Camelia, the Perl 6 bug

IRC log for #parrot, 2011-04-21

Parrot | source cross referenced

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:08 cotto_work whiteknight: you can look at the dir structure of previous version in docs.parrot.org
00:09 whiteknight yeah, that's what I'm doing
00:09 cotto_work can you get into parrotvm fine?
00:09 whiteknight if this were better documented, we wouldn't need any forensics
00:09 whiteknight yes
00:09 whiteknight thanks
00:09 cotto_work great
00:09 ShaneC left #parrot
00:09 cotto_work It's nice to not be the only one who can do that.
00:15 kid51 is now known as kid51_at_dinner
00:17 whiteknight what kinds of permissions do I have on this box?
00:18 cotto_work You should have enough to update docs.parrot.org and not too much more
00:18 cotto_work you can look at the trac config and probably lots of other stuff
00:18 whiteknight I don't have the ability to update the symlinks for latest and supported
00:19 whiteknight it doesn't look like I'm a member of the parrot group
00:19 davidfetter_ joined #parrot
00:20 cotto_work You should be able to do that.  Can you delete and create the symlinks?
00:21 whiteknight yes, I can
00:21 whiteknight and did
00:22 davidfetter left #parrot
00:23 davidfetter_ is now known as davidfetter
00:23 dmalcolm left #parrot
00:25 cotto_work ok
00:27 theory left #parrot
00:38 davidfetter left #parrot
00:42 * cotto_work goes home
00:47 soh_cah_toa what's the deal w/ the pbc_to_exe error? rohit_nsit08 had it before and so did i after a 'git pull'. what's causing this?
00:53 kid51_at_dinner is now known as kid51
00:55 jtpalmer joined #parrot
01:09 whiteknight what pbc_to_exe error?
01:12 soh_cah_toa whiteknight: when building parrot 3.3.0
01:13 soh_cah_toa whiteknight: remember? you told him to 'make realclean' and rebuild
01:18 whiteknight oh, did you make realclean and reconfigure?
01:18 soh_cah_toa whiteknight: oh yeah, i got it to build. i'm just curious as to what's causing this b/c before it sounded like this error was happening quite often
01:25 whiteknight we can look at it later. I'm heading to bed now
01:25 soh_cah_toa whiteknight: sure, see you later
01:25 whiteknight left #parrot
01:27 Eduardow left #parrot
01:32 Eduardow joined #parrot
02:04 dalek TT #2097 closed by jkeenan++: darwin/ppc build failure once gms is default GC
02:04 dalek TT #2097: http://trac.parrot.org/parrot/ticket/2097
02:24 kid51 left #parrot
02:38 theory joined #parrot
02:54 ShaneC joined #parrot
02:55 ShaneC left #parrot
02:58 jrtayloriv joined #parrot
02:58 woosley joined #parrot
03:04 cotto ~
03:05 hudnix left #parrot
03:17 bubaflub left #parrot
03:27 jrt4 joined #parrot
03:31 jrtayloriv left #parrot
03:31 soh_cah_toa left #parrot
04:49 benabik joined #parrot
04:50 wagle left #parrot
04:57 wagle joined #parrot
05:19 fperrad joined #parrot
05:58 contingencyplan left #parrot
06:00 theory left #parrot
06:45 UltraDM joined #parrot
07:22 mikehh apport pkg is still failing
07:22 mj41 joined #parrot
07:22 mikehh bah - wrong channel
08:13 wagle left #parrot
08:24 wagle joined #parrot
08:35 spinclad left #parrot
08:39 wagle left #parrot
08:41 spinclad joined #parrot
08:49 wagle joined #parrot
08:52 mikehh left #parrot
09:26 jrt4__ joined #parrot
09:26 birdwindupbird joined #parrot
09:30 jrt4 left #parrot
09:35 jrt4 joined #parrot
09:39 jrt4__ left #parrot
09:40 dalek nqp/ctmo: 0fbd19a | moritz++ | src/ops/nqp.ops:
09:40 dalek nqp/ctmo: track a parrot function rename.
09:40 dalek nqp/ctmo:
09:40 dalek nqp/ctmo: Enables building nqp on newer parrots
09:40 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/0fbd19a5b9
09:44 jrt4__ joined #parrot
09:48 jrt4 left #parrot
09:51 birdwindupbird left #parrot
09:51 jrt4__ left #parrot
09:52 jrt4__ joined #parrot
09:55 kuku joined #parrot
09:58 dalek nqp/ctmo: 8b1d3e1 | moritz++ | src/NQP/Actions.pm:
09:58 dalek nqp/ctmo: remove two Q:PIR blocks from NQP::Actions
09:58 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/8b1d3e1c4d
10:09 dalek nqp/ctmo: f837f1c | jonathan++ | src/NQP/Actions.pm:
10:09 dalek nqp/ctmo: Should have a $?PACKAGE in the initial scope. Fixes 09-var.t regression.
