Camelia, the Perl 6 bug

IRC log for #parrot, 2010-02-15

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 patspam joined #parrot
00:09 dhg joined #parrot
00:10 contingencyplan joined #parrot
01:04 abqar joined #parrot
01:44 kid51 joined #parrot
01:47 dalek rakudo/master: 5383a13 | (Solomon Foster)++ | src/core/Any-list.pm:
01:47 dalek rakudo/master: Protos for sub versions of min and max.
01:47 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/5​383a13d39480a2b4375df38a9b4da3b73703010
01:47 dalek rakudo/master: 542bfa5 | (Solomon Foster)++ | src/core/Str.pm:
01:47 dalek rakudo/master: Cheat version of Str.Num.
01:47 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/5​42bfa5084f3210fc26d8d57660009ebc6535b8b
01:47 dalek rakudo/master: ee93de4 | (Solomon Foster)++ | src/core/Any-num.pm:
01:47 dalek rakudo/master: We were getting an infinite loop in Any.abs, I think possibly because .Num was returning a float instead of a Num.  Work around that by calling the pir abs function directly on the value returned from self.Num.
01:47 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/e​e93de464efe057bb2279b36a35de20bdc4b5133
01:57 dalek parrot: r43958 | plobsing++ | branches/tt362 (2 files):
01:57 dalek parrot: eliminate ImageIO.visit_pmc_now
01:57 dalek parrot: also add svn properties on t/pmc/imageio.t
01:57 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43958/
01:58 jimk joined #parrot
02:14 dalek parrot: r43959 | plobsing++ | branches/tt362/include/parrot/pmc_freeze.h:
02:14 dalek parrot: remove unused structs from header
02:14 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43959/
02:14 hudnix joined #parrot
02:17 kid51 joined #parrot
02:23 dukeleto 'ello
02:23 plobsing yo dukeleto
02:33 dalek rakudo/master: 2eb5f98 | (Solomon Foster)++ | src/core/Any-list.pm:
02:33 dalek rakudo/master: Add simple-minded version of uniq.
02:33 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/2​eb5f98f7e2784aebb3fb502a204d99271a73544
02:44 dalek rakudo/master: 07e96ba | (Solomon Foster)++ | t/spectest.data:
02:44 dalek rakudo/master: Turn on uniq.t.
02:44 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/0​7e96bac1b1e78519b59deb682112796c19035a1
03:03 dalek parrot: r43960 | plobsing++ | branches/tt362 (2 files):
03:03 dalek parrot: eliminate ImageIO.id_list
03:03 dalek parrot: slight modification to t/pmc/imageio.t. re-usability is not a feature of ImageIO, and now it isn't.
03:03 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43960/
03:07 dalek rakudo/master: 3d02e7c | (Solomon Foster)++ | src/core/Complex.pm:
03:07 dalek rakudo/master: More Complex versions of infix:<**>.
03:07 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/3​d02e7cc977f1394280181fb6425721a76f60056
03:08 dalek rakudo/master: cf0cfea | (Solomon Foster)++ | t/spectest.data:
03:08 dalek rakudo/master: Turn on power.t.
03:08 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/c​f0cfea08beeb337a5370fe89edcddcb3cc2941f
03:08 dukeleto how evil is imcc_initialize ?
03:14 dukeleto i just read what it does. doesn't seem *so* evil at first glance
03:15 coke joined #parrot
03:16 dukeleto coke: ahoy
03:21 dalek plparrot: cf045c9 | dukeleto++ | src/handler/plparrot.c:
03:21 dalek plparrot: Intialize IMCC and properly destroy the interpreter object
03:21 dalek plparrot: review: http://github.com/leto/plparrot/commit/c​f045c96430a96b87780ece4cba75e4503732f8f
03:23 coke dukeleto: hio
03:36 dalek rakudo/master: 70667a0 | (Solomon Foster)++ | t/spectest.data:
03:36 dalek rakudo/master: Turn on lc.t, uc.t, and e.t.
03:36 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/7​0667a0a992606b8c409329e7c9443ca4627d24c
03:40 jsut joined #parrot
03:41 janus joined #parrot
04:25 dalek parrot: r43961 | plobsing++ | branches/tt362/src (2 files):
04:25 dalek parrot: abstract away ImageIO-specific details from visit and thawfinish loops and move these out into src/pmc_freeze.c
04:25 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43961/
04:58 dalek parrot: r43962 | plobsing++ | branches/tt362/src/pmc/imageio.pmc:
04:58 dalek parrot: eliminate ImageIO.extra_flags and ImageIO.pmc_result
04:58 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43962/
05:26 patspam joined #parrot
05:47 dalek parrot: r43963 | plobsing++ | branches/tt362 (3 files):
05:47 dalek parrot: add ImageIOSize PMC type to efficiently get the size of an image. see tt362 for why.
