalek rrot: r38294 | bacek++ | trunk/src/pmc/capture.pmc:
[cage] Mark vtable methods with VTABLE in capture.pmc
 12:40
cotto bacek, should the test also care about dynpmcs? 12:41
bacek cotto: yes, why not? 12:42
pmichaud is VTABLE simply decorative, or is it actually changing something here?
bacek can anyone confirm, that in class.pmc "void subclass" not used?
pmichaud: I'm going to strictly parse PMC. So, VTABLE has own semantic. 12:43
cotto bacek, it's dead code afaict 12:44
make cover doesn't even have information on it, which makes me think that pmc2c silently drops it on the floor. 12:45
bacek cotto: looks like it is.
cotto pmichaud, is there an easy way to sort a resizablestringarray, or do I have to use a rpa for that?
pmichaud cotto: I'm not aware of an easy way to do it. 12:46
cotto rpa ftw
although that's a little surprising
dalek rrot: r38295 | bacek++ | branches/pmc_pct/src/pmc/capture.pmc:
[cage] Mark vtable methods with VTABLE in capture.pmc
 12:50
Coke pmc2c drops a lot on the floor. =-)
bacek cotto: Parrot_quicksort sorts only PMCs 12:51
Coke: yeah. But I'm willing to fix it :) 12:52
Coke bacek: let me find you an RT.
bacek hates svn more and more... It's just incredibly slow.
dalek rrot: r38296 | bacek++ | branches/pmc_pct/src/pmc/class.pmc:
[cage] Remove dead "subclass" function, add more VTABLE to class.pmc
 12:53
Coke bacek: rt.perl.org/rt3/Ticket/Display.html?id=39313 12:54
summary: pmc2c only translates or passes through /part/ of the .pmc; some bits are just dropped. everything that is not explicitly translated should be passed through unchanged. 12:55
bacek Coke: It's taken by you. But I can overtake it : 12:56
Coke go ahead and steal it.
12:57 LylePerl joined
Coke spasiba 12:57
bacek Coke: Š½Šµ Š·Š° чтŠ¾ :)
Coke: but I disagree with "lax" parsing and rely on C compiler.
We can "strictly" parse PMC with PCT. 12:58
Coke bacek: we don't HAVE pct at that poitn.
bacek Coke: we have.
Coke no, pct requires a built parrot.
parrot requires pmcs.
bacek Coke: scrollback :) We can commit "compiled" PMCs in repo.
Coke /can/ 12:59
not /do/
bacek Similar to yacc/lex
Coke and that introduces a level of self-bootstrap that I personally have concerns about.
dalek rrot: r38297 | fperrad++ | trunk/docs/project/release_manager_guide.pod:
[release] update release_manager_guide
bacek Coke: yes, it is. But we can have bare minimum of PMCs for bootstrapping. 13:00
Hooray. Last comment in RT by allison exactly about this approach. 13:01
wayland76 Have the Perl 6 guys indicated their vague intentions about multithreading following that threads-on-top-of-event-loosp idea? 13:02
s/loosp/loops/
bacek imagine Erlang implemented in Parrot 13:03
wayland76 Does erlang do that?
bacek wayland76: they has own "threads". From my (external) point of view they doing almost same 13:04
wayland76 Ok 13:05
cotto bacek, the only PMC to pass so far is boolean 13:11
test incoming shortly
bacek cotto: thanks a lot
And I already has passed till class.pmc
cotto They're slooooooow. 13:12
dalek rrot: r38298 | bacek++ | branches/pmc_pct/compilers/pmc (3 files):
Add few parrot's adverbs to c_argument.
rrot: r38299 | cotto++ | branches/pmc_pct (2 files):
[pmc_pct] add a test to try parsing all pmcs and dynpmcs
cotto but that's always the first stage
bacek "Road in thousand steps" 13:14
cotto actually, it's not doing too bad
it doesn't like the dynpmcs, but all the non-dyn ones up to capture are working 13:15
cotto looks behind him and sees daylight
wayland76 tries that. No results 13:16
cotto It's funny that the release manager's guide always seems to need some kind of update. 13:17
The more we use Parrot to build Parrot, the more pressure we apply to make it faster, 13:19
Coke cotto: I've been trying to apply that sort of pressure for 7 years by running code on parrot. good luck. :P
cotto Coke, yes, but if it's required to build parrot it can't be avoided, 13:20
wait, 7?
