Parrot 2.7.0 "Australian King" Released! | parrot.org Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | close 13 tickets, merge outstanding branches, help NotFound clean up PBC tests
Set by moderator on 20 August 2010.
00:01 Psyche^ joined
bacek_at_work Good morning, humans 00:06
00:06 Patterner left, Psyche^ is now known as Patterner, mj41_ joined 00:09 mj41 left, mj41_ is now known as mj41
Paul_the_Greek Hey Bacek. 00:10
whiteknight hello bacek 00:11
dalek rrot-linear-algebra: 4e23a2f | Whiteknight++ | s (10 files):
add in the machinery to include some random .c files into the dynpmc build. Exercise this mechanism by moving some of the routines I have into separate files where they can (hopefully) be shared among multiple types
 00:12
rrot-linear-algebra: c20761a | Whiteknight++ | setup.nqp:
add logic to detect one version of lapack at one location, and nclude it in the build. No calls to it yet
Paul_the_Greek What determines whether a PMC should extend the scalar PMC? 00:24
sorear There are no rules. Sorry. 00:27
dalek rrot: r48599 | jkeenan++ | trunk/config/auto/pmc.pm:
Delete commented-out code.
 00:29
rrot: r48600 | jkeenan++ | trunk/t/steps/auto/pmc-01.t:
Test previously uncovered condition.
rrot: r48601 | jkeenan++ | trunk/examples/languages/squaak/doc/tutorial_episode_1.pod:
[codingstd] Correct excessive linelength.
00:29 dngor left
Paul_the_Greek Why does Integer extend scalar but BigInt doesn't? 00:30
00:31 brooksbp left
sorear I'd be very suprised if there was a reason 00:32
Paul_the_Greek Hmm. So if scalar has a useful method or two, I might inherit from it. If not, I wouldn't. 00:34
00:34 hudnix left
Paul_the_Greek That's confusing. 00:35
00:36 dngor joined
kid51 msg chromatic Can you look at include/parrot/hash.h with respect to codingstd requirements, e.g., c_macro_args ? Thanks. 00:37
purl Message for chromatic stored.
whiteknight Paul_the_Greek: As far as I am concerned nothing should inherit from Scalar, and Scalar should be deleted
Paul_the_Greek The docs say that only Foo provides Scalar, which is also a lie. 00:38
whiteknight: Sounds reasonable to me.
whiteknight probably requires a deprecation notice for some reason, but I can't imagine that Scalar PMC provides anything 00:39
Paul_the_Greek It's 970 lines of nothing, then. 00:40
Let me check for tickets ...
There is one ticket that says that scalar is the only PMC providing logical_not. 00:43
dalek rrot: r48602 | jkeenan++ | trunk/config/auto/pmc.pm:
[codingstd] Eliminate trailing whitespace.
 00:45
rrot: r48603 | jkeenan++ | trunk/t/op/gc-leaky-call.t:
Set file metadata properties.
rrot: r48604 | jkeenan++ | trunk/src (4 files):
[codingstd] Enforce c_parens rule.
00:46 darbelo left
Paul_the_Greek Seems like it should go away or else be refactored into something useful. 00:46
00:47 darbelo joined
Paul_the_Greek Maybe some things should be moved in the default PMC. 00:47
whiteknight it's a throwback to back when Parrot was more Perl-oriented 00:49
Paul_the_Greek Perhaps I'll add a ticket and undertake a little research project at some point. 00:51
whiteknight There really hasn't been too much discussion about it. I would like to see it removed but I haven't pursued it. I do wonder what other developers think about it
Paul_the_Greek I'll add a ticket. 00:52
dalek rrot-linear-algebra: b530931 | Whiteknight++ | src/ (4 files):
a few small cleanups
 00:54
00:59 whiteknight left 01:00 Paul_the_Greek left
dalek TT #1754 created by Paul_the_Greek++: The Scalar PMC 01:00
TT #1754: trac.parrot.org/parrot/ticket/1754
rrot: r48605 | jkeenan++ | trunk/tools/dev/as2c.pl:
Correct POD formatting error detected by t/codingstd/perlcritic.t.