10:09 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/f837f1c7b4
10:09 dalek nqp/ctmo: 42601fe | jonathan++ | src/NQP/Actions.pm:
10:09 dalek nqp/ctmo: Fix a couple of bugs that broke 33-init.t.
10:09 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/42601febaa
10:09 woosley left #parrot
10:13 moritz nqp: say(1 // 2);
10:13 p6eval nqp: OUTPUT«1␤»
10:29 mtk left #parrot
10:29 dalek nqp/ctmo: 4d845cd | moritz++ | src/NQP/Grammar.pm:
10:29 dalek nqp/ctmo: be a bit more idiomatic in NQP::Grammar; no functional changes
10:29 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/4d845cdd75
10:29 dalek nqp/ctmo: 57fa673 | moritz++ | src/HLL/Compiler.pm:
10:29 dalek nqp/ctmo: rewrite version string initialization with less PIR
10:29 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/57fa67384c
10:35 mtk joined #parrot
10:43 dalek nqp/ctmo: 5a5392e | jonathan++ | src/NQP/Actions.pm:
10:43 dalek nqp/ctmo: Install (what will be the) latest merged GLOBAL view. This unbreaks some uses of $GLOBAL::foo (and thus some of 43-package-var.t).
10:43 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/5a5392ea33
10:43 dalek nqp/ctmo: 51ea2f3 | jonathan++ | src/NQP/ (2 files):
10:43 dalek nqp/ctmo: Make sure we can always see the current $*GLOBALish.
10:43 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/51ea2f3698
10:51 bacek http://www.google.com/search?​q=judgement+day+april+21+2011
10:51 bacek good night, humans
10:55 jrt4__ left #parrot
11:01 woosley joined #parrot
11:19 oopus joined #parrot
11:21 oopus left #parrot
11:36 lucian joined #parrot
11:44 kuku is now known as birdwindupbird
11:56 Patterner left #parrot
11:57 Psyche^ joined #parrot
11:57 Psyche^ is now known as Patterner
12:02 Khisanth left #parrot
12:11 lucian aloha: tell whiteknight this is the level of granularity i was thinking of when we last talk about concurrency http://packages.python.org/greenlet/
12:11 aloha lucian: Okay.
12:11 * lucian waves hello
12:35 Coke left #parrot
12:35 Coke joined #parrot
12:37 dalek nqp/ctmo: 46e3495 | jonathan++ | src/NQP/Actions.pm:
12:37 dalek nqp/ctmo: A little refactor to the symbol lookup code to make it neater.
12:37 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/46e3495542
12:37 dalek nqp/ctmo: 3008506 | jonathan++ | src/ (2 files):
12:37 dalek nqp/ctmo: First cut of package auto-viv using the new package scheme. Enabled just for variable accesses in lexical packages for now.
12:37 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/3008506063
12:51 Coke left #parrot
12:51 Coke joined #parrot
13:04 whiteknight joined #parrot
13:05 whiteknight left #parrot
13:05 whiteknight joined #parrot
13:06 whiteknight good morning, #parrot
13:18 contingencyplan joined #parrot
13:19 Coke left #parrot
13:22 bubaflub joined #parrot
13:25 Coke joined #parrot
13:28 dalek nqp/ctmo: c789496 | jonathan++ | src/ (3 files):
13:28 dalek nqp/ctmo: Install our-scoped classes at compile time, and make sure the installation is reinstated during deserialization. An our-scoped package inside a lexically scoped package now ends up installed in the right way. However, nothing does lookups of anything other than variables this way yet...
13:28 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/c789496deb
13:28 dalek nqp/ctmo: 9708e54 | jonathan++ | src/Regex (2 files):
13:28 dalek nqp/ctmo: Set up a fake GLOBALish in the PIR Regex library.
13:28 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/9708e544e1
13:28 dalek nqp/ctmo: 7756ef8 | jonathan++ | src/stage0/ (7 files):
13:28 dalek nqp/ctmo: Update the bootstrap with the new package installation logic. Seems to work out OK.
13:29 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/7756ef8933
13:30 lucian whiteknight: aloha seems to have amnesia. http://packages.python.org/greenlet/
13:31 lucian this is the sort of granularity i think is necessary (from an API p.o.v.) to simulate mostly any concurrency
13:37 lucian_ joined #parrot
13:38 whiteknight good morning lucian
13:39 lucian_ 'morning
13:39 lucian_ well, good afternoon
13:39 lucian_ lucian is my ghost since my ISP kicked me in the teeth
13:41 lucian left #parrot
13:41 whiteknight ISPs are the best, aren't they?