13:20 gryphon joined
Coke or so. 13:20
cotto That is not a small number. 13:21
Coke no. so, good luck. =-)
13:22 bacek_ joined
bacek_ and me again. 13:22
Did I missed something?
cotto nope 13:23
wb
Infinoid good morning parroteers
cotto Hi, Infinoid 13:24
bacek_ Infinoid: it's almost tomorrow :)
dalek rrot: r38300 | bacek++ | branches/pmc_pct/compilers/pmc (4 files):
Add slurpy and named adverbs
Infinoid it's 6:30am, I still haven't gotten used to the idea of it being today!
Infinoid accuses bacek_ of premature optimization 13:25
bacek_ lives fast :)
Infinoid :)
bacek_ Failed 51/94 subtests 13:26
cotto++
cotto now you know what to fix 13:27
bacek_ I used "for src/pmc/*pmc; do ..." before :) 13:28
cotto although I suspect that that number will get small pretty quickly
bacek++
I'm curious what L1 will look like. 13:29
I personally vote for LOLCODE. 13:30
bacek_ +1 :) 13:31
Coke we should consider nqp.
13:32 iblechbot joined
Coke (parrot should have had scheme on parrot or something like it from early on.) 13:32
bacek_ is pheme really dead?
Coke pheme was a late addition, and is a HLL, not a core component. 13:33
and might be dead. Iunno.
bacek_ I used to hate lisp... Until implementing XQuery processor. 13:34
After few days of battle I gave up and start representing internals as lisp.
Coke pheme is chromatic's, he'll know better. 13:36
wayland76 purl, tell allison weren't we going to rename ports/fedora/parrot.spec to ports/fedora/parrot.spec.fedora ? 13:42
purl i haven't a clue, wayland76
wayland76 dalek, tell allison weren't we going to rename ports/fedora/parrot.spec to ports/fedora/parrot.spec.fedora ?
Is that how I do it?
moritz wayland76: s/tell/msg/ to purl
wayland76 purl, msg allison weren't we going to rename ports/fedora/parrot.spec to ports/fedora/parrot.spec.fedora ?
purl Message for allison stored.
wayland76 Sorry. I'm used to lambdabot :) 13:43
dalek rrot: r38301 | bacek++ | branches/pmc_pct/compilers/pmc (3 files):
Add abstract and no_init traits
 13:47
rrot: r38302 | bacek++ | branches/pmc_pct/src/pmc/default.pmc:
[cage] Remove dead elements_keyed_int and set_bool_keyed_int, add more VTABLE in default.pmc
rrot: r38303 | bacek++ | branches/pmc_pct/compilers/pmc (3 files):
Add singleton traits
bacek cotto: 44/94 13:53
purl 0.468085106382979
bacek purl: stupid girl...
purl bacek: what?
wayland76 Didn't you want to know that 44/94 is 46%? 13:54
purlion test :) 13:55
cotto bacek, nice. 13:56
13:57 Theory joined
bacek looks like dynpmc/md2 is dead... 13:59
13:59 Andy joined
cotto bacek, I'm adding VTABLE to dynlexpad. Should I fix the other dynpmcs too? 14:03
bacek cotto: I already did it.
But I can stop checking dynpmcs. 14:04
So we can share workload :)
cotto ok. I'll fix any that look broken. 14:05
bacek You can start from end of the list.
bacek git svn dcommitting now.
