Parrot 2.0.0 "Inevitable" released! | parrot.org | Priorities: merge branches, remove deprecations | Roadmap: icanhaz.com/parrotroadmap | Latest modified TT's: icanhaz.com/parrotbugs
Set by moderator on 9 February 2010.
bacek_at_work infoq.com for example 00:03
eg www.infoq.com/presentations/randal-...g-dynamism
cotto_work That's the one I thought of.
Coke I declare dibs on EvilScheme 00:10
Whiteknight just cleaned 2 feet of snow off his car 00:14
chromatic just cleaned 0 feet of snow off his car, because he has a garage. Also, it doesn't snow much here.
davidfetter fortunately it doesn't snow much here either 00:17
seeing as we have no gear to deal with same
Whiteknight chromatic: you made the effort to clean 0 feet of snow off something?
bacek_at_work just get another cold coke because it's +33 Celsius outside
cotto_work likes the snow but nobody here knows how to drive in it.
Whiteknight sounds like we need to put you to work on a coding project if you're that bored 00:18
00:20 AndyA joined
dalek rrot: r43914 | coke++ | branches/rm_cflags/config/gen/makefiles/root.in:
Avoid some warnings on ops.
 00:21
chromatic I'm trying to explain Perl 5 references to novices. That's a project. 00:28
Whiteknight urg 00:31
davidfetter can you include a lolcat to good effect here? 00:32
Whiteknight I wouldn't even consider myself adept with those beasts
darbelo I'm not entirely convinced it can be done in a way that preserves their novice status.
cotto_work Is this for fun or just profit?
chromatic It's a section in the Modern Perl book.
Tene chromatic: I was explaining Perl 5 references to a novice yesterday. 00:34
chromatic It's not too bad if you take the right approach.
davidfetter wonders whether anybody's got good tricks for remembering to repeat each question/comment back to the audience 00:45
davidfetter sometimes fails at this :P
chromatic Get the guy in the second row to yell out "Repeat the question?"
Tene shock collar 00:46
darbelo A sniper in the back row.
davidfetter hrm. that's a great idea, chromatic
darbelo, i kinda prefer lessons i can use later, dramatic as yours would be
darbelo You'll never make that mistake twice, I assure you. 00:47
cotto_work The next speaker would never forget to repeat questions.
davidfetter heh
darbelo That too.
davidfetter so get it for the previous speaker. hrm...
davidfetter has a little list of speakers he'd like to precede him ;)
00:47 AndyA joined
cotto_work A paintball gun would be similarly effective and less illegal. 00:48
davidfetter i'd still rather have it happen to somebody else ;)
the whole, "confederates in the audience" trope is one i haven't used as much as i probably should 00:49
chromatic Didn't I recommend that to you once upon a time?
davidfetter you did. i'm slow sometimes :)
chromatic I thought I had, or else I'm time traveling again. 00:50
davidfetter i'm using it more than i'm used to, just not enough yet
than i used to*
darbelo If it works for traveling fair magicians...
01:07 abqar joined
Austin Davidfetter: (re: repeating the question) 01:25
Try making a list of 5 or so expected questions.
Then "coerce" the first 5 questions into the ones on your list.
It helps if you show the list, and then make it kind of a joke that you're going to answer those - "What's that, you ask? Where can I download Parrot? Well, I have a slide for that..." 01:26
davidfetter hrm
heh
cotto_work I cried and had trouble breathing: languagelog.ldc.upenn.edu/nll/?p=2110
best minute EVAR
Austin By doing that, you get yourself into the habit of "repeating" the question, as part of the joke.
davidfetter is it getting /.d, cotto_work ?
cotto_work laughing 01:27
purl Yes, very funny. Now get back to work.
cotto_work not sure
Austin Then when you run out of slides, you've set the expectation with the audience that you're repeating the questions. (Also, that you're willing to avoid certain questions...)
davidfetter lol!
hrm 01:28
my goal here is to make sure my mike--and the audience--actually hears questions as asked
Austin ROFL.
Good one, cotto_work.
cotto_work++
davidfetter: Sure, but the hardest part is getting yourself to repeat them. 01:29
Making the repetition the setup for a little gag makes it fun for you.
davidfetter :)
Austin grrr... 01:39
Austin finds yet another place where parrot-nqp's change away from :init subs at the top level bites...