 01:02
01:11 contingencyplan left, contingencyplan joined
dalek rrot: r48606 | jkeenan++ | trunk/include/parrot/hash.h:
Correct spelling error in inline comment.
 01:35
rrot: r48607 | jkeenan++ | trunk/t/codingstd/c_operator.t:
[codingstd] Make test output more useful by indicating line in file at which standard was
kid51 Unfortunately, even with that improvement in c_operator.t, I can' identify why include/parrot/hash.h is failing the test. 01:39
Coke Ugh. what is putting "EXPORT" in my HLL namespace? 02:08
02:09 davidfetter joined 02:10 hudnix joined 02:12 kid51 left 02:35 janus left 02:41 janus joined 02:56 antivigilante joined
antivigilante how do I get squaak - i'm on debian 02:57
GeJ squaak? 03:01
purl i think squaak is languages/squaak or the language in kj's tutorial at parrotblog.org or svn.perl.org/parrot/trunk/languages/squaak/
Coke i don't think there's a separate install, antivigilante - it's just a demo language included in the tarball (or svn repo, if you prefer) 03:03
GeJ antivigilante: It looks like squaak a left the nest. github.com/ekiru/squaak-tutorial. 03:04
tcurtis It's in examples/languages/squaak in the Parrot repository. 03:05
antivigilante: GeJ: don't go to that github repo. I just had that there for a while when I was updating it. The version in Parrot's subversion repository is the most up-to-date version. 03:06
GeJ tcurtis: thanks for the update. 03:08
purl: no, squaak is examples/languages/squaak
purl okay, GeJ.
GeJ squaak? 03:09
purl squaak is, like, examples/languages/squaak
tcurtis squaak?
purl squaak is examples/languages/squaak
GeJ purl squaak is also the language in kj's tutorial at parrotblog.org
purl okay, GeJ.
GeJ botsnack
purl :)
antivigilante the tute sez make install-dev, would it be in parrot-dev on debian? 03:11
Coke the tute? 03:12
antivigilante tutorial
purl somebody said tutorial was the Parrot interpreter tutorial at en.wikibooks.org/wiki/Parrot_Virtua...troduction
Coke I have no idea how debian is packaging things, but "install dev" is really old.
it's just "install" these days.
does debian have a "parrot-dev" and a parrot? (I can't imagine you'd not want parrot-dev if they split it up.)
antivigilante parrot-devel and libparrot-dev 03:14
one sec
GeJ I'm no debian expert, but I think that in debian lingo, -dev means header files and such.
All you need when you want to compile additional software on your own against external libs. 03:15
But again, I'm no expert.
03:15 antivigilante left 03:36 antivigilante joined 03:42 Coke left, Coke joined 03:47 khairul left 03:56 Coke left 04:02 Coke joined, khairul joined
cotto hio khairul 04:08
khairul hio cotto. i'm slowly moving the code over to here, github.com/khairulsyamil/parrot-instrument 04:21
cotto cool 04:31
04:47 khairul left
antivigilante the ekiru link is the only one that works for me' 05:23
05:41 davidfetter left
cotto python-- 05:48
"multiple repeat" is not a valid way to tell me a regex has a syntax error 05:49
tcurtis antivigilante: docs.parrot.org/parrot/latest/html/...orial.html
05:52 antivigilante left 06:02 uniejo joined 06:40 fperrad joined
cotto msg dafrito Could you make map-svn also grab commit messages? 06:42
purl Message for dafrito stored.
dalek thub-trac: c2fcc54 | ++ | github/github.py:
fix code to force rereading of the revmap
 06:51
thub-trac: 3c9a116 | ++ | github/github.py:
Nuke code to force rereading of the revmap. This is infrequent enough that deleting the table will be fine.