*incoming* :) 14:06
cotto man. pmc2c is way too lax about VTABLE function declarations. 14:07
bacek cotto: indeed.
and METHODs as well 14:08
dalek rrot: r38304 | bacek++ | branches/pmc_pct/compilers/pmc (3 files):
Add dynpmc traits
rrot: r38305 | bacek++ | branches/pmc_pct/src/dynpmc/dynlexpad.pmc:
[cage] Mark vtable methods with VTABLE in dynlexpad.pmc
rrot: r38306 | bacek++ | branches/pmc_pct/src/dynpmc/gdbmhash.pmc:
[cage] Mark vtable methods with VTABLE in gdbmhash.pmc
rrot: r38307 | bacek++ | branches/pmc_pct/config/gen/crypto/digest_pmc.in:
[cage] Fix md.? template.
rrot: r38308 | cotto++ | branches/pmc_pct/src/dynpmc/gdbmhash.pmc:
[pmc] fix gdbmhash's remaining function
cotto I'll get the hash wrapper PMCs (md and sha)
It'll be easy since they're all generated. 14:10
or not 14:12
bacek cotto: they already fixed. I've just added "lib" and "group" traits. 14:13
pair.pmc is failing
cotto Can I take that one? 14:14
bacek I finished with dynpmc :) 14:15
dalek rrot: r38309 | bacek++ | branches/pmc_pct/compilers/pmc (3 files):
Add group and lib traits
bacek filehandle and fixedbooleanarray are mine :) 14:16
cotto dibs on... um... 14:17
bacek cotto: you can uncomment "say" in src/parser/action.pm. Then during parsing it will be easy to find "crash" point
cotto rotest and subproxy 14:18
dalek rrot: r38310 | cotto++ | branches/pmc_pct/src/dynpmc/pair.pmc:
[PMC] fix VTABLE declaration in Pair dynpmc
cotto bacek, thanks. So far it's been obvious.
bacek have to steal c_type from languages/c99.pg... 14:19
dalek rrot: r38311 | cotto++ | branches/pmc_pct/src/dynpmc (2 files):
[PMC] fix a couple more dynpmcs to use proper VTABLE function declarations
 14:21
cotto bacek, ROTest is failing because of 'VTABLE INTVAL get_integer() :write {' 14:23
bacek cotto: thanks. Will add handling soon.
cotto method attrs 14:24
That's what they are.
bacek, are you on svn >=1.5? 14:25
bacek cotto: no idea. I'm using git-svn 14:26
cotto It'd probably be a good idea to merge the VTABLE additions back into trunk, since they should have been there in the first place.
Infinoid ... it's snowing
cotto It's sunny!
Infinoid "I thought we already fixed that bug."
bacek cotto: will do. 14:28
purl i think will do. is there a more appropriate channel for me to ask question regarding trycatch in?
cotto There's a certain question mark in lib/Parrot/Pmc2c/Parser.pm:188 that says otherwise.
he Hm, first PR for my packaging of parrot for NetBSD: gnats.netbsd.org/41275/ 14:30
"Installed bits reference the build directory".
I beleive this one has its origins here(?)
(for 1.0.0) 14:31
14:33 PerlPilot joined
dalek rrot: r38312 | cotto++ | branches/pmc_pct/src/pmc/eventhandler.pmc:
[PMC] add missing VTABLE decoration
 14:34
rg he: you're right. i'm not even sure if it has been fixed yet. there should be a tt for it. 14:36
nope. it still behaves that way. 14:39
cotto bacek, working on figuring out why exception.pmc doesn't work... 14:42
bacek cotto: ok. I improved <c_type> slightly. It can probably help. 14:43
dalek rrot: r38313 | bacek++ | branches/pmc_pct/compilers/pmc/src/parser/grammar.pg:
Slightly improve c_type parsing
cotto nope. It's something else. 14:44
bacek np.
What the heck is ':read' and ':write'?
