Parrot 0.6.0 "P&P" released | Please mentor for SoC | parrotcode.org/ | YAPC::NA talks deadline is Mar 31 | tinyurl.com/2pmnlq
Set by moderator on 18 March 2008.
CrawfordComeaux Anyone around that might be able to answer some gsoc questions? 00:04
chromatic What would you like to know? 00:06
CrawfordComeaux What concerning perl6 implementation is needed and who would the mentor be? 00:07
chromatic The mentor would likely be particle.
Implementation-wise, I'm sure you could pick just about anything you want.
Do you have a particular interest?
www.perlfoundation.org/perl5/index....t_projects 00:08
shorten chromatic's url is at xrl.us/bhv9n
chromatic If you have a particular Perl 6 feature in mind, you could implement that. That's a fairly nice self-contained step actually. 00:20
Eevee is there a concise list of what's done and what's not? 00:22
or would that be the test suite
CrawfordComeaux ...beat me to the question
Limbic_Region getting rakudo to past more of spectest would be really cool from my perspective 00:23
chromatic The tests that pass are done, and what's left isn't done. 00:24
Not much of an answer, but the line changes every couple of days.
Eevee fair enough
Limbic_Region s/past/pass/ 00:25
00:25 jrockway_ joined
svnbotl r26486 | chromatic++ | trunk: 00:27
: [Lua] Added TODO to Parrot::Test::Lua so that Lua can have TODO tests.
diff: parrotvm.org/svn/parrot/revision/?rev=26486
CrawfordComeaux How can I go about finding out which tests pass and which don't? 00:29
chromatic If you have a Parrot download, build it, then go to languages/perl6 and run make spectest. 00:30
00:32 jrockway joined
svnbotl r26487 | chromatic++ | trunk: 00:32
: [Lua] Marked the one remaining failing Lua test (timed alarms) as TODO, because
: Parrot needs a little bit of work for Lua to pass this test correctly. See RT
: #51870.
diff: parrotvm.org/svn/parrot/revision/?rev=26487
CrawfordComeaux Thanks, chromatic 00:34
chromatic You're welcome. When/if particle shows up, he's the likely mentor. 00:35
Eevee holy. this is a lot of spec tests 00:36
00:37 Ademan joined
Infinoid I believe the official definition of "perl 6" is "any implementation that passes all the tests", so they try pretty hard 00:49
Eevee well, 70% pass! 00:50
unless mass numbers of them are todo'd 00:51
Infinoid awesome
Eevee but given the gsoc list I take it the tests are far from comprehensive 00:53
diakopter Larry estimates the test suite is about 1/3 complete
the one in the pugs repository
Coke particle is going to be in and out over the next 2 weeks.
diakopter of which rakudo pulls/uses/smokes a little of 00:54
Eevee ouch
diakopter don't let that dismay you. Eevee: what's your language implementation fancy? parser/grammars? object models? test suite completeness? 00:56
00:56 jrockway joined
Infinoid allright. I'm not a big fan of the "no cuddled elses" rule, but pdd07 is nonspecific about language, so I've expanded the test to cover perl files as well. 00:57
Eevee I'm not entirely sure yet; haven't really gotten into language design/compiler writing before 00:59
Infinoid Eevee: its funner than it sounds :)
Eevee if it didn't sound fun I wouldn't have come in here 8) 01:01
spinclad Parrot 0.6.0 'P&P': powers and principalities. 01:05
svnbotl r26488 | infinoid++ | trunk: 01:06
: [t/codingstd] c_cuddled_else.t can apply to perl sources too (pdd07 isn't specific)
: - rename c_cuddled_else.t to cuddled_else.t, it isn't just for C any more
: - add perl language files to the list
: - fix up all the failures (16 files)
Infinoid dinner time, bye! 01:07
davidfetter bon appetit, Infinoid 01:08
01:10 Theory joined
Coke Infinoid: there *is* a perlcritic test for that, I think. 01:30
01:34 AndyA joined 01:57 kid51 joined, jrockway joined 02:01 CrawfordComeaux joined 02:04 contingencyplan joined 02:06 Andy joined 02:34 Ademan joined 02:56 Eevee joined 02:58 kid51 joined
Infinoid Coke: for cuddled elses? nice. still had lots to fix up tho 03:16
03:20 particle joined 03:21 jdv79_ left
Coke curses lego star wars complete saga for having a %-done meter. 03:59
Eevee someone figured out that a lot of gamers have mild OCD 04:11
smash bros brawl has this giant wall of things to unlock
so now I can't stop playing until I have defeated it
04:17 petdance joined
petdance evening 04:23
04:43 contingencyplan joined
Coke hello, petdance 04:43
petdance darn IRC client at work
Coke ... you're at work? eesh
svnbotl r26489 | coke++ | trunk: 04:46
: [tcl]
: - Fix several of the diagnostic messages that have changed since we
: implemented them.