thub-trac: e65f5d0 | ++ | github/github.py:
initial somewhat working version of wiki links
thub-trac: 09f1817 | ++ | github/github.py:
remove some debugging code
thub-trac: b15d52d | ++ | github/github.py:
add stub tooltip support
rrot: r48608 | NotFound++ | trunk/t/pmc/exception.t:
test Exception is_equal
 06:54
06:58 tcurtis left
dalek rrot: r48609 | NotFound++ | trunk/t/pmc (7 files):
update run instructions in packfile test files, TT #1750 TT #1712
 07:27
07:35 tadzik joined 08:24 tadzik left
ttbot Parrot trunk/ r48610 i386-linux-thread-multi make error tt.taptinder.org/file/cmdout/373738.txt ( tt.taptinder.org//buildstatus/pr-Pa.../rp-trunk/ ) 08:25
dalek rrot: r48610 | mikehh++ | trunk/include/parrot/hash.h:
fix codetest failures - unwrapped macro arguments and
 08:34
rrot: r48611 | mikehh++ | trunk/include/parrot/hash.h:
does not seem to like wrapping _code
 08:51
09:36 ruoso left 09:59 tadzik joined 10:07 khairul joined 11:45 whiteknight joined 11:57 lucian joined
whiteknight good morning, #parrot 12:03
12:03 nwellnhof joined 12:04 particle left 12:07 particle1 joined 12:15 ruoso joined 12:16 Paul_the_Greek joined
Paul_the_Greek Good morning, folks. 12:16
12:27 lucian left, lucian joined 12:28 lucian left 12:29 lucian joined
Coke hio. 12:29
nwellnhof Is anyone knowledgeable about src/call/args.c around?
moritz I can't offer knowledge, only ops :-) 12:30
Coke ugh. I have 17 namespaces in .HLL 'tcl' that shouldn't (eventually) be there.
Paul_the_Greek nwellnhof: I'm not knowledgeable, but I am looking at it. Yikes. 12:31
Coke we've hit our goal of closing N tickets, yes? 12:32
Paul_the_Greek I think we're certainly close.
moritz I think bacek_at_work and whiteknight did some hacking on the calling conventions, no?
Paul_the_Greek I didn't help by opening two new ones. :D 12:33
nwellnhof I think we can close TT#945 and TT#1465.
Paul_the_Greek Is there a test for #1465, nwellnhof? 12:35
nwellnhof Yeah, I added tests for both tickets last week. 12:36
t/op/gc-leaky-call.t and t/op/gc-leaky-box.t
Paul_the_Greek Let me look at those just to see how you tested a leak. 12:37
moritz it would be even more robust to compare memory usage before and after the loop 12:39
instead of using a hard limit 12:40
nwellnhof Currently, the memory usage of all smaller programs will be between 1MB and 2MB. 12:41
Coke TT: 662 active tickets.
we have closed 10 tickets in the past 6 days. 12:42
3 tickets left to meet the goal.
nwellnhof Did you count TT#1121?
I closed it last week as invalid. 12:43
Paul_the_Greek Coke: I have a patch that is ready to commit. If you want to do it, you can close the ticket.
Coke I counted everything that was closed in the past six days according to: trac.parrot.org/parrot/timeline?fro...ate=Update 12:45
Paul_the_Greek: ... don't you have commit privs?
Paul_the_Greek I do, but I'm on vacation and will wait to start committing until I return home. 12:46
Here is the ticket: trac.parrot.org/parrot/ticket/481; cotto has tested it.
NotFound Paul_the_Greek: I've just read your last comment in TT #1753. If the fixed size allocator has a problem is the fixed size allocator who needs a fix, no the attr struct size.
Coke looks like #17222 is closable. 12:47
looks like #1722 is closable.
Paul_the_Greek NotFound: Agreed. The PMC attributes allocator has to deal with the issue.