13:41 lucian_ the best at kicking people in the teeth? sure
13:42 whiteknight right, that thing you said
13:42 tadzik or not delivering their services for 3 months and threatening you with courts for not paying .30 PLN another 3 months after breaking the agreement? Hell yeah
13:44 lucian_ is now known as lucian
13:44 whiteknight I don't understand how firefox can be eating up so much of my RAM and CPI
13:44 whiteknight CPU
13:45 lucian whiteknight: don't worry, it's normal
13:45 whiteknight it's absurd that simply being open for any length of time slowly eats up all resources
13:45 lucian its GCs are very leaky, for one
13:45 lucian and it caches things it really shouldn't
13:46 lucian XPCOM and XUL are to blame for much of it
13:46 whiteknight I don't know much about either of them
13:46 whiteknight I use chatzilla, so eventually if I want to use my computer again for real productivity stuff, i'm going to have to close that too
13:47 lucian i know much more than i'd like too about both ...
13:47 lucian (Browse, Sugar's browser, embeds xulrunner with pyxpcom & hulahop. dreadful)
13:47 whiteknight actually, that's happening now. brb
13:47 whiteknight left #parrot
13:49 whiteknight joined #parrot
13:49 atrodo whiteknight> I solved that problem: http://chatzilla.rdmsoft.com/xulrunner/
13:49 whiteknight ah, so this is what it feels like with an extra gig of free RAM, and 25% of my CPU not tied up
13:49 atrodo well, i found a solution to that problem
13:50 whiteknight it's like I've never breathed before
13:50 hudnix joined #parrot
13:51 atrodo Then again, I use App::Bondage to always be on
13:52 lucian atrodo: how intensive is chatzilla on its own for an IRC client?
13:53 atrodo lucian> Not at all.  It's at about 75M right now, really doesn't bloat a whole lot
13:54 atrodo I think i've only seen it get really large once or twice, but those seemed like odd cases
13:58 lucian i see
14:05 whiteknight I only use chatzilla at work because it's easy to install
14:05 whiteknight and I only use one chatroom, don't do any DCC crap, etc
14:07 lucian whiteknight: you might try pidgin portable or something
14:08 bubaflub on Mac I prefer Colloquy
14:09 * lucian uses LimeChat on osx
14:10 lucian oh, this too http://portableapps.com/ap​ps/internet/pchat_portable
14:10 whiteknight lucian: I don't have any problem with Chatzilla, I'm just going to stick with that here
14:11 lucian well, your problem is resource munching
14:11 * atrodo likes using telnet for irc and im and http and ssh
14:12 atrodo that last one gets tricky somedays
14:13 whiteknight I have to keep firefox open for very long periods of time for it to eat that many resources
14:13 whiteknight I've had it open since Monday morning
14:14 lucian heh. are you in a position to switch browsers?
14:14 atrodo I usually have to restart it about every day, but then again, I do web dev, so I stress it a little more than most
14:14 whiteknight I'm annoyed that it grows without bounds, though the rate of growth is not prohibitive
14:14 whiteknight atrodo: I do web dev too
14:14 atrodo Oh, neat, didn't know that
14:14 whiteknight although I open other instances of firefox for intensive testing
14:15 lucian i tend to use chrome pretty much exclusively these days
14:15 lucian rarely do i need firefox
14:15 whiteknight lucian: I'm not interested in switching browsers, not yet anyway
14:15 atrodo I have come to the conclusion that firefox doens't like javascript that misbehaves.  Those expiremental days seem to make it grow more than normal
14:16 lucian atrodo: its JS gc is among the crappier, yes
14:16 atrodo Chrome has some weird bugs, plus it's tools are not as good as firebug
14:16 atrodo although, i think i saw firebug has a chrome version now
14:17 atrodo lucian> The problems don't seem to be GC so much as bad javascript (exceptions)
14:24 lucian atrodo: hmm. exceptions objects need to be collected too, i guess
14:24 lucian well, i just know its various GCs are bad, from being bitten by them
14:27 Andy_ joined #parrot
14:49 davidfetter joined #parrot
14:49 cotto ~~
14:49 theory joined #parrot
14:50 Andy_ left #parrot
14:50 Andy_ joined #parrot
14:54 Andy_ left #parrot
14:54 Andy_ joined #parrot
15:13 woosley left #parrot
15:16 UltraDM left #parrot
15:16 fperrad left #parrot
15:21 dukeleto ~~
15:22 bubaflub morning dukeleto
15:22 dukeleto bubaflub: how goes it?
15:22 bubaflub dukeleto: not bad; just workin' and what not.  how are you?