05:47 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43963/
05:50 woosley joined #parrot
06:33 kurahaupo joined #parrot
07:02 AndChat| joined #parrot
07:03 uniejo joined #parrot
07:27 viklund joined #parrot
07:59 iblechbot joined #parrot
08:14 dalek parrot: r43964 | gerd++ | trunk/config/gen/makefiles/root.in:
08:14 dalek parrot: extend the changeset r43950 to add all dependency; reflecting ticket 644
08:14 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43964/
09:00 dalek TT #644 closed by gerd++: target 'make hello' broken
09:15 nbrown joined #parrot
09:32 bacek joined #parrot
09:45 bacek aloha
09:45 Austin Good morning, Bacek
09:45 bacek Austin, it's 20:45 here :)
09:46 Austin Yeah, but when did you wake up?
09:46 Austin Morning is a state of mind.
09:47 bacek about 12 hours ago...
09:48 Austin Oh, well.
09:48 Austin Good evening, then.
09:48 Austin How's things in the southern hemisphere?
09:49 bacek Pretty well. Today is first dry day in about couple of weeks..
09:49 Austin Really?
09:49 Austin And you posting those sunny day at the beach photos!
09:50 bacek Just to remember my self how it should be in summer.
09:50 Austin :)
09:50 Austin Did you cut your grass today?
09:50 bacek (And annoy people from other side of the World of course)
09:50 * Austin sings, "Conjunction Junction! What's your function?"
09:51 KingOfKarlsruhe joined #parrot
09:52 bacek I don't have front/back yard anymore.
09:52 bacek So, no grass cutting. Just enjoying beer.
09:52 mikehh t/benchmark/benchmarks.t - Failed test:  9 - Failed test 'examples/benchmarks/freeze.pasm' - Segmentation fault
09:52 mikehh all other tests PASS (pre/post-config, make corevm/make coretest, smoke (#32203), fulltest) at r43964 - Ubuntu 9.10 amd64 (g++ with --optimize)
09:56 Austin Is this because of your move closer to work?
09:56 Austin Or did you make some kind of horrific lawn maintenance error?
09:56 bacek I moved close to work. 5 minutes to drive.
09:57 Austin Yeah, I remember you mentioning that.
10:01 AndyA joined #parrot
10:08 bacek mikehh, I can't reproduce it on i368...
10:09 mikehh also fails on gcc (with --optimize) at r43964, passed at r43954
10:12 bacek I suspect r43956
10:12 bacek sigh...
10:15 chromatic That was my fear too.
10:15 nbrown joined #parrot
10:21 mikehh also fails gcc without --optimize
10:22 nopaste "bacek" at 114.73.21.243 pasted "Patch for buffers aligned size" (87 lines) at http://nopaste.snit.ch/19621
10:22 bacek mikehh, can you test nopasted patch?
10:23 bacek (I will commit it anyway, but I want to be sure that this is actually fix crash)
10:24 nopaste "mikehh" at 94.11.186.149 pasted "backtrace for examples/benchmarks/freeze.pasm" (37 lines) at http://nopaste.snit.ch/19622
10:25 mikehh applying the patch now
10:26 bacek ermm.. backtrace looks exactly as without r43956 applied.
10:26 * bacek confused
10:26 purl You won't be after this episode of Soap!
10:27 bacek Austin, (TT#1445) Just add it and get cheap karma!
10:30 joeri joined #parrot
10:32 chromatic We have six open defects in Coverity; I think I've closed three of them.
10:33 chromatic I'll commit after the release on Tuesday.
10:35 bacek We have 619 open trac tickets. Can you close 300 of them? :)
10:35 bacek It's same ratio as in Coverity!
10:36 gaz joined #parrot
10:39 dalek TT #1445 created by Austin_Hastings++: Add ResizableStringArray.get_number VTABLE function
10:41 dalek parrot: r43965 | bacek++ | branches/sys_mem_reduce:
10:41 dalek parrot: Branch to remove/reduce usage of sys_mem_allocate set of functions. Just because there is no such thing as 'system memory' in GC world. TT#1402
10:41 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43965/
10:44 mikehh bacek: same results with the patch
10:45 bacek sigh...
10:46 nopaste "bacek" at 114.73.21.243 pasted "Additional patch" (25 lines) at http://nopaste.snit.ch/19623
10:47 bacek mikehh, can you also apply second patch?