I'm going to look at that allocator to possibly speed it up, so I'll deal with the alignment issue too. 12:48
NotFound Paul_the_Greek: ok
nwellnhof Paul_the_Greek: The allocation is pretty fast already. The garbage collection is where most time is spent. 12:49
moritz rakudo leaks on simple loops (again) :(
NotFound Coke: I think we can close it, 3 weeks without answers is more than enough 12:50
Paul_the_Greek One source of ticket closure are the ones with pending patches. Some of them are quite old and possibly irrelevant now.
nwellnhof moritz: Which parrot revision?
moritz nwellnhof: r48587 - but I fear it's a rakudo bug, not parrot 12:51
Paul_the_Greek For example: trac.parrot.org/parrot/ticket/951
NotFound Coke: closed
Coke if a patch no longer applies, the ticket is rejectable, yes, along with a nice note that we're sorry it took so long and please rebase and resubmit if this is still an issue." 12:52
NotFound++
2 to go to meet our target.
nwellnhof moritz: r48587 contains my leak fixes
dalek rrot-linear-algebra: 020f27c | Whiteknight++ | setup.nqp:
add shebang to setup.nqp. I think I also gave it +x
 12:54
rrot-linear-algebra: d4ef78b | Whiteknight++ | setup.pir:
remove the old setup.pir. Everything should be done with the new setup.nqp now
rrot-linear-algebra: 1277e1c | Whiteknight++ | ROADMAP.pod:
delete ROADMAP.pod. The items discussed there are already completed or in current progress. Additional issues and TODO items will be tracked in the github issue tracker at github.com/Whiteknight/parot-linear-algebra.git
rrot-linear-algebra: c3fe221 | Whiteknight++ | PLATFORMS:
Update PLATFORMS with inforation about Ubuntu10.04 with clang1.1. Some of these older records need to be updated/deleted as well. In the future I will probably only discuss recent LTS versions of Ubuntu and the current development version I am using, unless I find a testing champion for other systems.
nwellnhof Coke: I can close #945 and #1465. Both have a test case now. I just wanted to wait a little in case there's something wrong with my patch.
Paul_the_Greek nwellnhof: I'll benchmark any allocation speed-ups that I make. You may be right that I can't make much of a dent.
nwellnhof Paul_the_Greek: Every little helps.
But, of course, there's Amdahl's law. 12:55
Paul_the_Greek nwellnhof: Which benchmark do you think I should use that is PMC allocation/free heavy?
NotFound Loose pressusre on the gc or the allocators is good, even if simple benchmarks don't show clear improvements.
moritz compiling rakudo is a pretty good GC benchmark :-) 12:56
nwellnhof Paul_the_Greek: Basically, everything that isn't pure integer and number benchmarks. 12:57
Paul_the_Greek I also have to invent a custom PMC to demonstrate the attribute alignment problem. 12:58
NotFound Compiling winxed is a simple benchmark, and a lot short than compiling rakudo ;)
Paul_the_Greek Is it legitimate to add an experimental PMC to the system and then remove it later?
What is winxed? 12:59
purl hmmm... winxed is not javascript, just uses a familiar syntax. or code.google.com/p/winxed/ or winxed.org/
nwellnhof Paul_the_Greek: You should that in an own branch.
s/should/should do/
NotFound Paul_the_Greek: yes, but you need to bump PBC_COMPAT in both steps.
Better in a branch, yes 13:00
moritz and each time you break my rakudo setup, and that damn error message from parrot doesn't even tell me which PBC it failed to read
Paul_the_Greek Okay, I'll make a branch. Good to learn how to do that.
nwellnhof Back to my src/call/args.c issue: The code starting at line 491 uses a really baroque way to iterate over a hash. 13:01
I think it could be replaced by the new hash iteration macro. But maybe I'm missing something. 13:02
moritz it was probably written to avoid calls 13:03
nwellnhof That code is also the only user of parrot_hash_get_idx which could go away then.
moritz which would have made it recurse
dalek TT #1722 closed by NotFound++: osx build failed caused by errant '\\c'
TT #1722: trac.parrot.org/parrot/ticket/1722
nwellnhof It looks like remnants from a strange past. 13:05
moritz just replace it, and if all tests pass, commit it
dalek rrot-linear-algebra: bd96745 | Whiteknight++ | README:
Updates to README
 13:06
Paul_the_Greek NotFound: I have a pending patch that adds opcodes. Do I need to deal with PBC_COMPAT?