15:24 dukeleto bubaflub: just woke up to a sunny day in Portland
15:26 davidfetter dukeleto, wow!
15:26 * davidfetter woke up to a sunny day in oakland, but this is not so unusual ;)
15:26 dukeleto davidfetter: indeed
15:26 davidfetter well, it wasn't quite sunny when i woke up, but it is now
15:29 dukeleto davidfetter: i hope to learn about PGXN later today and do some PL/Parrot hacking with theory
15:30 davidfetter dukeleto, PGXN and EXTENSIONs in general, are pretty awesome
15:30 davidfetter PGXN took some important lessons from CPAN :)
15:32 redicaps joined #parrot
15:32 redicaps left #parrot
15:32 dukeleto interesting language based on ruby-llvm : https://github.com/rferraz/spell
15:34 redicaps joined #parrot
15:34 redicaps left #parrot
15:35 redicaps joined #parrot
15:38 mj41 left #parrot
15:51 redicaps left #parrot
15:56 ambs joined #parrot
15:57 theory dukeleto: Right on.
16:04 mikehh joined #parrot
16:14 dalek nqp/ctmo: cc282da | jonathan++ | src/ (3 files):
16:14 dalek nqp/ctmo: Move symbol installation code into the grammar; needs to happen earlier than it's currently doing, but also this'll help to better parse without doing the actions too.
16:14 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/cc282da67d
16:14 dalek nqp/ctmo: de0e355 | jonathan++ | src/metamodel/how/NQPClassHOW.pm:
16:14 dalek nqp/ctmo: Add a way to set the default parent class in NQPClassHOW, to try and help resolve a circular dependency.
16:14 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/de0e355080
16:14 dalek nqp/ctmo: 4371629 | jonathan++ | src/NQP/Actions.pm:
16:14 dalek nqp/ctmo: Use set_default_parent from actions.
16:14 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/4371629a87
16:14 dalek nqp/ctmo: 7e325a4 | jonathan++ | src/Regex/Cursor.pir:
16:14 dalek nqp/ctmo: Cursor calls set_default_parent.
16:14 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/7e325a4f76
16:14 dalek nqp/ctmo: dc42c1e | jonathan++ | src/stage0/ (7 files):
16:14 dalek nqp/ctmo: Update bootstrap.
16:14 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/dc42c1e881
16:14 dalek nqp/ctmo: a4e9240 | jonathan++ | src/Regex/Cursor.pir:
16:14 dalek nqp/ctmo: Cursor locates NQPMu lexically in the setting.
16:14 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/a4e9240567
16:18 dod left #parrot
16:26 jnthn__ Anyway know what happens if you supply multiple --library= arguments to Parrot?
16:26 jnthn__ Will it use the order that they were passed as its search order?
16:27 dukeleto jnthn__: good question
16:27 dukeleto jnthn__: i doubt we have a test for that
16:27 whiteknight I suspect they stack
16:28 moritz is there an easy way to print the library search path?
16:28 dmalcolm joined #parrot
16:30 jnthn__ My situation is that I have a bootstrapping compiler. At the start of each stage it compiles a "setting" (standard library, essentially)
16:31 cotto_work ~~
16:31 jnthn__ From that point on I want it to use the previous stage in the bootstrap for most things, but find the "setting" from a different place
16:33 whiteknight moritz: There probably is a way to get the array of search paths from the interp. I would need to look at it
16:33 jnthn__ Maybe stacking library paths ain't really the way to do it though, now I think more about it.
16:33 whiteknight they are definitely stored in the iglobals array
16:33 jnthn__ Hmm
16:33 * jnthn__ takes a little break from the hacking to work out what he really needs
16:33 dalek nqp/ctmo: e65afd7 | jonathan++ | src/HLL/Actions.pm:
16:33 dalek nqp/ctmo: Make actual values of symbols in the setting available to the compiler. Will be useful for all sorts of stuff. :-)
16:33 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/e65afd7b4e
16:40 whiteknight look in frontends/parrot/main.c:parseflags or whatever it's called to see, but I suspect it's stacked
16:47 whiteknight left #parrot
16:58 birdwindupbird left #parrot
17:09 whiteknight joined #parrot
17:10 ShaneC joined #parrot
17:11 Eduardow left #parrot
17:14 cotto_work php--
17:16 Eduardow joined #parrot
17:17 whiteknight what's wrong with it?
17:20 plobsing_ left #parrot
17:20 plobsing joined #parrot
17:22 whiteknight VisualBasic--
17:22 cotto_work its nature
17:23 dukeleto cotto_work: no comment ;)
17:23 cotto_work a self-hosted php compiler will be an interesting endeavor
17:29 ShaneC left #parrot
17:30 dukeleto interesting is a word
17:46 plobsing left #parrot
17:50 ShaneC joined #parrot
17:50 ShaneC left #parrot
18:02 whiteknight I don't think a self-hosted PHP compiler would be too big a problem.