01:41 AndyA joined, patspam joined, bacek_at_work joined, hudnix joined, Patterner joined, ascent joined, Tene joined, ingy joined, KatrinaTheLamia joined, solarion joined, allison joined, leto joined, rhr joined, GeJ joined, workbench joined, slavorgn joined, Essobi joined, cxreg joined, dukeleto joined, Util joined, Coke joined, nopaste joined, szabgab joined, kvorg joined, cotto joined, integral joined, preflex joined, Infinoid joined, he joined, jjore joined, dalek joined, jan joined, elmex joined, baest joined, Maddingue joined, spinclad joined, frodwith_ joined, jsut_ joined, tewk joined, treed_ joined, redbrain_ joined, Hunger joined, athomaso1 joined, NotFound joined, wagle joined, _sri joined, kjeldahl__ joined, mj41 joined, Austin joined, eiro joined, darbelo joined, confound joined, TonyC joined, Khisanth joined, purl joined, hicx174 joined, jhelwig_ joined, PacoLinux joined
dalek rrot: r43915 | mikehh++ | branches/rm_cflags/compilers/pirc/src/pircapi.c:
fix codetest failure - add missing ASSERT_ARGS
 01:42
01:45 AndyA joined, patspam joined, bacek_at_work joined, hudnix joined, Patterner joined, ascent joined, Tene joined, ingy joined, KatrinaTheLamia joined, solarion joined, allison joined, leto joined, rhr joined, GeJ joined, workbench joined, slavorgn joined, Essobi joined, cxreg joined, dukeleto joined, Util joined, Coke joined, nopaste joined, szabgab joined, kvorg joined, cotto joined, integral joined, preflex joined, Infinoid joined, he joined, jjore joined, dalek joined, jan joined, elmex joined, baest joined, Maddingue joined, spinclad joined, frodwith_ joined, jsut_ joined, tewk joined, treed_ joined, redbrain_ joined, Hunger joined, athomaso1 joined, NotFound joined, wagle joined, _sri joined, kjeldahl__ joined, mj41 joined, Austin joined, eiro joined, darbelo joined, PacoLinux joined, jhelwig_ joined, hicx174 joined, Khisanth joined, TonyC joined, confound joined, TonyC joined, Khisanth joined, hicx174 joined, jhelwig_ joined, PacoLinux joined, darbelo joined, eiro joined, Austin joined, mj41 joined, kjeldahl__ joined, _sri joined, wagle joined, NotFound joined, athomaso1 joined, Hunger joined, redbrain_ joined, treed_ joined, tewk joined, jsut_ joined, frodwith_ joined, spinclad joined, Maddingue joined, baest joined, elmex joined, jan joined, dalek joined, jjore joined, he joined, Infinoid joined, preflex joined, integral joined, cotto joined, kvorg joined, szabgab joined, nopaste joined, Coke joined, Util joined, dukeleto joined, cxreg joined, Essobi joined, slavorgn joined, workbench joined, GeJ joined, rhr joined, leto joined, allison joined, solarion joined, KatrinaTheLamia joined, ingy joined, Tene joined, ascent joined, Patterner joined, hudnix joined, bacek_at_work joined, patspam joined, AndyA joined, sjn joined, pjcj joined, moritz joined, TimToady joined, zibri joined, PerlJam joined, ttbot joined, janus joined, silug_ joined, kurahaupo joined, davidfetter joined, payload joined, slavorg joined, abqar joined, purl joined 01:47 AndChat| joined, AndyA joined, patspam joined, bacek_at_work joined, hudnix joined, Patterner joined, ascent joined, Tene joined, ingy joined, KatrinaTheLamia joined, solarion joined, allison joined, leto joined, rhr joined, GeJ joined, workbench joined, slavorgn joined, Essobi joined, cxreg joined, dukeleto joined, Util joined, Coke joined, nopaste joined, szabgab joined, kvorg joined, cotto joined, integral joined, preflex joined, Infinoid joined, he joined, jjore joined, dalek joined, jan joined, elmex joined, baest joined, Maddingue joined, spinclad joined, frodwith_ joined, jsut_ joined, tewk joined, treed_ joined, redbrain_ joined, Hunger joined, athomaso1 joined, NotFound joined, wagle joined, _sri joined, kjeldahl__ joined, mj41 joined, Austin joined, eiro joined, darbelo joined, PacoLinux joined, jhelwig_ joined, hicx174 joined, Khisanth joined, TonyC joined, confound joined
davidfetter hrm. is something whacking irc.perl.org ? 01:47
01:49 jsut joined
cotto_work probably under attack again 01:53
Austin Ladies and gentlemen, I give you, trac.parrot.org/parrot/wiki/Parrot%20Dictionary 01:54
cotto_work Why the space?
oh wow. That's a lot.