10:47 bacek On top of first one
10:58 mikehh bacek: ok that works
10:58 bacek mikehh++
10:59 mikehh make test passes, benchmark_tests PASS
10:59 mikehh haven't tried the rest of fulltest ;-}
10:59 bacek Too late. I've committed it into svn :)
11:00 bacek r43966
11:14 dalek parrot: r43966 | bacek++ | trunk/src/gc (3 files):
11:14 dalek parrot: Remove aligned_size. Always use aligned_string_size for buffers to avoid crashes in compact_pool
11:14 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43966/
11:31 mikehh joined #parrot
11:31 payload joined #parrot
11:32 mikehh had to reboot - checking i386 first
11:47 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#32205), fulltest) at r43966 - Ubuntu 9.10 i386 (g++ with --optimize)
11:48 mikehh ok hrading back to amd64 - bbiab
11:58 bluescreen joined #parrot
12:17 Austin Where is Exception::new being defined?
12:18 bacek VTABLE_init(_pmc)?
12:19 Austin Hmm...I'll say no.
12:22 Austin The only things in the namespace at startup are 'annotations' and 'backtrace'.
12:22 Austin Somebody is putting it there...
12:23 bacek new "Exception" == op new
12:25 Austin Right. But there's a "new" method in the namespace.
12:25 Austin Apparently I'm putting it there, but I don't know how.
12:25 Austin :(
12:27 mikehh joined #parrot
12:27 bacek "runtime/parrot/library/Parrot/Exception.pir"
12:27 bacek this one?
12:27 purl i heard this one was bugged too now
12:27 bacek purl forget this one
12:27 purl bacek: I forgot this one
12:30 Austin Could be, since my test code is run inside nqp.
12:31 Austin Yep. Right at the top of my file, it's already defined.
12:43 Austin Bacek, Exception is declared as "pmclass Exception auto_attrs {..."   Does that mean it inherits only from default.pmc ?
12:59 sri joined #parrot
13:00 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#32207), fulltest) at r43966 - Ubuntu 9.10 amd64 (gcc with --optimize)
13:03 kid51 joined #parrot
13:06 dalek TT #1446 created by Austin_Hastings++: Exception PMC 'clone' does not work
13:50 whiteknight joined #parrot
13:50 Austin Howdy, Whiteknight.
13:52 whiteknight good morning, Austin
13:54 Austin Effing git. Why would anyone expect that having cloned a repository, I might want to push back to it? What possible reason could there be for sharing the write info at the same time? Idiots.
13:55 whiteknight Yeah, I've noticed that in a lot of cases Git is not nearly so easy and intuitive as I had been lead to believe
13:55 whiteknight I won't argue with it's power or capability, but I'll be damned if it's easy
13:55 Austin :)
13:55 Austin git--
13:56 whiteknight For instance, I almost never use branches on git because it's far too complicated to work with them
13:56 Austin :-O
13:56 whiteknight and if I recall, branches are supposed to be the big selling point
13:57 Austin Actually, I don't mind the branching model. It's free-form, but it's supposed to be. It's all the "stodgy corporate luser" stuff that I desperately need that doesn't work without a great deal of crank turning and swearing.
13:57 dalek kakapo: bffafc6 | austin++ |  (42 files):
13:57 dalek kakapo: Created a much smaller 'base' library with minimal structure - mainly pmc types. Reorged some source, moving things into Internals and Parrot dirs. Added more bootstrap tests.
13:57 dalek kakapo: Signed-off-by: Austin Hastings <Austin_Hastings@Yahoo.com>
13:57 dalek kakapo: review: http://gitorious.org/kakapo/kakapo/commit​/bffafc6402172da6c18c46f109d428beb863e011
13:58 Austin And now I'll have to create a release history file for Kakapo, since I've just tagged release-1. (And discovered that it's not a good release, since it lacks a release history file, among other things.)
13:58 whiteknight Austin: exactly. I routinely work on two computers, and I find sharing branchwork between them is a huge hassle
14:03 whiteknight I do still prefer git in many cases, but my workflow has changed dramatically because I just can't easily do the things I want (or remember all the millions of commands and CLI arguments that make those things happen)
14:05 Austin Put it in your makefile.
14:05 Austin :)
14:06 whiteknight I'm going to slowly build up a library of aliases to make git act more like svn when I need it
14:10 bluescreen joined #parrot
14:11 shockwave joined #parrot
14:13 shockwave What's the name the print function equivalent to C's print, which takes format specifiers?
14:14 shockwave This is in PIR.
14:16 Austin No such animal.
14:16 whiteknight parrot has an sprintf
14:16 Austin shockwave: Try the String sprintf op, then use say
14:16 whiteknight sprintf + print does what you need
14:16 Austin (or print)
14:17 shockwave I'll look at sprintf.