18:05 bubaflub whiteknight: are most of the PHP functions written in PHP or are they C?  i think fleshing out the core library might be difficult
18:11 cotto_work The compiler doesn't scare me too much.  The massive mess of builtin functions will be a lot of work.
18:12 bubaflub cotto_work: precisely my point - the language and syntax are fairly straightforward, but the thousand of global functions might not be so easy to handle
18:13 whiteknight yeah, that's definitely the hard part
18:13 whiteknight but then again, if we have PHP syntax we can avoid the huge mess of builtins and provide our own awesome runtime
18:13 whiteknight with double awesome
18:14 whiteknight or we can have a "compatability mode" to add in a more familar runtime instead
18:16 mtk left #parrot
18:19 ambs left #parrot
18:19 ambs joined #parrot
18:19 ambs left #parrot
18:20 Khisanth joined #parrot
18:22 mtk joined #parrot
18:31 mj41 joined #parrot
18:58 dukeleto whiteknight: blarg. Org admins have to make hard decisions.
19:01 jrt4__ joined #parrot
19:08 dukeleto jrt4__: welcome
19:14 whiteknight dukeleto: he's been trying extremely hard to find a mentor, apparently
19:18 dukeleto whiteknight: i just asked in #gsoc, too
19:18 dukeleto whiteknight: i am conflicted
19:18 dukeleto whiteknight: i don't want to set up a student to fail, and not having a mentor is kind of like that.
19:22 cotto_work Sometimes it just doesn't work.
19:23 jnthn__ https://gist.github.com/935285 # this patch appears to be a simplification. In theory it should be a tiny bit faster, and break nothing. It also removes a blocker that's in my way with nqp...
19:24 theory Everything is faster in me.
19:25 jnthn__ :P
19:32 benabik jnthn__: I haven't (yet) dug into the guts of PAST::Compiler, but it seems like that would completely change what PAST::Block does if given a namespace.  Is the current behavior wrong?
19:33 jnthn__ benabik: I'd argue that it is.
19:34 dukeleto in theory, practice and theory are the same, but in practice, they are different.
19:34 * dukeleto just wanted to bother theory
19:34 benabik dukeleto: Don't you mean practice?
19:34 * theory kicks dukeleto
19:35 hercynium joined #parrot
19:35 dukeleto jnthn__: does the parrot test suite pass with that patch? what behavior does it change?
19:36 * whiteknight suspects "theory" only picked that username for the lulz
19:36 dukeleto jnthn__: we can apply it, as long as it doesn't subtley break other HLL's
19:36 theory whiteknight: No, that was a pleasant side-effect.
19:37 jnthn__ dukeleto: testing at the moment...
19:37 dukeleto jnthn__: coolio
19:37 dukeleto jnthn__: i would recommend "make fulltest" to make sure
19:38 jnthn__ *nod*
19:38 jnthn__ tbh it looks like it was some kind of workaround.
19:38 jnthn__ Maybe for another issue that is long gone.
19:39 whiteknight if the patch provides behavior which is more correctish than what we have now, and causes few failures, I'm for it
19:39 whiteknight what's the blocker?
19:40 jnthn__ whiteknight: The inconsistency the patch fixes is that it normally lets you identify a sub outside of a namespace by subid, which is going to be unique.
19:40 jnthn__ But inside of a namespace it doesn't look it up by subid.
19:40 whiteknight oh. I wonder why it wouldn't use subid inside the ns? That seems weird
19:40 jnthn__ Right, to me too.
19:41 jnthn__ I mean, imagine you wanted to use this to identify a multi candidate uniquely.
19:41 jnthn__ It's odd that if the multi lived in a namespace, this'd break.
19:41 PerlJam maybe there was a time when subids weren't globally unique?
19:41 * moritz fulltests and rakudo tests
19:42 jnthn__ PerlJam: But that could equally apply to things outside of a namespace too...
19:42 whiteknight okay, I'm sold. Apply the damn patch already!
19:42 whiteknight full speed ahead and damn the torpedos!
19:43 * PerlJam is still waiting for rakudo to finish compiling.
19:43 jnthn__ Same ;)
19:43 jnthn__ It fixes my nqp issue and passes make test in Parrot.
19:43 PerlJam src/gen/core.pm compilation seems to be taking even longer than normal
19:43 jnthn__ :/
19:44 jnthn__ It completed here.
19:44 jnthn__ I think it just...takes a long time.