Austin Just in time for Whiteknight to have to edit it to get rid of all the opcodes.
cotto_work We clearly need a zoom op. 01:55
Austin zip ? 01:56
purl i think zip is actually built-in to many prebuilt systems. or zip(@a, @b) == ($a[0], $b[0], $a[1], $b[1], ...) or at www.7-zip.org/ or �.
Coke Austin++ on that wiki page. email the list about that one. 01:58
Austin okay
01:59 bacek joined
mikehh rm_cflags branch: 02:03
All tests PASS (pre/post-config, make corevm/make coretest, smoke (#32162), fulltest) at r43915 - Ubuntu 9.10 amd64 (g++ with --optimize)
cotto_work an ad every 90 seconds is irritating
Austin++ 02:04
mikehh Coke: lots more warnings compared to trunk - mostly unused vars 02:05
dalek tracwiki: v153 | Austin_Hastings++ | WikiStart 02:07
tracwiki: Added dictionary link
tracwiki: trac.parrot.org/parrot/wiki/WikiSta...ction=diff
tracwiki: v1 | Austin_Hastings++ | Parrot%20Dictionary
tracwiki: Imported
tracwiki: trac.parrot.org/parrot/wiki/Parrot%...ction=diff
Austin Hmm. Do classes have to "addattribute" all their inherited attributes? Or are those magically inherited from the parents? 02:21
02:25 plobsing joined
Tene I'm not certain, but I'd be a little surprised if not inherited. 02:29
Coke mikehh: yes. that's because trunk disabled it by defaults and enables it on a few files. I'm enabling it everywhere, fixing what I can and disabling the rest. 02:33
mikehh: it should be gcc warnings clean before mergeback. =-)
mikehh Coke: anyway all test PASS at the moment :-} 02:34
Coke: let me know before and I'll check for g++ 02:35
02:36 AndChat| joined 02:37 jsut_ joined
dalek tracwiki: v3 | mikehh++ | CFunctionDocs 02:40
tracwiki: bring up to date at r43915</a>
tracwiki: trac.parrot.org/parrot/wiki/CFuncti...ction=diff
tracwiki: v4 | mikehh++ | CFunctionDocs
tracwiki: up to r43915</a>
tracwiki: trac.parrot.org/parrot/wiki/CFuncti...ction=diff
Austin Tene: Q: Given a string-or-protoobject, I turn it into a class with P6object.get_parrotclass. What do I call to create a new instance? 03:01
Tene Austin: use the 'new' opcode. 03:03
Austin Actually, I'm trying to get back to the proto-object, since I know that has a new method that I want to run. 03:05
Tene if you have a string, get_hll_global the namespace. 03:06
If you already have a protoobject, you're done.
Austin: you might try the 'get_proto' method. 03:07
it's listed in the POD immediately before 'get_parrotclass'
get_proto(name)
Retrieve the protoobject for "name". Return null if no protoobject
exists, or whatever is present isn’t a protoobject.
Austin This is (one of the many reasons) why I hate venturing into the swamp that is P6object... 03:10
aha. parrotclass has "setprop" 'metaclass'. 03:15
so $protoobject.HOW.WHAT 03:16
Grr 03:20
But I'm starting from parrotclass, which has no methods. :(
Grr
03:22 theory joined 03:23 bacek joined
Austin So it's $metaclass := Q:PIR { %r = getprop 'metaclass', $P0 } ; then $metaclass.WHAT() 03:25
03:35 janus joined, davidfetter joined 03:43 theory joined 03:55 theory joined 03:56 jsut joined 04:12 theory joined 04:22 theory joined 04:43 theory joined, kurahaupo joined 04:44 AndChat| joined 04:57 kurahaupo joined
Util Austin++ # for Dictionary 04:59
Austin Thank you.
04:59 Khisanth joined
Austin And now, an open-source reboot. 05:10
git--
05:21 Austin joined
Austin Mmmm...Kakapo's UnitTest::Testcase now has a TAP listener. 05:35
05:40 theory joined
dalek tracwiki: v2 | Util++ | Parrot%20Dictionary 05:42
tracwiki: Added .str, CONST_STRING, CONST_STRING_GEN
tracwiki: trac.parrot.org/parrot/wiki/Parrot%...ction=diff
bacek_at_work OW 05:57
WOW
Austin++ # Dictionary is really awesome!