NotFound Paul_the_Greek: yes 13:09
With experimental opcodes it may not have effect, but better to be on the safe side. 13:10
Paul_the_Greek NotFound: So I see the PBC_COMPAT history file that needs to be edited. Is there something else to do?
NotFound Paul_the_Greek: I hope not, the problems with the native_pbc test files should be out of the way right now. 13:12
Paul_the_Greek NotFound: What about ops.num?
NotFound Paul_the_Greek: experimental ops have its own numbering, not affected. 13:13
Paul_the_Greek Sorry, talking about adding permanent opcodes.
NotFound Paul_the_Greek: what opcodes? 13:14
purl opcodes are Good!
Paul_the_Greek NotFound: Here's the ticket: trac.parrot.org/parrot/ticket/481
Two native integer opcodes and two Float PMC opcodes. 13:15
NotFound Fine then. But I don't know what's the current state on ops numering policy and tools.
Paul_the_Greek I'll ask cotto, since he reviewed the patch. At least I have to edit PBC_COMPAT. 13:17
msg cotto Regarding ticket #481, I need to edit PBC_COMPAT. Anything else regarding opcode compatibility or numbering? 13:18
purl Message for cotto stored.
dalek TT #945 closed by nwellnhof++: Sub invocation consumes all memory. (needs testing) 13:20
TT #945: trac.parrot.org/parrot/ticket/945
TT #1465 closed by nwellnhof++: Parrot GC leaks memory for basic programs
TT #1465: trac.parrot.org/parrot/ticket/1465
Paul_the_Greek Can I tell online whether I have commit privilege? 13:25
moritz if you can commit to parrot with your trac username and password, you have commit privs 13:26
but I would be suprised if somebody gave you commit privs without telling you :-)
so if you don't know, the answer is likely "no"
Paul_the_Greek I was supposed to get commit privilege, but I'm not at home so I haven't tried one yet. 13:27
13:34 Paul_the_Greek left
dalek rrot: r48612 | nwellnhof++ | trunk (3 files):
Use hash iteration macro in dissect_aggregate_arg
 13:35
13:55 uniejo left 14:20 Andy joined
whiteknight ping ash_, plobsing 14:43
Coke msg paul_the_greek - the typical way to check is to add yourself to AUTHORS. 14:45
purl Message for paul_the_greek stored.
whiteknight Does anybody know if ash_'s GSoC project completed successfully? I haven't heard from him in a while 14:50
Coke Hit the mentor? 14:58
dalek rrot: r48613 | nwellnhof++ | trunk/src/hash.c:
Fast path for hash get/put with string keys

Also remove the COW test from STRING_compare. It was wrong (should have used strstart instead of bufstart) and didn't give a speedup.		 14:59
moritz seen plobsing 15:00
purl plobsing was last seen on #parrot 6 days, 9 hours, 22 minutes and 4 seconds ago, saying: remind me never to give you patches on release day [Aug 17 05:38:47 2010]
15:08 theory joined
Coke partcl-nqp: puts [namespace children] 15:13
p6eval partcl-nqp: OUTPUTĀ«ā¤Ā»
Coke moritz: what was your suggestion for tracking version information in partcl in the evalbot? 15:14
moritz Coke: make the default 'make' target write a revision file
which p6eval then can read
Coke: have you fixed --gen-parrot in partcl-nqp? 15:15
Coke ".revision"?
moritz wfm
Coke moritz: no, I never use it. how is it broken?
moritz partcl-nqp: puts [namespace children]
p6eval partcl-nqp: OUTPUTĀ«::Partcl ::String ::ResizablePMCArray ::Hash ::Info ::Array ::FileGlob ::StringGlob ::_tcl ::Dict ::Regex ::PCT ::ARE ::PAST ::HLL ::Namespace ::EXPORTā¤Ā»
Coke (none of those should be there. =-) 15:16
moritz Coke: last time I tried it didn't execute make or make install or so
Coke: point is, if I want to install a cron job for updating partcl-nqp, I need a reliable way to get the correct parrot revision
and automatic 15:17
purl automatic is probably youtube.com/watch?v=tIRCzbBB99E
Coke ok. I just stole gen-parrot from rakudo at some point. I'll open a ticket so i don't forget this time.