19:44 PerlJam aye, but I just compiled it for the release and it didn't seem to take this long
19:44 PerlJam oh!  there it finished
19:45 PerlJam maybe my subjective time is off because I'm watching it too closely  :)
19:45 whiteknight it's like mother always said: a watched rakudo never builds
19:46 PerlJam exactly
19:46 moritz #   Failed test 'Parrot_PMC_delete_keyed_str'
19:46 moritz #   at t/src/extend_vtable.t line 343.
19:46 moritz # Exited with error code: [SIGNAL 11]
19:46 whiteknight blah, that test always fails
19:46 cotto_work lta
19:46 benabik whiteknight: It stopped failing for a couple of days for me...  But then it started again.
19:47 whiteknight we should put in an assertion to prove that extend_vtable.t fails. Otherwise, something might be wrong
19:47 * moritz recently fixed that one spectest that often failed
19:47 cotto_work I don't see the failure.
19:48 whiteknight cotto_work: maybe your setup is borked
19:48 moritz cotto_work: do you configure with --gc=gms ?
19:48 cotto_work I'm running on a make-believe computer.
19:48 benabik jnthn__: It's adding a reference to the sub being built to an enclosing sub?  And you're changing it from searching for a global to using the constant table?  (Trying to understand the code for GSoC.)
19:48 jnthn__ benabik: Yes
19:49 cotto_work moritz: it should be running gms.  that's the default post 3.3
19:49 moritz \o/
19:52 cotto_work cute hack in that test
19:53 whiteknight moritz: yeah, changing the default GC was the first thing we did post-release
19:53 benabik jnthn__: The old way seems to ignore subid, but now it seems to ignore namespace...  Arg, maybe I should leave trying to understand this for when I don't have school projects.
19:53 cotto_work That test shouldn't work.  It uses an uninitialized string.
19:53 whiteknight now that we have GMS as the default and everybody is using it, we can get down to improving and optimizing it
19:53 jnthn__ benabik: subid is more specific than namespace+subname.
19:54 jnthn__ benabik: That is, at least as unique
19:56 benabik jnthn__: POST actually deals with the namespace and such, this is just a reference...  Okay, I think I get it.  Thanks.
19:56 whiteknight cotto_work: I want to get together sometime soon and start seriously talking about concurrency
19:56 whiteknight cotto_work: I've got some ideas on what the system could look like that I'm writing up, and want to start seriously bouncing them off people
19:58 cotto_work whiteknight: excellent
19:58 cotto_work Things bounce off of me all the time.
19:59 cotto_work meeting time
19:59 whiteknight what, right now?
20:02 dukeleto moritz: you can ignore the extend_vtable tests for now
20:03 dukeleto it doesn't "always fail", it fails on some combinations of OS's and compilers right now because I merged after the 3.3.0 release and haven't made it pass everywhere yet
20:03 dukeleto benabik: because I added more tests :)
20:03 benabik dukeleto: More tests good.  Failing tests bad.  I'm torn.  ;-)
20:04 dukeleto benabik: they are tests which compile C programs, so they are sensitive to OS and compiler. They all pass on my machine :)
20:04 dukeleto benabik: merging to master is the quickest way to shake out the bugs
20:05 dukeleto and I like breaking master
20:05 * dukeleto is a bad bad boy
20:08 moritz jnthn__: all rakudo spectests pass with your patch
20:08 jnthn__ zo/
20:08 jnthn__ moritz++, PerlJam++
20:09 whiteknight commit already!
20:10 dalek parrot: 3c7fb39 | jonathan++ | compilers/pct/src/PAST/Compiler.pir:
20:10 dalek parrot: Make block references have consistent semantics (always use subid), whether they live in a package or not.
20:10 dalek parrot: review: https://github.com/parrot/parrot/commit/3c7fb39da0
20:23 Caelum joined #parrot
20:24 Caelum anyone seen alester?
20:24 Andy_ Yup
20:24 * Andy_ waves
20:24 Caelum hi :)
20:24 Caelum I'm writing a patch for vim-perl
20:24 Andy_ ok
20:25 Andy_ please write a ticket first
20:25 Andy_ so we can tie the patch or pull request to a ticket.
20:25 Caelum the issue is that "sub foo # comment \n {" does not highlight "# comment" properly
20:25 Andy_ That's fine.  Just need a ticket.
20:25 Andy_ So I can track history.
20:26 Caelum where should the ticket go, github issue?
20:26 Andy_ yes, please
20:26 Caelum ok, got it
20:26 dalek nqp/ctmo: c46e9f1 | jonathan++ | src/ (2 files):
20:26 dalek nqp/ctmo: Make it possible to specify a path to load the setting from, so we can make sure we get the correct one during the bootstrap procedure.