They are exists only in rotest.pmc 14:45
Failed 32/94 subtests
nopaste "cotto" at 96.26.202.243 pasted "minimal failing PMC" (3 lines) at nopaste.snit.ch/16395 14:46
cotto bacek, they're "method attributes". I'd ask allison if we still care about them.
bacek cotto: ok 14:47
dalek rrot: r38314 | bacek++ | branches/pmc_pct/compilers/pmc (3 files):
Add no_ro traits
 14:49
14:52 rdice joined
he rg: it appears that at least TT #540 touches on that issue. 14:52
cotto bacek, the pared-down exception.pmc failure is in the above nopaste. 14:56
I'll take Null next.
bacek where is nopaste?
purl nopaste is at nopaste.snit.ch/ (ask TonyC for new channels) or rafb.net/paste or poundperl.pastebin.com/ or paste.scsys.co.uk/ or App::Nopaste or tools/dev/nopaste.pl or at www.extpaste.com/ or paste.scsys.co.uk (for #catalyst, #dbix-class, #moose and others)
cotto bacek, nopaste.snit.ch/16395 14:57
14:58 cognominal joined
dalek rrot: r38315 | bacek++ | branches/pmc_pct/compilers/pmc/src/parser/grammar.pg:
Allows pointer-to-pointer in c_type.
 14:59
rrot: r38316 | bacek++ | branches/pmc_pct/src/pmc/fixedpmcarray.pmc:
[cage] Add VTABLE to splice in FPA.pmc
rrot: r38317 | bacek++ | branches/pmc_pct/src/pmc/packfile.pmc:
[cage] Fix METHOD declarations in packfile.pmc
bacek cotto: METHODs shouldn't have return type AFAIU
or I can make them optional.
cotto bacek, I don't think it makes sense for METHODs. 15:00
bacek cotto: me either. Check r38317. I actually removing types from METHODs.
cotto looks good 15:01
purl O_O
dalek rrot: r38318 | cotto++ | branches/pmc_pct/src/pmc/null.pmc:
[PMC] remove an unneeded function decoration
 15:02
cotto bacek, I'll take PMCProxy next. 15:03
bacek cotto: ok
I'll take *hash.pmc 15:04
rg he: #495 is also related, but i can't find a definite match. 15:05
cotto bacek, it looks like the same problem as exception.pmc 15:06
bacek cotto: fix it! :) 15:07
cotto bacek, is it safe to commit a grammar.pg change? 15:09
bacek cotto: yes
cotto nm. fix doesn't appear to work. 15:10
bacek cotto: for which pmc?
dalek rrot: r38319 | bacek++ | branches/pmc_pct/src/pmc/hash.pmc:
[cage] Add VTABLE to vtable methods in hash.pmc
 15:11
rrot: r38320 | bacek++ | branches/pmc_pct/src/pmc/orderedhash.pmc:
[cage] Add VTABLE to vtable methods in orderedhash.pmc
cotto the example I nopasted, which happens in exception.pmc and pmcproxy.pmc
bacek should we allow return type for METHODs? 15:12
cotto It doesn't really make sense. I'm inclined to disallow it and fix anything that breaks because of it. 15:13
It's all PCC'd out the wazoo anyway.
he rg: mm, yes, in the same class of problem. I guess this issue is well enough documented, so there's no need for a new ticket?
bacek so, just fix exception.pmc
and pmcproxy
purl hmmm... pmcproxy is basically Class for PMCs.
bacek I'll take *sub 15:14
cotto my brain's going. I think I'm done for now. 15:15
rg he: whatever you feel is best. You could at least comment on one of the tickets to point at the netbsd one. 15:16
cotto It looks like comments after the closing bracket mess up the parser
15:17 fperrad1 joined
bacek cotto: it's possible... 15:17
dalek rrot: r38321 | bacek++ | branches/pmc_pct/src/pmc/multisub.pmc:
[cage] Fix multisub.pmc VTABLE and METHODs declarations.
 15:18
rrot: r38322 | bacek++ | branches/pmc_pct/src/pmc/sub.pmc:
[cage] Fix sub.pmc VTABLE declarations.
15:19 fperrad1 left
bacek ok, *array.pmc and enough for today. 15:19
dalek rrot: r38323 | bacek++ | branches/pmc_pct/src/pmc/string.pmc:
[cage] Fix string.pmc VTABLE declarations.