15:18 theory left
Coke moritz++ #tickets opened. I'll cc you when they get closed. (code.google.com/p/partcl/issues/list) 15:20
Coke finds an old ticket pondering something like pir::say hi ;# in partcl. 15:23
moritz that's mostly valuable if you want to write the built-ins in TCL 15:27
if not, why bother?
Coke it's mainly helpful for developer diagnostics. 15:28
old partcl let you do: [inline PIR {trace 1}] , etc.
this is actually less useful with PCT, I suppose.
Coke ponders just killing all the old partcl tickets. 15:29
ugh, partcl is failing tests again. 15:30
15:31 cotto_work joined
cotto_work ~~ 15:32
15:34 brianwisti joined
dalek rtcl: 22bda21 | coke++ | runtime/tcllib.pir:
These are now dyn-PMCs, load them.
 15:42
15:42 pyrimidine joined
15:45 mikehh_ is now known as mikehh
mikehh opbots, names 15:45
15:57 ruoso left 16:01 Paul_the_Greek joined
Paul_the_Greek purl,messages 16:01
ping Coke 16:06
Where is the file called AUTHORS? 16:09
16:11 tcurtis joined
cotto_work What makes the assumption that we have such a file? 16:12
16:13 nwellnhof left
Paul_the_Greek Coke mentioned it in a msg to me. Did he mean CREDITS, I wonder? 16:13
cotto_work probably
Paul_the_Greek cotto_work: For my patch that adds the math opcodes, I also need to edit PBC_COMPAT, right? 16:14
16:14 tadzik left
cotto_work which tt was it? 16:14
Paul_the_Greek Getting ...
cotto_work found it 16:15
Paul_the_Greek Ticket: trac.parrot.org/parrot/ticket/481
Is there a way to tell, online, whether I now have commit privilege? 16:16
cotto_work You don't need to bump PBC_COMPAT because just adding dynops.
Adding core ops (src/ops/*.ops) requires a PBC_COMPAT bump. 16:17
Paul_the_Greek Right, that makes sense.
cotto_work Hmmm. Coke can tell online but I don't think trac tells you if you have an active commit bit. 16:18
(Coke has Trac admin privileges.)
Paul_the_Greek When I asked him about it, he said "the typical way to check is to add yourself to AUTHORS." 16:19
What do you suppose he meant?
cotto_work Ah. He definitely meant CREDITS.
Paul_the_Greek Oh, so he meant I should try to commit a change to CREDITS. 16:20
cotto_work Yup.
Paul_the_Greek It'll wait until I get home. 16:21
cotto_work ok 16:22
Paul_the_Greek Are you the one who suggested a benchmark for PMC allocation/free timing? 16:23
cotto_work I don't think so.
16:24 bubaflub joined
Paul_the_Greek primes2 looks like a good choice. 16:32
16:32 ambs joined 16:34 ambs left 16:46 ruoso joined 16:52 khairul left 16:55 theory joined 16:57 Paul_the_Greek left 16:58 icarroll left
Coke yes, I said the wrong file. 17:14
That's why I suggested that one; seemed like a very low risk commit. =-)
cotto_work though reverts are cheap 17:19
17:22 brianwisti left 17:23 nwellnhof joined 17:36 Paul_the_Greek joined
Coke msg chromatic Parrot::Embed failure: rt.perl.org/rt3/Ticket/Display.html?id=64124 17:37
purl Message for chromatic stored.
Coke msg chromatic (though i did no diligence other than asking Gabor if he was sure. =-)
purl Message for chromatic stored.
Paul_the_Greek Which file has all the ASSERT macro definitions? 17:40
Coke they are spread out. 17:41
cotto_work They're generated by headerizer. 17:42
Coke if you're in a .c file, look in the corresponding .h (listed near some HEADERIZER comment)
Paul_the_Greek But the base ones such as ASSERT and ASSERT_ARG?