Austin :) 05:58
cotto bacek_at_work, ping 06:21
06:28 theory joined
cotto anyone know why two-stage PMC initialization is necessary? 06:35
Austin Prolly because you can't count on ordering
Or do you mean internally? 06:36
cotto yeah, internally
Austin link?
purl or "Link is ... like ... this pointy eared goblin that walks around in midi-music land with a letter opener attacking circles and things and wooing princesses but not bannon, you know?" or preaction is Error.
cotto I want to get it down to one stage and then make it lazy
Austin Sure.
But if you're talking about init from a hash, you can't count on key ordering. 06:37
Otherwise, give me a link or a nopaste, and I'll see what I can see.
cotto I mean like filling in VTABLEs and MROs.
Austin Is there something pluggable down at a low level? 06:38
cotto as in the initialization code for the String PMC isn't called util the first String PMC is instantiated.
Austin Ahh.
cotto It'll speed up startup.
Austin Maybe they're trying to make it really, really lazy
:)
Would the single-stage init for stuff that doesn't get used slow down startup? 06:39
cotto It's a bunch of function calls and memory copying that doesn't strictly need to happen. 06:40
Austin Well, for a while there chromatic was running around chasing femtoseconds. Ask him. 06:41
cotto I would but he's not on atm. He usually joins about this time though.
Austin Yeah.
cotto It's hard being lazy.
Austin Okay, time for another git-induced reboot. 06:44
bacek_at_work cotto, pong 06:51
06:56 Austin joined
cotto bacek_at_work, what kind of snags do you think I might run into trying to make PMC initialization lazy? 06:57
(just picking your brain)
bacek_at_work cotto: no idea. What for? 07:00
cotto It make startup faster.
Mostly because I feel like it and am waiting for pmichaud to find some time to look at ops_pct. 07:01
(and it seems like an area where Parrot could improve)
bacek_at_work I doubt that it will save more than 0.01s 07:13
cotto Perhaps. I should really know better than to optimize before profiling. 07:15
cotto goes off to profile.
Looks like Parrot_initialize_core_pmcs takes up about 25% of Parrot's startup time. 07:19
.04s seems more realistic. I may have to put this off until a more compelling reason presents itself. 07:22
07:22 uniejo joined 07:26 chromatic joined
cotto chromatic, ping 07:27
chromatic cotto, pong 07:31
cotto chromatic, can you think of a reason other than a .04s reduction in parrot's startup time why lazy PMC initialization would be a good idea?
chromatic What do you mean by lazy PMC initialization? 07:32
cotto don't initialize a PMC's _vtable, etc until the first time an instance is instantiated. 07:33
chromatic There's more complexity to that; you have to initialize all of its parents too. 07:34
cotto Yeah. So far I'm not thinking it'd be worthwhile.
I thought I'd bounce it off you before I abandoned the idea. 07:35
chromatic If we could cut 40% of startup time (not registering a lot of MMD that we don't need) we gain more. 07:36
I've thought about that for a while; skip the stage two initialization of PMCs until we first instantiate one.
cotto My profiling shows it at about 25% of startup time.
(i.e. looking at Parrot_initialize_core_pmcs when running a pir file with a single empty sub) 07:37
chromatic Creating MMD and NCI and METHODs and so forth?
cotto That's when the function is called with pass=1 iirc.
chromatic Right.
cotto that counts both passes 07:38
Why is the two-stage initialization required? 07:39
chromatic We use some PMCs to store data for other PMCs, and we have to be able to create instances of the first. 07:41
cotto ok
For now it sounds like it wouldn't be worthwhile. 07:44
chromatic I wouldn't say that, but it's not our biggest priority. 07:51
cotto I'll hold off for now and keep it in the back of my mind. I don't think it'd be more than a couple hours' hacking to get it implemented. 07:55
chromatic Let's hope, but who knows? 07:56
cotto The guy who does it, that's who.
I'm glad bacek++ had the good sense to remind me to profile before optimizing. 07:57
cotto sleeps 07:58
dalek parrot: f00a51a | dukeleto++ | src/handler/plparrot.c:
Import the rest of the Postgres header files and check if called as a trigger
 08:00
parrot: c5e7750 | dukeleto++ | src/handler/plparrot.c:
Various improvements to the language handler

data, stolen and tweaked from PL/Perl.
chromatic msg tewk Do you know Matthew Might? 08:04
purl Message for tewk stored.