20:26 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/c46e9f1fe0
20:26 dalek nqp/ctmo: 2d19edc | jonathan++ | build/PARROT_REVISION:
20:26 dalek nqp/ctmo: Bump to latest Parrot to get PCT block ref fix.
20:26 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/2d19edc2f8
20:26 dalek nqp/ctmo: 35c39f0 | jonathan++ | build/Makefile.in:
20:26 dalek nqp/ctmo: Use --setting-path flag to make sure we get the correct version of the setting during the bootstrap process.
20:26 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/35c39f051f
20:26 Caelum Andy_: how would I write a test for this? At first glance, I don't really understand how t/ works in vim-perl
20:27 Andy_ We don't have anything that tests the highlighting.
20:27 Andy_ I'd love if ew did.
20:27 Caelum ok, I see
20:29 davidfetter http://upload.wikimedia.org/wikipedia/co​mmons/5/5e/Caelum_constellation_map.svg ?
20:30 whiteknight left #parrot
20:34 jrt4__ left #parrot
20:34 jrtayloriv joined #parrot
20:40 dalek nqp/ctmo: 2944b43 | jonathan++ | src/NQP/Actions.pm:
20:40 dalek nqp/ctmo: Add a way to find a lexical that is known at compile time. Use this mechanism for setting the default parent NQPMu. (This is the first use of the mechanism that will let us be able to deal with all of these bits at compile - or in Perlier terms BUILD - time).
20:40 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/2944b43fa7
20:40 dalek nqp/ctmo: 820c806 | jonathan++ | src/metamodel/how/NQPClassHOW.pm:
20:40 dalek nqp/ctmo: Remove a workaround.
20:40 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/820c8067a4
20:40 dalek nqp/ctmo: 2cab3bf | jonathan++ | / (8 files):
20:40 dalek nqp/ctmo: Update the bootstrap with latest changes. Also incorporates a Makefile update, since now the --setting-path is available (and needed) in stage1.
20:40 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/2cab3bfb4f
20:41 soh_cah_toa joined #parrot
20:42 dalek nqp/ctmo: 1d737e4 | jonathan++ | src/core/NQPMu.pm:
20:42 dalek nqp/ctmo: NQPMu is in the setting and should be lexical (and now all blockers to it being are gone, so it can be).
20:42 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/1d737e48a3
20:51 plobsing joined #parrot
20:51 soh_cah_toa left #parrot
20:52 cotto_work dukeleto: the Parrot_PMC_delete_keyed_str test uses an uninitialized string. That might have something to do with the failures.
20:53 cotto_work in t/src/extend_vtable.t
20:56 Andy_ Caelum: Got it, thanks.
20:59 mj41 left #parrot
21:01 soh_cah_toa joined #parrot
21:10 tewk left #parrot
21:14 pranq joined #parrot
21:14 pranq left #parrot
21:18 soh_cah_toa cotto_work: hey
21:21 cotto_work soh_cah_toa: hio
21:21 soh_cah_toa cotto_work: is there anything else in the release process that needs to be automated
21:22 soh_cah_toa cotto_work: i'll nopaste what i have so far
21:23 nopaste "soh_cah_toa" at 192.168.1.3 pasted "Automated Release Manager (So Far...)" (240 lines) at http://nopaste.snit.ch/41370
21:28 cotto_work I didn't know we had release_check
21:29 soh_cah_toa cotto_work: it checks that the tarball builds on its own. whiteknight says it also runs 'make fulltest'
21:29 cotto_work nice
21:30 cotto_work that right there takes out a good chunk of the annoyance of a release
21:30 soh_cah_toa yay
21:30 soh_cah_toa what else could i add? i want to automate as much as possible
21:31 jrtayloriv left #parrot
21:31 cotto_work it doesn't run fulltest though, just test
21:31 soh_cah_toa sure it does
21:31 cotto_work soh_cah_toa: have you read through the release manager guide looking for automatable pieces?
21:31 soh_cah_toa good idea
21:31 jrtayloriv joined #parrot
21:32 cotto_work thanks for working on this
21:32 cotto_work soh_cah_toa++
21:32 soh_cah_toa sure
21:32 tcurtis soh_cah_toa++
21:32 moritz one thing I noticed is that you need to update the version in more than one place
21:32 moritz that could surely be automated somehow
21:32 soh_cah_toa moritz: where else?
21:32 cotto_work It's really nice to have a gsoc student who's not focused exclusively on his project, but is actually looking for other ways to improve parrot.
21:33 cotto_work (though I'm sure that'll change once the program starts in earnest)
21:33 soh_cah_toa cotto_work: well, i'd like to stick around afterwards so i got to learn this stuff sometime
21:34 moritz soh_cah_toa: on re-skimming the guide it seems it's only one location, and once 'make release VERSION=$version'
21:35 tcurtis soh_cah_toa: which project(s) did you submit a proposal for? I've probably heard it mentioned before but I've since forgotten.