14:17 shockwave Thanks.
14:18 whiteknight Austin: TT #1446, what do you suggest to be the solution? implement clone, or implement freeze/thaw?
14:18 Austin I didn't look that far, and I don't really understand what freeze/thaw is about.
14:19 Austin I do know a bunch of people hate it.
14:19 Austin For Exception, clone seems like a no-brainer.
14:19 whiteknight freeze/thaw is the serialization feature. Freeze turns a PMC into a byte buffer suitable for saving to a file, for example. Thaw turns that byte buffer back into a live PMC
14:19 whiteknight Okay, we can probably add Clone to it
14:21 lucian joined #parrot
15:12 mikehh joined #parrot
15:15 PacoLinux joined #parrot
15:21 cognominal joined #parrot
15:23 bubaflub joined #parrot
15:28 mikehh joined #parrot
15:32 payload joined #parrot
15:37 Psyche^ joined #parrot
15:37 dalek kakapo: 1900cf2 | austin++ |  (6 files):
15:37 dalek kakapo: Added README, VERSION files, plus basic tests for String.nqp
15:37 dalek kakapo: Signed-off-by: Austin Hastings <Austin_Hastings@Yahoo.com>
15:37 dalek kakapo: review: http://gitorious.org/kakapo/kakapo/commit​/1900cf24212cdedfc5a62ebb62ff8ecbc2754aea
15:52 shockwave !paste
15:56 nopaste "shockwave" at 76.119.137.239 pasted ".sub 'foo' :method # var x = 2" (10 lines) at http://nopaste.snit.ch/19624
15:57 shockwave Does anyone know if IMCC, or some underlying system of Parrot, optimizes code like this:
15:57 shockwave http://nopaste.snit.ch/19624
15:58 whiteknight shockwave: what kinds of optimizations, like constant folding or pre-computing?
15:59 whiteknight in either case, the answer is probably "no"
15:59 shockwave dang
15:59 shockwave ok
15:59 whiteknight shockwave: does seem like the kind of thing that could be optimized though
16:00 shockwave @whiteknight, at some point, I'm hoping my compiler will perform those optimizations. But for now, it doesn't do them.
16:00 shockwave And that's the kind of code that it currently spits out.
16:00 shockwave Hopefully, it will be good enough.
16:00 whiteknight shockwave: if you wrote an optimization add-on to PCT that does this, all languages would benefit
16:00 shockwave I'm not using PCT.
16:01 whiteknight well, I'm sure if you wrote a class to do it, it could be used in PCT anyway
16:01 whiteknight my point is, the more people who could benefit from the work, the more help and support you could get
16:01 shockwave No doubt.
16:02 shockwave What language is PCT written in?
16:02 whiteknight PIR and NQP mostly, I think
16:03 theory joined #parrot
16:03 shockwave My compiler is written in C++. Because of its architecture, I don't know if an optimizer for it would translate to PCT. But if it could, and when I do it, maybe it can be added to PCT in some way.
16:04 whiteknight does your compiler link to libparrot?
16:04 shockwave no
16:04 whiteknight oh, so it would be hard then. Nevermind
16:07 rblackwe joined #parrot
16:08 NotFound joined #parrot
16:08 NotFound hi
16:08 purl hey, NotFound.
16:34 payload joined #parrot
16:38 payload joined #parrot
16:39 darbelo joined #parrot
16:55 dalek rakudo/master: c65c28b | (Martin Berends)++ | docs/compiler_overview.pod:
16:55 dalek rakudo/master: [docs/compiler_overview.pod] over 80% updated for new master (former ng)
16:55 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/c​65c28ba503e8a10587b1da39ad08d16eed9d9e2
17:22 * whiteknight just found a beautiful function that's over 550 lines long
17:23 darbelo So much beauty!
17:25 tetragon joined #parrot
17:26 cotto whiteknight, where?
17:26 purl i think whiteknight, is the win64 build working?
17:26 whiteknight cotto: in code at work, not Parrot
17:26 cotto forget whiteknight,
17:26 purl cotto: I forgot whiteknight,
17:27 whiteknight luckly all Parrot developers know better
17:27 whiteknight whiteknight?
17:27 purl rumour has it whiteknight is mailto:wknight8111@gmail.com or the grand master funk or http://wknight8111.blogspot.com/
17:27 cotto whiteknight, do you read the daily wtf?  It sounds like you've got some submission-worth stuff.