Austin .msg tewk Do you know Donna Does? 08:06
Austin vamps.
chromatic What, was she a co-author of a paper on bounded CFA for dynamic languages? 08:07
Austin Better, she's a Ruby hacker. 08:08
chromatic Don't do what Donny Dont does.
Austin '' is not a valid vtable function name. 08:21
What does *that* mean?
08:22 iblechbot joined
chromatic Something called VTABLE_add_vtable_override() on a Class PMC with a sub that doesn't override a vtable. 08:29
Austin Hmm. I was calling "register" on Float pmc type after injecting a 'new' method. If I call it before adding 'new', no problem.
08:35 he joined 08:41 fperrad joined
dalek rrot-plumage: 47acafd | leto++ | t/03-util.t:
Add a few tests for all_matches
 08:47
09:12 Austin joined 09:25 AndyA joined 09:54 payload joined 09:56 bacek joined 10:00 cognominal joined 10:04 payload joined 10:45 cognominal joined 10:53 pjcj joined
dalek rrot: r43916 | bacek++ | branches/boehm_gc_2 (11 files):
Sync branch up-to-date with master.
 10:59
rrot: r43917 | bacek++ | branches/boehm_gc_2 (6 files):
Rename GC_DEBUG into PARROT_GC_DEBUG to avoid clash with Boehm GC
 11:32
11:40 AndyA joined 11:44 kurahaupo joined 12:19 ruoso joined 12:27 woosley joined 12:29 bluescreen joined
dalek rrot: r43918 | gerd++ | trunk/lib/Parrot/Configure/Messages.pm:
only update the configure message also to the new year
 12:37
12:39 Whiteknight joined, tetragon joined
Whiteknight good morning #parrot 12:39
Austin Guten morgen, Whiteknight. 12:43
Are you all dug out from the snow?
Ahh..
dalek kapo: 36a0418 | austin++ | (6 files):
Fixed parrot_nqp -> parrot-nqp in Makefile.in
 12:44
Austin sings "Glorious! Gitorious! One keg of booze for the four of us..."
dalek kapo: e17a820 | austin++ | (56 files):
Catching up after the holidays.
kapo: 7c3c9a5 | austin++ | (19 files):
Added COMMON.nqp for Pmc methods. Defined COMMON Pmc test case. Build testcases for Undef, Float, Sub. Extended UnitTest code.
Austin sings "Glory be to God that there are no more of us - for one of us could drink it all alone!" 12:45
Whiteknight Austin: define "dug out?" 12:52
Austin Umm... are you at work?
Able to go to work? 12:53
Able to see the top of your car?
Whiteknight I am able to go to work, yes. I'm going in a little late to let the roads thaw out
Austin Good enough. You're dug out. 12:54
close-parrot.blogspot.com/2010/02/progress.html 13:02
UnitTest works, finally.
13:15 woosley1 joined 13:18 woosley joined 13:35 woosley joined
Austin Does nqp have a logical xor operator? 13:36
Perl6 has ?^ 13:38
Rats.
Pmichaud, what's the hold-up on the ?^ op? 13:39
13:40 woosley joined, jsut joined 13:41 fperrad_ joined
Coke Austin: perhaps a fork on github with a pull request would be a better answer there. =-) 13:42
Austin Coke: " return $got == self.expected " 13:43
TRIMTOWTDI
Whiteknight okay, I'm getting ready and going to work. later 13:46
13:48 mj41_ joined 13:52 Austin_Hastings joined 13:58 iblechbot joined 14:10 clinton joined, plobsing joined
Coke oooh, exciting errand. back later. 14:11
14:14 davidfetter joined
pmichaud (?^ op) NQP doesn't add operators until there's a definite need for them. 14:41
Austin :) 14:42
On the one hand, I could work around it. On the other hand, the fact that I had to work around it indicates a need, no?
pmichaud sure 14:43
I don't have an issue with adding it
Austin Please don't.
pmichaud assuming Parrot has an opcode that dtrt already
Austin I don't expect to need it again for a couple of years.