 15:21
rrot: r38324 | cotto++ | branches/pmc_pct/src/pmc/pmcproxy.pmc:
[PMC] add a missing VTABLE decoration
cotto bacek, it would be nice if the parser eventually got smart enough to say what's wrong. 15:22
bacek cotto: yeah. I'll try to do something with it. 15:23
cotto "eventually" is fine by me
bacek cotto: done :) 15:27
cotto or that 15:28
btw, have you run make test?
bacek for selected pmcs
r38326 is "smart" parser
15:29 fperrad joined
moritz you can easily get error messages from PCT with the ~ syntax 15:29
dalek rrot: r38325 | bacek++ | branches/pmc_pct/src/pmc/resizablepmcarray.pmc:
[cage] Fix RPA VTABLE and METHOD declarations.
 15:30
rrot: r38326 | bacek++ | branches/pmc_pct/compilers/pmc/src/parser/grammar.pg:
Add little bit brains to parser as cotto++ suggested.
rrot: r38327 | bacek++ | branches/pmc_pct/src/pmc/resizablestringarray.pmc:
[cage] Fix RSA VTABLE and METHOD declarations.
bacek oops.
It's wrong parser.
cotto make test looks good 15:31
dalek rrot: r38328 | cotto++ | branches/pmc_pct/src/pmc/exception.pmc:
[PMC] fix some METHOD declarations in the Exception PMC
 15:33
bacek all test successful on my box 15:35
not compilers/pmc/t/02 :)
cotto same here 15:36
How many failing PMCs remain?
dalek rrot: r38329 | bacek++ | branches/pmc_pct/compilers/pmc/src/parser/grammar.pg:
Fix "smart" parsing. bacek--
 15:37
cotto The test takes 6.5m to run on my box.
bacek 02-all? 15:38
cotto Heh. I only see 2 failures: pccmethod_test and rotest. Maybe there's a pattern. 15:40
bacek Failed 2/94 subtests 15:41
Hooray!
pccmethod_test expose bug in parser 15:42
cotto ok. I'll let you fix it. I'm going to sleep. 15:44
gg
bacek here we go 15:45
I'm going to sleep too
one failing test left :)
dalek rrot: r38330 | bacek++ | branches/pmc_pct/compilers/pmc/src/parser/grammar.pg:
Add just :named adverb.
 15:46
bacek cotto: good night. It was good mini-hackatron :)
15:51 jan joined 15:57 darbelo joined 16:29 flh joined 16:31 jhorwitz joined 17:13 ruoso joined 17:43 mikehh joined 18:17 davidfetter joined
Coke pmichaud: reading a blog entry about what would happen to the electoral college if texas exercised its right to split up. =-) 18:21
18:34 he left 18:40 cognominal joined
Coke allison: ping. 18:40
dalek rrot: r38331 | coke++ | trunk/config/gen/platform/netbsd (2 files):
[cage] fix file properties
rtcl: r328 | wcoleda++ | trunk/ (14 files):
Move closer to building against an installed parrot.
 18:46
18:46 wayland76 joined
dalek rtcl: r329 | wcoleda++ | trunk/Configure.pl:
Give up sooner if we have no parrot_config
 18:52
Infinoid just gave his 2 weeks notice at $job 18:57
wayland76 Infinoid: What are you doing next? (If we're allowed to ask) 18:58
Another $job?
Infinoid I assume so, I don't have anything lined up yet. 18:59
wayland76 OK
Infinoid Maybe I can find someone to pay me for waving my patches angrily at people on the internet, I'm good at that. :)
NotFound I'd like that someone paid me for ranting on irc 19:00
wayland76 It might be just as much fun if you did it without the anger :)
Infinoid hehe
NotFound Even better, for doing quality control of beds
wayland76 I'd like someone to pay me for doing whatever open-source work I thought important :) 19:01
Infinoid Yeah, that sounds nice, but unlikely
wayland76 NotFound: Took me a while, but I just got it :)
Infinoid oh well, I'll figure something out
NotFound wayland76: me also, but I think I have more probablities with the other two X-)
wayland76 Infinoid: What kind of job was it? Sysadmin/Web/Coding? 19:02
darbelo Got to school and apply for GSoC. That's exactly what it is.