Coke or: vi -t ASSERT_ARGS_PackFile_Constant_pack 17:43
Paul_the_Greek No source tree here at the beach; no decent editor.
Coke if you have no source tree, I'm wondering why you're asking these questions. =-) 17:44
particle1 he's reading parrot fanfic
Paul_the_Greek No source tree on this laptop that I can search with a decent editor. 17:45
Coke where do you see a bare ASSERT? 17:46
PARROT_ASSERT_ARG is defined in include/parrot/exceptions.h 17:47
17:47 theory left
Paul_the_Greek There it is. Thanks. And, oh look, PARROT_ASSERT. 17:49
Coke ah. so you didn't mean "ASSERT".
if you have network access, you might want to request an account on feather.perl6.nl from Juerd. Then you can have whatever you want available, if you have a network cxn.
(as long as it's linux-colored) 17:50
moritz or has putty
Coke moritz: ... I mean to say that feather is a linux box.
Paul_the_Greek Right, didn't see plain ASSERT. The 'PARROT_' prefixes kind of fade into the background. 17:51
17:58 tcurtis left
cotto_work msg dafrito Here's a good test case: make a page with ~1000 md5sums and see how the github plugin handles it. 17:59
purl Message for dafrito stored.
18:00 TimToady left
dalek kudo: 689bb75 | moritz++ | src/builtins/Parcel.pir:
fix %() segfault. Patch courtesy of TiMBuS++
 18:01
cotto_work 53 seconds 18:02
with plenty of room for optimization
Paul_the_Greek What takes 53 seconds?
cotto_work a trac wiki page with >1000 md5sums rendered with the github plugin's wiki syntax extension 18:03
(add a link and tooltip to any mentions of svn revisions or git changesets) 18:04
vs .7 seconds with the plugin disabled
18:20 Paul_the_Greek left
dalek rrot: r48614 | nwellnhof++ | trunk/src/hash.c:
Optimize expand_hash

update of free list. Also don't assume that new_size = 2 * old_size.		 18:20
rrot: r48615 | nwellnhof++ | trunk/src (2 files):
Don't recompute hash value of strings in expand_hash
whiteknight cotto_work: is the plugin written in Parrot? 18:29
cotto_work Mu 18:30
18:31 tcurtis joined
cotto_work hopes he wasn't supposed to take that question seriously 18:32
18:48 sorear left 18:59 sorear joined 19:11 tcurtis left
dukeleto 'ello 19:11
whiteknight hello dukeleto 19:26
19:30 bubaflub left
whiteknight dukeleto: Can you add people to the parrot organization? 19:31
and by "people", I mean "me"?
cotto_work me too 19:35
atrodo And update the org icon to the parrot head? 19:36
dukeleto whiteknight: i don't know if I can add people, i will check. 19:39
dalek p-rx: 705f966 | moritz++ | t/p6regex/ (8 files):
[p6regex-test] fudge it to be clean; fix a few tests along the way
 19:43
dukeleto has anyone played around with the new github mirror? 19:44
whiteknight what new mirror? 19:45
moritz github.com/parrot/parrot/ 19:46
whiteknight oh, I haven't "played" with it at all, no
19:52 darbelo left
dalek p-rx: dfbe437 | moritz++ | build/Makefile.in:
include p6regex-test in "test" target
 19:55
20:00 whiteknight left 20:01 lucian_ joined 20:04 lucian left 20:05 nwellnhof left
GeJ Bonjour everyone. 20:08
moritz bon soir 20:09
dukeleto it looks like only pmichaud can add people to the parrot team on github 20:13
dukeleto just updated the github mirror as well 20:14
moritz that would be an unfortunate case of low bus number 20:15
dukeleto is very close to having it know when branches are rm'ed from svn and do the same on github, but is busy with work 20:16
GeJ servus moritz.