15:03 lucian joined 15:20 clinton joined 15:24 barney joined 15:35 theory joined 15:38 Andy joined 15:45 cognominal joined 15:51 davidfetter joined 15:54 Psyche^ joined 15:55 cognominal joined 16:09 barney joined
pmichaud does anyone have an example of overriding :vtable('concatenate') from PIR? 16:25
nopaste "pmichaud" at 66.25.4.52 pasted "first attempt to override :vtable('concatenate') from PIR" (37 lines) at nopaste.snit.ch/19578 16:27
pmichaud oh, maybe it's just... 16:28
16:28 payload joined
pmichaud okay, got it. :) 16:29
nopaste "pmichaud" at 66.25.4.52 pasted "working :vtable('concatenate') from PIR" (35 lines) at nopaste.snit.ch/19579
pmichaud needs a dummy 'dest' parameter
17:00 jan joined 17:07 davidfetter joined 17:14 cotto_work joined 17:41 cotto_work joined
dukeleto 'ello 17:45
cotto_work 'hi 17:48
dukeleto jolly good day, fine Parrot folk 17:57
18:09 cognominal joined 18:12 plobsing joined
cotto_work plobsing, is all your code for the opengl_dynamic_nci branch committed? 18:29
It sounds like an improved approach.
plobsing cotto_work: I have no local changes if that's what you mean 18:31
cotto_work That's what I mean.
18:36 iblechbot joined 18:42 cognominal joined 18:43 Maddingue joined 18:46 Maddingu1 joined 18:47 cognominal_ joined, Maddingue joined 18:56 joeri joined 19:03 chromatic joined
davidfetter mornin' chromatic 19:06
chromatic morning
19:13 Psyche^ joined 19:18 Psyche^_ joined 19:19 whiteknight joined 19:23 ash_ joined
tewk yep, I was think crazy things like a meta runloop that could do arbitrary computations based on the bytecode stream being executed. 19:47
cotto_work and now you're replying to people who aren't there 19:51
ustream isn't very smart about pausing and unpausing after several minutes. 19:53
dukeleto cotto_work: yeah, it gets farty 20:00
20:01 theory joined 20:04 cotto_work joined
Coke alas? 20:18
purl And a lad, as well!
Coke alias?
purl i guess alias is Adam Kennedy
Coke supposes I can just use the cpan addy.
alias is also adamk at cpan dot org 20:26
purl okay, Coke.
20:40 bluescreen joined
dalek rrot: r43919 | mikehh++ | branches/vtable_massacre/t/pmc/bigint.t:
adjust plan (and copyright date)
 20:51
mikehh vtable_massacre branch: 21:03
All tests PASS (pre/post-config, make corevm/make coretest, smoke (#32168), fulltest) at r43919 - Ubuntu 9.10 amd64 (g++ with --optimize)
dalek rrot: r43920 | coke++ | branches/rm_cflags (13 files):
Only use platform specific slashes where necessary.
 21:08
21:09 bacek joined
Coke kills replace_slashes in the makefile generation. 21:11
(branch) 21:12
nopaste "mikehh" at 94.11.186.149 pasted "rakudo test failures - from vtable_massacre branch at r43919" (31 lines) at nopaste.snit.ch/19582 21:26
mikehh vtable_massacre branch: rakudo couple of failures in make test and a also in make spectest - nopaste.snit.ch/19582
s/a also/a few/ or s/a also/also/ 21:29
21:38 shamu joined 21:41 cotto_work joined 22:09 clinton joined
cotto_work How difficult is it to port something like Parrot to VMS? 22:24
Austin Heh
chromatic Fairly involved, I think. Don't count on anything POSIXy.
Austin How difficult was it to port to Windows?
cotto_work clock? 22:26
purl cotto_work: LAX: Fri 2:26pm PST / CHI: Fri 4:26pm CST / NYC: Fri 5:26pm EST / LON: Fri 10:26pm GMT / BER: Fri 11:26pm CET / IND: Sat 3:56am IST / TOK: Sat 7:26am JST / SYD: Sat 9:26am EST /
chromatic Windows is easier in comparison.
cotto_work so VMS (or OpenVMS) is something that will take a motivated and knowledgeable individual 22:27
Austin Or someone too naive to know better. 22:28
Maybe a GSoC project...
darbelo OpenVMS will need someone with access to an OpenVMS box. 22:38
That's probably harder to find than a motivated sucker^W person.
22:39 payload joined 22:40 cotto_work joined
Austin Is there some kind of --stop-with-the-segfaults-already option for parrot that isn't documented? 22:46
cotto_work We're accepting patches.