17:27 whiteknight cotto: yeah, I read it every day. There are a few gems I have considered posting
17:28 bubaflub the code on the dailywtf really hurts my soul sometimes
17:35 chromatic joined #parrot
17:40 kurahaupo joined #parrot
17:46 whiteknight I have finally gotten suncc to work on my opensolaris VM
17:46 whiteknight apparently it doesn't play too nicely with gcc
17:47 chromatic OpenSolaris doesn't play nicely with GCC?
17:49 darbelo I would guess that we don't know the proper incantations to gcc on OpenSolaris.
17:50 darbelo Sun has this habit of shipping several subtly different libraries, and always linking to the wrong one.
17:51 whiteknight chromatic: well, on install they both try to gobble up the alias "cc"
17:51 whiteknight suncc was having some kind of conceptual difficulty realizing that it wasn't called "cc"
17:52 darbelo Oh, I misread. Sorry,
17:52 chromatic Other than that, how did you enjoy the play, Mrs. Lincoln?
17:53 whiteknight how do I tell Configure.pl not to build the cgoto core?
17:53 whiteknight suncc freaks out trying to build that
17:54 whiteknight nevermind, --cgoto=0
17:55 whiteknight I tell you what, the more I use opensolaris, the less I like it
17:58 cognominal joined #parrot
17:59 cotto_work good morning #parrot
17:59 davidfetter joined #parrot
18:15 gaz joined #parrot
18:26 cotto_work darbelo, you ready to rock the release?
18:32 whiteknight_ joined #parrot
18:33 whiteknight_ joined #parrot
18:45 cotto_work alternaltely, you could release the rock.
18:49 darbelo cotto_work: Haven't updated NEWS yet. Other thatn that I think we're good to go tomnorrow.
18:49 darbelo I'm aiming for a few hours before #parrotsketch.
18:51 payload joined #parrot
18:52 bacek joined #parrot
19:01 joeri joined #parrot
19:07 whiteknight joined #parrot
19:10 KingOfKarlsruhe joined #parrot
19:15 bacek morning... good morning
19:15 whiteknight hello bacek
19:16 bacek aloha whiteknight
19:16 whiteknight I saw that TT #1443 fix, very nice
19:18 bacek whiteknight, yeah. Apparently I forgot to update couple of other places. So final fix was done yesterday with help of mikehh++
19:18 whiteknight nie
19:18 whiteknight nice
19:18 * bacek cursing day when I decided to look at parrot GC
19:19 whiteknight you're doing a great job with it though
19:19 whiteknight can never have too much cleanup
19:20 bacek Why? Ideal program with 0 lines code requires no cleanup!
19:20 whiteknight the whole string compactor thing is such a damn mess. That's where we need to do most cleanup
19:20 cotto_work no garbage, no collection
19:20 whiteknight somehow I don't think Rakudo will run on that "ideal" parrot
19:21 darbelo whiteknight: Sure it will, they'll just need to write a few wxtensions ;)
19:21 darbelo s/wx/ex/
19:22 bacek string compactor isn't so bad actually.
19:23 bacek Only one problem - looks like it's called too early in mark&sweep phase.
19:23 whiteknight really? I dont remember where it gets called
19:26 dalek TT #1447 created by allison++: [NQP-rx] failure to capture two duplicate subrules in the same rule
19:27 bacek I just watched total_size inside this function. Pattern is like big chunk/small chunk/big chunk/small chunk
19:30 chromatic How much time do we spend compacting pools?  Does it help?
19:32 bacek no idea
19:33 payload joined #parrot
19:40 whiteknight i've wanted to get profiling numbers on it for a while
19:40 whiteknight the alternative is, I suppose, just using raw malloc/free calls to allocate string guts
19:41 whiteknight I would actually suspect that's faster
19:44 plobsing joined #parrot
19:47 cotto_work Parrot_gc_compact_memory_pool isn't showing up in the callgrind profile, in spite of being called 11 times when compiling tools/dev/pprof2cg.nqp
19:48 cotto_work actually, all that shows up is a single call to it, but no info about the function itself.
19:49 ash_ joined #parrot
19:49 cotto_work nm
19:52 cotto_work just needed to make kcachegrind display more items.
19:54 cotto_work Parrot_gc_compact_memory_pool takes up ~.869% of the time spent compiling tools/dev/pprof2cg,nqp.
19:56 whiteknight okay, so it's obviously not too too expensive
19:57 whiteknight as an experiment I would eventually like to see what happens to runtime performance if we rip it out completely and replace with free() and malloc() calls
19:58 cotto_work branches are cheap
20:02 kurahaupo joined #parrot
20:04 whiteknight yeah, I'll get around to doing it eventually
20:05 kurahaupo pwd
20:06 * kurahaupo is obviously lost -- not only wrong directory, but wrong window too.