NotFound I'm too old for school
Infinoid embedded coding
NotFound ... too young to die
wayland76 Right. I found the more Perl you do, the more you hate embedded :) 19:03
Infinoid strangely, I love them both
I tend to love the low level parts of both... probably why I'm so good at XS
wayland76 Ah. I like to ignore resource constraints :) 19:04
Infinoid sadly, noone around here seems to need a bios or driver written, it's all web nonsense here
hehe
wayland76 I did some embedded, with no prior experience
Infinoid (the low level stuff is also why I love parrot)
NotFound wayland76: then I recommend you z80 assembler programming X-)
Infinoid I haven't done z80, done some 8051 and some avr tho 19:05
wayland76 My Dad needed a 4-line terminal programmed for a robot program he'd written
Infinoid cool
wayland76 And he was running out of time on his project
NotFound I do the fun part: write the assembler :) 19:06
wayland76 (When I say robot, I mean a big machine that picks up racks, not humanoid in any way)
Infinoid (wayland76's dad)++
dalek rrot: r38332 | coke++ | trunk/ports/macports/Portfile:
[port] bump macport to 1.1
wayland76 I got to do it in C, with co-routines
darbelo Pfft. REAL embedded work is done with an hex editor. 19:08
Infinoid but a jtag debugger sure helps
NotFound Someone can verify that tools/dev/mk_language_shell.pl doesn't work without an installed parrot? 19:09
Infinoid I don't have an installed parrot, should I just run it and see what happens?
NotFound Can't open perl script "/usr/local/lib/parrot/1.1.0-devel/tools/dev/gen_makefile.pl": No such file or directory
That is what happens for me
Infinoid first it complained about not having a configured parrot 19:10
then I got:
[snip]
creating /home/infinoid/parrot-dev/test/languages/foo/t/00-sanity.t
Unable to locate parrot_config. at Configure.pl line 45.
so I guess it wants a fully built parrot 19:11
NotFound It does not complain for me, just fails.
Infinoid did you run it from the parrot source checkout directory? 19:12
NotFound Yes
Infinoid ok, me too.
NotFound With parrot already built, but not installed
Infinoid it created languages/foo/ but failed to configure it (because parrot wasn't built yet)
I'll try again when parrot is built
NotFound Sorry if don't explain well
Infinoid Reading configuration information from ../../parrot_config
Creating Makefile
Can't open perl script "/usr/local/lib/parrot/1.1.0-devel/tools/dev/gen_makefile.pl": No such file or directory 19:13
is that what you got?
NotFound Yes
Infinoid ok, verified.
NotFound I suppose that is not intended, and I must open a ticket
Coke didn't pmichaud just check in a replacement for mk_lan* ? 19:14
Infinoid I dunno, my git bounce box is gone and I haven't set up a new one yet
NotFound I'm using trunk r38331
Infinoid so I'm still at r38284
Coke s/just/last week or so/ 19:15
darbelo Coke: you mean tools/dev/create_language.pl ?