: - Fix [lsort -dictionary] so that items that differ only by case sort properly.
Coke starts a gcdebug runcore test of languages/tcl and goes to sleep. *maybe* it'll be done when he wakes up. 05:00
petdance no, the IRC client at work is Andy 05:19
05:32 CrawfordComeaux joined
Tene Okay, what do I actually need to do to mentor for gsoc? 06:16
06:19 kraai joined
chromatic Sign up to the list, I believe. 06:55
07:05 petdance joined, slightlyoff joined
svnbotl r26490 | chromatic++ | trunk: 07:09
: [lib] Changed c_cuddled_else.t to cuddled_else.t in list of default tests
: (fallout from r26488).
diff: parrotvm.org/svn/parrot/revision/?rev=26490
07:09 uniejo joined
Tene chromatic: which list? 07:10
purl it has been said that which list is that?
Tene purl: forget which list
purl Tene: I forgot which list
chromatic Let me see. 07:11
07:11 jrockway_ joined
chromatic Hm, looks like you need an invitation from the project administrator. 07:13
ewilhelm, ping
Or maybe just code.google.com/soc/2008/mentor_step1.html 07:14
07:21 jrockway_ joined
svnbotl r26491 | chromatic++ | trunk: 07:26
: [src] Changed CONTEXT(interp->ctx) to CONTEXT(interp), which seems clearer and
: fulfills RT #41881. If we need to rearrange how we store contexts in the
: interpreter, we can rearrange it in only the macro now.
diff: parrotvm.org/svn/parrot/revision/?rev=26491
Tene chromatic++ # will look after sleep 07:43
08:10 Coke joined
ewilhelm Tene: you want to know what *work* you have to do to mentor or what paperwork? 08:11
08:14 toft joined 08:29 AndyA joined 09:31 IllvilJa joined 09:39 kj joined 10:44 avar joined 10:47 ruoso joined 12:30 jrockway_ joined 12:37 orbiscerbus joined
svnbotl r26492 | coke++ | trunk: 12:47
: [codingstd]
: Resolve RT#51880, Patch Courtesy Matt Kraai
diff: parrotvm.org/svn/parrot/revision/?rev=26492
12:55 orbiscerbus left 13:07 gryphon joined 13:17 skids joined
Coke hurm. I would like to create a stub parrot Sub that when invoked, goes out to disk, reads in some HLL, compiles it to PIR, replaces the existing method, and then basically tailcalls to the new version. 13:21
Ah. If instead of trying to tailcall to the replaced version, I do so to the *just compiled* version, that works. And then the next time it just invokes the new one. 13:23
er, not the just compiled version. find_name('sub name') seems to get me a hook to the right version when I need it. 13:24
13:43 particle joined 13:59 kraai left 14:41 b2gills joined 15:07 sjansen joined 15:16 Theory joined 15:25 uniejo joined 16:16 davidfetter joined 16:20 Dave joined 16:21 jrockway joined
Tene ewilhelm: both 16:47
Coke (unified testing) it would be nice if we had a way to have the default test for prove t/ be the same regardless of whether you ran "prove t" or "make test" (esp. if t is some sub-sub directory) 16:51
ewilhelm just pass make args to prove? 16:52
16:55 ambs joined
Coke "how" ? 17:04
purl "how" is different for different languages
Coke basically, I want {{ "make test" but only those tests in t/codingstd }}
Infinoid ...which are a subset of all the .t files in t/codingstd, right? 17:05
ewilhelm this is where xt/ comes in handy 17:06
svnbotl r26493 | coke++ | trunk:
: [tcl]
: Don't *always* compile [parray] when running tclsh; only do so if
: parray is ever invoked; if it is, load it, compile it, and have it
: replace the stub. (AUTOLOAD)
diff: parrotvm.org/svn/parrot/revision/?rev=26493
ewilhelm or possibly something of a manifest since you need to have a ./prove anyway 17:07
Coke ewilhelm: I could see that. 17:08
ewilhelm tene, sorry my scrollback is broken - palm &irssi
what is "both"?