Austin That's what they all say 22:47
cotto_work Do you have a repeatable one?
Austin I'm not sure
looks like it 22:48
purl No it doesn't, shut your hole!
Austin smooch purl
It's one where modifying bytecode changes the fault, so it seems like a wild pointer 22:49
darbelo o/ Wild pointers, couldn't drag me away o/ 22:51
o/ Wild wild pointers couldn't drag me away o/
Austin Rolling Darbelo? 22:52
darbelo Was listening to that sound on the radio right now. Couldn't help it.
Austin Keep in mind that "repeatable" is not the same thing as "testcase". Its somewhere in 25k lines of pir.
darbelo Austin: I'm planing to write a replacement for pbc_dump do you have any gripes with that too you would like addressed? 22:53
s/too/tool/ 22:54
Austin Yeah: 1. Consistent offset numbers (hex, decimal, whatever). 2. In the -d mode, don't skip silently over the varargs. 22:55
dalek kapo: 0bd97cc | austin++ | (11 files):
This snapshot shows a segfault.
Austin 3. Sub and method boundaries should be delimited in -d mode 22:56
4. Generate some kind of branch target indicator.
I think those were the big ones.
You want tickets for those? 22:57
darbelo Nah, I haven't started coding yet. Just wanted to make you feel like we listen to you every once in a while ;) 22:58
We need to be more responsive to HLL developers getting shagged by rare parrots. 23:00
23:01 mikehh_ joined
Austin I'm on a horse! 23:04
23:06 mikehh_ joined 23:07 iblechbot joined
Austin TT#1437 23:16
TT# 1438 - Indicate sub/method boundaries 23:25
darbelo Hm. How long is the usual "stop breaking stuff and get some testing done" period beofre a release? 23:29
dalek TT #1437 created by Austin_Hastings++: pbc_dump: don't skip varargs words in -d mode 23:30
TT #1438 created by Austin_Hastings++: pbc_dump: Indicate sub/method boundaries in -d mode
chromatic darbelo, probably Saturday night until Tuesday. 23:31
darbelo Guess I'll send a mail to the list then. 23:33
chromatic Mostly it depends on how much excitement you want. 23:35
Austin TT#1439 - mark branch targets 23:38
Sweet. 23:41
It's Friday, it's after 6, the sun's sunk below the yard-arm, and it's Jim's birthday tomorrow. 23:42
dukeleto 'ello
Jim who?
purl rumour has it Jim is being helpful and polite
Austin Any one of which would be a justification for some drinking.
Jim's a buddy of mine.
No parrot connection, other than me.
dukeleto darbelo: what is your plan to replace pbc_dump? 23:43
darbelo Actually that's not the end goal. I want to give the packfile PMCs some use. 23:44
Austin Dukeleto: How'd it go? Anything online yet?
darbelo Figured inspecting and dumping bytecode was a good way to dogfood it. 23:45
dukeleto Austin: well and yes
darbelo I'll probably do it from PIR for maximum dogfooding.
dukeleto Austin: i forgot to mail parrot-dev the info. duh
Austin Don't worry. Nobody reads that list. 23:46
Nobody Austin: No I don't.
Austin :)
Nobody++
darbelo karma Nobody 23:47
purl nobody has karma of 1
dalek TT #1439 created by Austin_Hastings++: pbc_dump: Mark branch targets in -d mode
dukeleto Nobody-- ;) 23:48
darbelo What'd you do that for? Now nobody's got no karma. 23:49
23:51 payload joined 23:52 tetragon joined
Austin I can only use one printf for debugging this code, because having two printfs converts the error into a segfault. 23:52
darbelo Austin: Can you nopaste a backtrace from that segfault? 23:53
Austin I can only use one printf for debugging this code, because having two printfs converts the error into a segfault.
dukeleto likes "Just another point in the solution space"
Austin bah, wrong window
nopaste "Austin" at 68.37.46.53 pasted "segfault backtrace" (46 lines) at nopaste.snit.ch/19584 23:55
darbelo Ouch. Died in Parrot_str_new_COW.
Austin Yeah, but 165 for me is "165\t if (PObj_constant_TEST(s)) { 23:56
darbelo And THAT sucks even more. 23:57
chromatic Do you have any custom PMCs or ops? 23:58
Austin Nope.
With three exceptions, everything is in nqp.
And the exceptions are small. 23:59
(and pir)

Powered by Cro and the Raku® Programming Language.

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

Thank you!