20:10 cotto_work It's too bad purl doesn't reply to short messages
20:11 darbelo pwd?
20:11 purl /home/darbelo/private/naked-monkey-photos
20:11 darbelo pwd
20:12 kurahaupo :-)
20:12 darbelo That '?' makes the difference, apparently.
20:12 ash_ doncha hate when your you change windows and are apparently writing over what your working on with cd somedir
20:13 kurahaupo huh?
20:14 ash_ yeah, apparently i fail at english today
20:15 ash_ i meant to say, don't you hate when you change windows and are apparently writing over what your working on, as in you accidently end up on your text editor instead of the terminal
20:15 kurahaupo (purl doesn't know "huh")
20:17 * kurahaupo wonders how 1 missed focus and 4 keystrokes can derail #parrot
20:18 kurahaupo ash_: don't ask "who is purl", you don't want to know :-)
20:19 tewk Anyone know of a cpan package that wraps git for scripting etc?
20:20 ash_ as a simplification layer or as a library call?
20:21 tewk ash_, I'm interested in both, but a simplification layer
20:22 ash_ cogito is something worth looking into if you want a simplification, its even made by the developers of git
20:22 ash_ url: http://git.or.cz/cogito/
20:23 ash_ oh, umm wow, just realized it's apparently been deprecated so... that might not be worth looking into, sorry about that, i need to stop using it (or find a replacement)
20:25 ash_ tewk: http://git.wiki.kernel.org/index​.php/InterfacesFrontendsAndTools has a list of front ends to git if you want to look there
20:26 kurahaupo joined #parrot
20:28 hercynium joined #parrot
20:53 whiteknight purl msg Coke can you take a look at the "depends" branch? It's 7 months old and appears to be dealing with dependencies, which I know you've been taking a fresh stab at
20:53 purl Message for coke stored.
20:53 whiteknight bacek: ping
20:54 bacek whiteknight, pong
20:54 whiteknight bacek: the "pmc_i_ops" branch, what's the status on that?
20:54 whiteknight I suspect it's getting unviable
20:54 * chromatic sets the linker on fire
20:54 bacek Waiting for architect decision.
20:55 chromatic In theory, you can tell GNU ld to use an improved hashing algorithm to reduce relocation collisions.
20:55 chromatic In practice... will it do so?  You tell me.
20:55 bacek It's basically replacing 3-arg VTABLEs in terms of 2-args
20:55 whiteknight bacek: can we ping allison to make her remember the question?
20:56 whiteknight purl msg Util what's the status of the "autoprops" branch? Doesn't look like any activity in 4 months. Probably better to delete and replace with a fresh branch if needed?
20:56 purl Message for util stored.
20:57 bacek whiteknight, why not?
20:57 whiteknight bacek: does it have a ticket associated with it?
20:57 bacek whiteknight, probably no...
20:57 whiteknight seen rblasch?
20:57 purl I haven't seen 'rblasch', whiteknight
20:58 whiteknight purl msg pmichud: what's the status of the dectx branch? Looks like no movement in 2 months. Might be better to delete this and create a fresh branch if you need to do work?
20:58 purl Sorry, I've never seen pmichud before.
20:59 whiteknight purl msg pmichaud: what's the status of the dectx branch? Looks like no movement in 2 months. Might be better to delete this and create a fresh branch if you need to do work?
20:59 purl Message for pmichaud stored.
21:01 whiteknight I really wonder what's the value of the old vc9 branch?
21:01 GeJ Good morning everyone!@#
21:01 whiteknight It's basically 10,000 revisions old, and predates 1.0.0
21:02 cotto_work You can always take a diff from the most recent sync point and see what it changes.
21:02 cotto_work btw, whiteknight++ for spring cleaning
21:05 dalek parrot: r43967 | whiteknight++ | branches/vtable_massacre:
21:05 dalek parrot: this branch has been merged to trunk
21:05 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43967/
21:05 dalek parrot: r43968 | whiteknight++ | branches/tt_1020:
21:05 dalek parrot: deleting an old branch that was never used and is now woefully out of date
21:05 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43968/
21:06 cotto_work There's not much love for TT #1020.
21:06 chromatic allison had some ideas about it.
21:06 whiteknight it was quite ambitious. Maybe overly so
21:06 whiteknight I like the idea but I worry about the performance implications
21:11 whiteknight anyway. Heading home. Later
21:15 payload joined #parrot
21:33 Psyche^ joined #parrot
21:37 dalek parrot: r43969 | bacek++ | trunk/compilers/pirc/src (9 files):
21:38 dalek parrot: Add headerizer directives to PIRC
21:38 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43969/
21:38 dalek parrot: r43970 | bacek++ | trunk/src/pmc_freeze.c:
21:38 dalek parrot: Add headerizer into pmc_freeze.c
21:38 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43970/
21:38 dalek parrot: r43971 | bacek++ | trunk/compilers/pirc/src (9 files):
21:38 dalek parrot: Add more headerizer directives into PIRC and rerun headerizer
21:38 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43971/
21:38 dalek parrot: r43972 | bacek++ | trunk (2 files):
21:38 dalek parrot: Make headerizer happy with Parrot_gc_sys_name.