(or, how do i pgup? :-) 17:09
Tene /alias su sb goto -15 17:10
/su
;)
Tene also uses irssi on a PalmOS device.
both = "What is involved in being a mentor?" and "How do I sign up to do it?" 17:11
ewilhelm ah... be available and helpful + ask in #soc 17:12
avar You also have to fill out periodic status reports and such
it's all in the gewgle faq
ewilhelm thinks google url is there
y. read the wiki 17:13
basically, make sure success/failure is entirely up yo the student 17:14
17:21 jhorwitz joined
svnbotl r26494 | fperrad++ | trunk: 17:44
: [Lua]
: - add lfs.symlinkattributes()
: - and test
diff: parrotvm.org/svn/parrot/revision/?rev=26494
18:15 Psyche^ joined
Coke acme? 18:24
purl acme is the yapc::Europe guy or known to type 'i am fat' when he steals terminals or mininat or gnat's bitch or orange or mini_gnat or the American Company that Makes Everything or Wile E. Coyote's preferred supplier of goods of all sorts or the Plan 9 emacs-without-the-lisp or kane's bitch or Agent Orange or in lurve (or at least in Amsterdam) or know as Leon Brocard
18:57 uniejo joined
cotto[work] how often does coverity testing happen to parrot? 19:07
19:15 Andy joined
Coke coverity? 19:35
purl coverity is probably a commercial tool for Automated Error Prevention and Source Code analysis, See, www.coverity.com/main.html or it has been used to measure the quality of the LAMP stack and other major source projects
Coke was hoping it would be on the list. 19:36
ISTR it's daily.
some parrot committers have coverity logins to check such things.
cotto[work] thanks 19:39
19:42 lichtkind joined 19:53 slightlyoff joined
TimToady phone in 1 19:59
20:12 marmic_ joined 20:14 barney joined 20:19 b2gills left
Coke Eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeek 20:20
Coke gets his phone. Coke's phone is wedged. Coke reboots his phone... 20:22
Coke waits...
Eevee I used to laugh at the prospect of rebooting a phone 20:29
then I got a Treo
Coke Normally don't have to reboot the iphone, but something wedged Safari... 20:46
svnbotl r26495 | chromatic++ | trunk: 20:56
: [src] Changed Parrot_Context_info struct such that its file member holds a
: STRING *, not a char *. Not only does this simplify code that fills in this
: struct, but it simplifies code that uses the struct, as the latter no longer
: needs to worry about if it needs to free the file member.
21:23 mire joined
svnbotl r26496 | chromatic++ | trunk: 21:33
: [PDB] Minor cleanups and C++ compatibility fixes.
diff: parrotvm.org/svn/parrot/revision/?rev=26496
21:49 Limbic_Region joined 21:52 particle joined
svnbotl r26497 | bernhard++ | trunk: 21:53
: #51380: [PATCH] Code layout of imcc.y
: Try to make code layout of imcc.y more consistent.
diff: parrotvm.org/svn/parrot/revision/?rev=26497
Infinoid wow, upgrading to perl-5.10.0 is a great reminder of the interconnectedness of the universe 21:54
21:55 LimbicRegion joined 22:01 jrockway joined 22:13 kid51 joined, marmic joined 22:27 skids joined 22:30 Theory joined 22:33 DarkWolf84 joined
svnbotl r26498 | jkeenan++ | trunk: 22:48
: 1. Per rt.perl.org/rt3/Ticket/Display.html?id=39823, remove
: tools/dev/run_indent.pl from Parrot distribution.
: 2. Adjust MANIFEST, capturing 1 name change and 1 addition not
: previously caught.
22:54 Pabellon joined
Infinoid admits to failing to update the MANIFEST 22:57
thanks, kid51 :)
kid51 ... and for your punishment, I'm giving you a ticket! "Heh, heh, heh," the evil wizard cackled 23:00
Infinoid woohoo, my very own ticket 23:01
svnbotl r26499 | chromatic++ | trunk: 23:07
: [src] Make PARROT_ASSERT() somewhat cheaper in the default case. It's around a
: 6.5% speedup in some of the Tcl tests, and I suspect that's likely true
: elsewhere.
diff: parrotvm.org/svn/parrot/revision/?rev=26499
r26500 | jkeenan++ | trunk: 23:26
: 1. Set SVN properties on lib/Parrot/Test/Util/Runloop.pm.
: 2. Apply patch to t/pmc/objects.t per
: rt.perl.org/rt3/Ticket/Display.html?id=39196 (revised version of
: patch first submitted by kjs).
diff: parrotvm.org/svn/parrot/revision/?rev=26500
23:27 cout joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!