21:35 moritz I remembered either wrong, or a much older version
21:35 soh_cah_toa tcurits: the debugger
21:36 soh_cah_toa tcurits: okay, VERSION should be of the form RELEASE_n_n_n? just like git?
21:36 soh_cah_toa i mean moritz, whoops
21:36 cotto_work x.y.z
21:36 soh_cah_toa cotto_work: okay
21:37 cotto_work make release VERSION=a.b.c (in the release manager guide)
21:37 tcurtis soh_cah_toa: very nice.
21:37 soh_cah_toa cotto_work: yeah, it's time i look at that
21:37 cotto_work indeed
21:38 cotto_work whiteknight found a piece that needed rewriting while he was cutting the release
21:39 soh_cah_toa cotto_work: i'll ask him. maybe i can take care of it
21:39 bubaflub soh_cah_toa++
21:39 cotto_work soh_cah_toa: http://irclog.perlgeek.de/p​arrot/2011-04-19#i_3545579
21:42 soh_cah_toa there it is, okay
21:42 * soh_cah_toa must go make some coffee now
21:45 theory left #parrot
21:46 theory joined #parrot
22:01 dalek nqp/ctmo: acf33e3 | jonathan++ | / (69 files):
22:01 dalek nqp/ctmo: A little reorganization. Move the (NQP-specific) meta-objects out of the metamodel directory, and rename it to 6model, to start to better indicate what is in the 6model core and what is using it. No functional changes.
22:01 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/acf33e3696
22:05 lucian left #parrot
22:11 cotto_work jnthn__: what does 'ctmo' stand for?
22:15 jnthn__ cotto_work: compile time meta-objects
22:16 jnthn__ cotto_work: It's easiest to describe as "the set of enabling refactors for gradual typing".
22:18 cotto_work ok.  That makes sense from what you've posted in your blog.
22:20 jnthn__ Yes, finally getting around to actually implementing this stuff.
22:33 Andy_ left #parrot
22:38 bubaflub left #parrot
22:42 Coke left #parrot
22:42 Coke joined #parrot
22:55 hercynium left #parrot
22:56 Coke left #parrot
22:56 Coke joined #parrot
23:04 davidfetter left #parrot
23:06 bubaflub joined #parrot
23:07 Coke left #parrot
23:07 Coke joined #parrot
23:16 Coke left #parrot
23:16 Coke joined #parrot
23:20 dalek nqp/ctmo: 3404565 | jonathan++ | src/HLL/Se (2 files):
23:20 dalek nqp/ctmo: Start to rename SettingManager to ModuleLoader, and move the current (very minimalistic) module loading to there, in anticipation of expanding it with extra pieces soon.
23:20 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/3404565942
23:20 dalek nqp/ctmo: b5bf667 | jonathan++ | / (3 files):
23:20 dalek nqp/ctmo: Start renaming of SettingManager => ModuleManager.
23:20 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/b5bf6670e9
23:20 dalek nqp/ctmo: e3209ee | jonathan++ | src/Regex.pir:
23:20 dalek nqp/ctmo: Update a HLL::SettingManager reference in the REgex library.
23:20 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/e3209ee5e3
23:20 benabik left #parrot
23:31 kid51 joined #parrot
23:32 soh_cah_toa left #parrot
23:35 dalek nqp/ctmo: c179b99 | jonathan++ | src/stage0/ (7 files):
23:35 dalek nqp/ctmo: Update bootstrap to eliminate HLL::SettingManager usage.
23:35 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/c179b99d1f
23:35 dalek nqp/ctmo: 9863370 | jonathan++ | src/ModuleLoader.pm:
23:35 dalek nqp/ctmo: Rip out HLL::SettingManager compat.
23:35 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/9863370a50
23:35 dalek nqp/ctmo: 11717b0 | jonathan++ | build/Makefile.in:
23:35 dalek nqp/ctmo: Build SettingManager.pbc as ModuleLoader.pbc as well, so we can transition.
23:35 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/11717b0767
23:38 janus left #parrot
23:38 janus joined #parrot
23:38 Coke left #parrot
23:38 Coke joined #parrot
23:46 whiteknight joined #parrot
23:48 whiteknight good evening, #parrot
23:52 kid51 is now known as kid51_at_dinner
23:54 dalek nqp/ctmo: 6201d0d | jonathan++ | / (12 files):
23:54 dalek nqp/ctmo: Final cleanup after SettingManager => ModuleLoader changes.
23:54 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/6201d0d19d

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

Parrot | source cross referenced