21:38 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43972/
21:38 dalek parrot: r43973 | bacek++ | trunk/src (2 files):
21:38 dalek parrot: Rerun headerizer
21:38 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43973/
21:40 chromatic Rakudo precomiling Test.pm to Test.pir spends 9% of its execution time managing COW STRINGs.
21:42 chromatic 11.66% of its time sweeping GC pools.
21:48 bacek joined #parrot
21:49 bacek chromatic, can we have immutable strings please?
21:51 chromatic Love to.
21:54 cotto_work headerizer is more complainey than normal
21:54 cotto_work but it works, so bacek++
21:56 dalek TT #1423 closed by cotto++: pirc breaks headizer
21:56 cotto_work and only now do I notice that I spelled "headerizer" incorrectly
22:06 patspam joined #parrot
22:07 Whiteknight joined #parrot
22:08 chromatic Hm, some low-hanging fruit in the String PMC.
22:10 shillo joined #parrot
22:12 Whiteknight ?
22:12 Whiteknight do tell
22:14 shillo joined #parrot
22:15 chromatic Removing extraneous COW copies by replacing get_string() calls with attribute access.
22:24 Coke msg Whiteknight see http://trac.parrot.org/parr​ot/wiki/BranchDescriptions - that's petdance's branch. http://trac.parrot.org/parrot/changeset?new​=40391@branches/depends&amp;old=40372@trunk shows that there's not much worth saving over the current state. (I'd ask petdance if he wants to save it.)
22:24 purl Message for whiteknight stored.
22:25 Whiteknight Coke: okay, thanks.
22:27 dalek parrot: r43974 | plobsing++ | branches/tt362 (5 files):
22:27 dalek parrot: use ImageIOSize in Parrot_freeze_size
22:27 dalek parrot: use Parrot_freeze_size in Packfile_Constant_pack_size.
22:27 dalek parrot: this completes the todo item in this function and satisfies tt362
22:27 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43974/
22:28 Coke Whiteknight: I wrote that page/tool to show who is working/worked on the branch so you can target those sorts of "who is the person to talk to about this branch" emails. (I'm fine with getting pinged on depends since I am working on that; the vc9 one you could probably have just bounced off Bob.
22:29 Whiteknight Coke: I've been trying to ping the correct parties on most branches, I pinged you on this one because I haven't seen petdance in a while and you're working on the makefile
22:41 pmichaud purl msg whiteknight dectx branch can be destroyed -- I don't need it for anything
22:41 purl Message for whiteknight stored.
22:41 Whiteknight thanks, will do
22:44 Whiteknight done
22:47 cotto_work pmichaud, when will you have time to look at the ops_pct branch?
22:50 pmichaud cotto_work: don't know.  I'm a bit behind on rakudo stuff at the moment and have a slight cold to boot :(
22:50 pmichaud I'm only operating at about 60% efficiency, I think :(
22:50 pmichaud I'll try to take a look tonight or tomorrow morning (more likely tomorrow after sleep)
22:57 cotto_work sounds fine
22:59 dalek rakudo/master: 8564837 | (Solomon Foster)++ | src/core/Any-str.pm:
22:59 dalek rakudo/master: Add trim-leading and trim-trailing, and reimplement trim in terms of them.
22:59 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/8​5648371b452f94e255ba76c19c41d7edb2d5a85
22:59 dalek rakudo/master: 4246be2 | (Solomon Foster)++ | t/spectest.data:
22:59 dalek rakudo/master: Turn on trim.t.
22:59 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/4​246be20f4d32663c8f8fb3f98b36b8f7c444a7f
22:59 dalek parrot: r43975 | whiteknight++ | branches/dectx:
22:59 dalek parrot: removing unused branch on request from pmichaud++
22:59 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43975/
23:01 payload joined #parrot
23:05 dalek rakudo/master: cfe8cb3 | chromatic++ |  (7 files):
23:05 dalek rakudo/master: Merge branch 'master' of git@github.com:rakudo/rakudo
23:05 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/c​fe8cb3f34208f7b653dcc5667b239c8dcd96f11
23:06 dalek rakudo/master: 1c51479 | chromatic++ | build/Makefile.in:
23:06 dalek rakudo/master: [build] Fixed PMC include path when building bind.c.