NotFound Trying that... 19:16
Unable to locate parrot_config. --> I suppose it needs to have it in the PATH ? 19:17
Infinoid if it's expecting an installed parrot, I would guess so
Coke rant: having to approve messages from google code commit to a google groups list. 19:20
NotFound Setting the path it works, but make test seems to have the language name 'perl6' hardcoded
$ ack perl6
t/harness
53: exec => ['./perl6'],
purl rumour has it t/harness is setting extra parrot flags
NotFound Created TT#580 about this 19:27
19:34 gryphon joined
Coke hurm. that portfile for 1.1.0 for mac fails to run. 19:40
19:55 ruoso joined 20:28 gryphon joined
flh little question about PCC: does anyone know which data hide in INTERP->current_args? I think it's a set_args op (i.e. the signature is at current_args + 1, and actual args are at current_args + 2), but can someone confirm? 20:53
I'm a bit confused because Parrot_pcc_invoke_sub_from_c_args (for example) doesn't seem to set the opcode in *current_args, nor a pointer to the signature in current_args+1 20:55
Infinoid I think the current_args handling is done within the stuff autogenerated by pmc2c, see rewrite_pccmethod() in lib/Parrot/Pmc2c/PCCMETHOD.pm for example 20:57
20:58 braceta joined 21:01 braceta left, braceta joined
flh ok, so I guess your answer means that I'll have to dive into Parrot_process_args 21:02
Infinoid I don't know the formats or offsets of what it points to, sorry
flh and, well, let's try another question: IMCC handles ".const <type> foo = 'bar'" differently when <type> is 'Sub' 21:03
i.e., when <type> is not 'Sub', but another PMC, this calls the VTABLE instantiate_str
is there a good reason for not doing the same thing (i.e., doing everything in the Sub's intantiate_str) for Subs? 21:04
21:05 Casan joined, whoppix joined
flh this seems to prevent me from playing with my pmc, which extends Sub (it cannot inherit the behaviour of the Sub when I do ".const 'mysub' foo = 'bar'") 21:05
Infinoid flh: if you don't get an answer here, I suggest you ask the parrot-dev list. I think the right people to ask are allison or maybe chromatic, but they aren't always paying attention to irc 21:06
flh ok, it was worth trying to get an answer before going to bed :) 21:08
NotFound I think this was mentioned some time ago, but don't remember if there is a ticket.
particle flh: that's a known limitation of imcc, and it's on our roadmap to fix. not sure the timeline, though 21:09
mailing list will tell you, i'm sure
flh thanks, I'll take some time tomorrow to look in RT and Trac if there is a ticket about that, and ask on parrot-dev 21:12
21:18 Whiteknight joined 21:21 szabgab joined
dalek rrot: r38333 | whiteknight++ | trunk/docs/book/ch08_dynops.pod:
a few changes to chapter 8 for readability
 21:33
allison driving to speak at a conference 21:42
22:10 braceta joined 22:14 he joined 22:19 Casan joined 22:38 Casan joined
dalek rrot: r38334 | whiteknight++ | trunk (2 files):
Remove some weird macros for converting a pointer value to an opcode_t value. These macros were only used in one place and were used incorrectly (although I don't think the incorrect usage raised any warnings, which is weird). I've replaced it with a better macro for handling ptr->INTVAL conversions
 22:51
23:00 Xrblackwe joined 23:12 tetragon joined 23:19 Coke joined, davidfetter joined
dalek rtcl: r330 | wcoleda++ | trunk/Configure.pl:
We like lowercase.
 23:22
cotto darbelo, ping 23:28
darbelo pong
dalek rtcl: r331 | wcoleda++ | trunk/config/makefiles/root.in:
Define variable used in makefile
cotto darbelo, an easy way to get the svn repo started would be to use mk_language_shell.pl. 23:31
darbelo but that creates a language :) 23:33
cotto yes, but it also does the dynpmc stuff
darbelo I created a dynpmc with tools/dev/gen_class.pl 23:35
cotto That doesn't give you a nice makefile template, though.
just a stub pmc 23:36
darbelo Had it build with a hacked up Makefile (half-copied from src/dynpmc/), but the I broke it. 23:37
23:38 amoc joined 23:39 rdice joined
darbelo also, I'm not entirely sure where I need to put the .o and .so for parrot to find. 23:44
cotto Parrot should only need the .so (or dll for windows) 23:45
darbelo I'm guessing runtime/parrot/dynext/ then. Right? 23:46
cotto sounds likely 23:48
Is that what make install does? 23:49
darbelo Thats where the 'built-in' dynpmc's go. 23:50
the README there claims "This directory holds dynamic Parrot extension files." 23:51
23:52 bacek joined
Infinoid happy weekend! 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!