20:21 icarroll joined
icarroll Where can I find a list of built-in rules or named character classes for nqp-rx? 20:22
20:22 Paul_the_Greek joined
Paul_the_Greek purl,messages 20:23
moritz icarroll: there's a list in the test suite, in the nqp-rx repo in t/p6regex/rx_subrules
icarroll moritz: thank you
moritz there's also perlcabal.org/syn/S05.html#Predefined_Subrules but I'm not sure if nqp-rx implements them 20:24
icarroll k
20:25 Paul_the_Greek left
dalek kudo: 819e2ce | colomon++ | src/core/List.pm:
Remove List.batch method, as it is obsolete.
 20:42
icarroll I'm trying to define an adjacency operator (two terms with only whitespace between them), but I must be doing something wrong 20:53
token infix:sym<apply> { <.ws> <O('%apply, :pasttype<asplode>')> } 20:54
doesn't work
purl Look buddy, doesn't work is a strong statement. Does it sit on the couch all day? Is it making faces at you? Does it want more money? Is it sleeping with your girlfriend? Please be specific!
icarroll meaning it gives a syntax error when I try to parse "say 5" 20:55
any suggestions?
purl "I suggest you don't smoke cigarettes ... and don't touch yourself when you're chillin' at the bus stop."
moritz icarroll: what about token infix:sym<appl> { <.ws> <term> } ?
icarroll I'll try that
moritz note that in Perl, 'say 5' parses the 'say' as listop, and expects a term after that anyway 20:56
icarroll Null PMC access in get_string_keyed()
moritz if the programming language you try to parse doesn't follow the term - op - term pattern, then maybe it's easier if you override the operator precedence parser 20:57
icarroll that sounds rough
mostly it does follow term op term
but function application is adjacency
if the existing operator parser doesn't like that though, I guess I'll have to roll my own 20:58
20:58 kj joined
moritz rakudo: say 5 21:00
p6eval rakudo 689bb7: OUTPUTĀ«5ā¤Ā»
21:01 bubaflub joined
moritz icarroll: it sounds more like you should parse differently after a function name, not like you want a general two-terms-in-a-row rule 21:01
icarroll well, the first item doesn't have to be a name. it could be an expression that returns a function 21:02
moritz ugh
icarroll heh
moritz what is this programming language?
icarroll made up
moritz ok
icarroll this piece of the grammar is similar to haskell though
moritz if you allow two terms in a row, think about whether you'll still get decent parse error messages out of it 21:03
icarroll hadn't thought about error messages
moritz anyway, it might be worth investigating where that null PMC access comes from 21:04
maybe the concept of calling <.ws> <term> in an infix operator isn't flawed, there might just be an inconvenient bug 21:05
anyway, TTFN
icarroll thanks
laters
21:07 contingencyplan left 21:10 pyrimidine left 21:14 kj left 21:18 theory joined 21:21 lucian_ left, ruoso left 21:24 kj joined
cotto_work seen dafrito 21:35
purl dafrito was last seen on #parrot 2 days, 19 hours, 56 minutes and 29 seconds ago, saying: Yeah, he seems to have forgotten your name [Aug 21 01:38:59 2010]
21:42 perlite left, perlite joined 21:52 Paul_the_Greek joined 21:54 tcurtis joined 22:01 tadzik joined 22:03 Andy left 22:04 fperrad left 22:20 kj left 22:22 kid51 joined 22:24 davidfetter joined, icarroll left 22:25 bubaflub left 22:31 jsut joined 22:32 whiteknight joined, jsut_ left 22:35 tcurtis left 22:45 senf_statt_oel joined 22:48 Paul_the_Greek left 22:54 dngor_ joined 22:55 dngor left 22:59 davidfetter left
dalek rrot: r48616 | jkeenan++ | trunk/tools/util/ncidef2pasm.pl:
Improve documentation.
 23:13
23:14 dngor_ is now known as dngor 23:42 TonyC left, nopaste left
dalek rrot-linear-algebra: 2233f07 | Whiteknight++ | .gitignore:
update .gitignore to stop messing with my library files
 23:43
rrot-linear-algebra: c596ed3 | Whiteknight++ | src/lib/mat (2 files):
add the two library files that I thought I added last time
23:49 ruoso joined
cotto_work quiet day today 23:54
23:57 kid51 is now known as kid51_at_dinner

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!