23:06 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/1​c51479e3ca9c3766c865244644289a7e4db519b
23:06 dalek rakudo/master: e5f4b23 | chromatic++ | src/binder/bind.c:
23:06 dalek rakudo/master: [binder] Fixed compiler warnings.
23:06 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/e​5f4b23e923e9e13c79387ff6bdde68b6375357a
23:06 dalek rakudo/master: 448a825 | chromatic++ | src/pmc/perl6multisub.pmc:
23:06 dalek rakudo/master: [PMC] Fixed compiler warnings.
23:06 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/4​48a8253015474c3ecf4f972c9bf318e551234a8
23:06 dalek rakudo/master: ee1e518 | chromatic++ |  (2 files):
23:06 dalek rakudo/master: Merge branch 'master' of git@github.com:rakudo/rakudo
23:06 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/e​e1e518df5f981621e287cd08f3d1d153a6723ad
23:06 dolmen joined #parrot
23:10 cotto_work joined #parrot
23:12 payload joined #parrot
23:16 chromatic Signs you have a lousy hashing implementation: when you reduce the number of GCables created, but performance decreases because of more hash collisions and more hash bucket chain walking.
23:17 cotto_work lta
23:24 Whiteknight hash algorithm should be easy enough to change, right?
23:24 chromatic Yeah, it ought to be.
23:25 darbelo We don't know if it actually is, though.
23:25 Whiteknight it should be properly encapsulated behind a clean function or two
23:25 darbelo Maybe it is, I try to not look at hash internals too much.
23:26 chromatic It's pretty precarious.
23:28 wagle if i want to do a find_name, and want to catch the exception if it doesn't exist, which of the exception handling instructions do i use?
23:28 Whiteknight goddamnit, why can't any of these older subsystems have been coded to any level of quality?
23:29 Whiteknight wagle: push_eh and pop_eh
23:30 * wagle goes off to stare at those the hardest for a bit
23:34 Whiteknight wagle: find_method_not_null might be interesting too
23:35 wagle generic pmc..  actually, i want to die if its set, not die if its not..  i think..
23:36 wagle or else I might change my tactics entirely..  but thanks, will look at that
23:36 lucian joined #parrot
23:37 cotto_work joined #parrot
23:39 wagle docs need examples..  whats the simplest way to say something like: throw "aieeee!"
23:40 * wagle is learning pir
23:40 Whiteknight I think that's exactly it: throw "aieee!"
23:40 wagle doesnt like it
23:40 * davidfetter catches an "aieee!"
23:40 wagle The opcode 'throw_sc' (throw<1>) was not found. Check the type and number of the arguments
23:40 * darbelo segfaults
23:40 purl No whammies!
23:41 darbelo wagle: You want to throw an Exception, not a string.
23:41 Whiteknight no
23:41 Whiteknight $P0 = new ['Exception']
23:41 Whiteknight $P0['message'] = "aieee"
23:41 Whiteknight throw $P0
23:41 wagle makes sense thanks
23:42 Whiteknight I thought we had a throw_s form
23:42 Whiteknight we do have die_s, that makes for lousy symmetry
23:42 cotto_work Such a form could be added easily enough.
23:43 cotto_work but then it wouldn't have any type information or severity
23:44 Whiteknight cotto_work: for most uses, those things aren't needed or wanted anyway
23:44 Whiteknight and it's not like you couldn't construct an exception "the long way" too
23:44 darbelo Does anyone know what was the purpose of bacek's orederedhash_revamp branch was?
23:44 darbelo Other than revamping OrderedHash.
23:45 bacek_at_work darbelo, replace hackish implementation with not-so-hackish one
23:45 cotto_work make the internals less hacky iirc
23:45 Whiteknight yeah, the internals are much cleaner now
23:45 darbelo But the interface changed, didn't it?
23:46 bacek_at_work darbelo, little bit.
23:46 bacek_at_work It actually same apart from iterating over OH with removed keys.
23:47 darbelo Ok. I guess I'm done with NEWS for today then.
23:52 wagle howcome I cant do: $P0['message'] = concat "aieeee!","foo"
23:53 Whiteknight wagle: it's an assembly language. Too low-level
23:53 Whiteknight $S0 = concat "aieee!", "foo"
23:53 wagle isnt that a basic/simple assignment?
23:53 Whiteknight $P0['message'] = $S0
23:53 wagle guessnot
23:53 Whiteknight no, it's two operations: the concat and the indexed assign
23:54 wagle hokeh
23:54 payload joined #parrot

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

Parrot | source cross referenced