Parrot 1.7.0 "African Grey" is out! | Testing hackathon November 14 and 15 -- trac.parrot.org/parrot/wiki/Opcode...gHackathon | find out what's up with the slice opcode | Latest modified TT's: icanhaz.com/parrotbugs
Set by moderator on 13 November 2009.
japhb darbelo, reading ... 00:01
plobsing japhb: src/call/pcc.c seems to be relevant
japhb plobsing, OK, perldoc on that file is quite revealing, thank you. The only doc for these signatures should not A) be in the src/ tree, and B) be listed under a function marked DEPRECATED. Even if the signatures aren't. 00:03
(not blaming you, btw)
plobsing completely agree 00:06
japhb Oh for heaven's sakes. The wiki page darbelo pointed to says to look at the docs for Parrot_ext_call, which don't include all the information in the deprecated Parrot_PCCINVOKE function.
darbelo japhb: Want to get even more riled up? 2.0 (in January) is our "Production Ready (TM)" release. 00:08
plobsing do we get to slap a big "enterprise ready" sticker on it? 00:09
darbelo I'm lobbying to name that release NCC-1701 00:10
:)
japhb I wonder if a parrot has ever figured prominently in a Star Trek (episode or movie) 00:11
darbelo memory-alpha.org/en/wiki/Parrot is the only relevant refernce I can dig up 00:12
diakopter seems to recall a parrot in a holodeck scene
cotto_work I have to question if 2.0 is really going to be "production ready" in anything other than name.
japhb OK, seriously? Someone took the time to change all my glut_callbacks code to use a function that is listed in the wiki page as "will be deprecated in the future"? What exactly does that even mean? You don't tell people something will be deprecated someday, you just deprecate it! 00:15
japhb is getting frustrated 00:16
time to talk a little walk
darbelo I think it won't be. Knowing waht I know about the dark corners of the codebase (that we are required to keep in place by the dep policy), I can't honestly say 2.0 will be 'production ready'. 00:17
cotto_work It'll be fun to kick up that anthill at the next #ps.
although knowing PHP, code quality isn't really an indicator of production-readiness 00:18
diakopter be sure to bring along the pesticide
darbelo But I've had that opinion of other 'production ready' projects before. So remember to take my opinions with a grain of salt. 00:19
chromatic I wouldn't mind rethinking the deprecation policy, especially in light of what actual users (Partcl, Rakudo) want and need and expect.
cotto_work +a lot
though not terribly likely 00:20
chromatic Look at it this way: the six month cycle is, in theory, to make life easier for distributors.
If Rakudo and Partcl and Pynie and Cardinal and whomever find that cycle onerous, what's the point of making Parrot easy for distributors to distribute? 00:21
Parrot won't be that useful for *any* language developer, and shipping Parrot won't be that interesting for any distributor.
cotto_work We'd get a lot of ground covered if we could kill the darkpan argument.
or at least weaken it 00:22
darbelo The problem, as I see it, with the deprecation policy, is that keeping some interfaces *unchanged* for six monts gets in the way of *improving parrot*.
plobsing does the deprecation policy cover itself?
darbelo I'd *LOVE* to add the deprecation policy to DEPRECATED.pod, and work out a new one for 2.1 00:23
But I don't think that'll fly too far.
chromatic The PaFo board has discussed some ideas around this, but the decision fizzled out last week. Ping any of Coke, allison, particle, or jhorwitz for their thoughts. 00:26
japhb OK, bak. OK. Since I don't want to do this more times than I need to, what function *should* I be using to call into PIR from C in the glut_callbacks.c code? (In other words, what is the "least deprecated", most likely to last version of the available APIs)? 00:31
From the Wiki page, I suspect it is Parrot_ext_call, but I'm not sure that is normative. 00:32
dalek rrot: r42495 | mikehh++ | trunk/ext/nqp-rx/src/stage0 (2 files):
fix codetest failures - pod syntax (and a load of trailing spaces)
 00:36
allison japhb: see trac.parrot.org/parrot/wiki/PCCMigrationNotes 00:40
00:40 mikehh joined
japhb allison: is that normative? 00:40
allison japhb: and yes, it's Parrot_ext_call
japhb OK, I'll work through that then 00:41
allison yes, normative, definitive, and official
japhb allison, All of the various docs should really be merged into something in docs/
allison (it's a magical function that does everything all the old functions used to do, and more)
japhb (and by "various docs" I include the embedded POD in the relevant C files)
allison japhb: there's docs/embed.pod, which is our official API definition, but it's lacking in descriptions for the API functions 00:42
00:43 Whiteknight joined
japhb allison, nod. And it's also lacking the signature syntax and type info 00:43
allison japhb: indeed. I added some of that to the C-level documentation for Parrot_ext_call, but the documentation needs to spread 00:44
japhb: I don't recall if I added type info to Parrot_ext_call, but if not it can be found in src/call/pcc.c under Parrot_PCCINVOKE 00:45
(and needs to be copied/migrated over to Parrot_ext_call)
japhb allison, unfortunately, the deprecated function Parrot_PCCINVOKE seems to have better info than Parrot_ext_call.
jinx
allison :)
00:47 tetragon joined 00:50 Whiteknight joined
japhb allison, if I make a TT for reviewing the src/ tree and moving any general docs/comments out of the src/ tree and into the proper places in docs/, any chance can we make this a release blocker for 2.0? 00:55
allison japhb: not sure what you mean. A lot of docs/ is generated from src/
japhb: but we could definitely use a doc cleanup run for 2.0 (it's the primary milestone for 1.9) 00:56
japhb My mental image of "production ready" does not include "must read raw parrot sources to find out how to use the official embedding/NCI interfaces"
GOOD
Good choice of milestones.
allison japhb: well, no, but the reason we use POD in src/ is so it can be extracted, and read as text or html 00:57
japhb: (but, that does imply keeping the POD in src/ up to date and readable by ordinary humans)
japhb sure, but that's the wrong place to put stuff that's public API. It's arguable appropriate for private API, but not public.
and that too.
s/arguable/arguably/ 00:58
allison japhb: there is definitely also room for introductory documentation separate from src/
japhb Let me rephrase: If it's subject to the deprecation policy, the primary documentation for it should not be within src/, but instead in docs/
oh, quite.
Distributors should be able to ship a parrot-docs package, using just the contents of docs/, and know that that will be enough for anyone using parrot-bin or parrot-dev packages. 01:02
dalek rrot: r42496 | japhb++ | trunk/config/gen/opengl.pm:
[OpenGL] glut_callbacks.c PCC calls fixed, but not yet modernized
japhb Only people hacking on parrot itself should need to read docs in src/
mikehh pmichaud: ping 01:06
plobsing japhb: that's a good start, but might not do it
japhb huh what?
plobsing japhb: for example, you have to cast short's to INTVAL
otherwise they won't get sign-extended
and negative values will fail badly 01:07
(I just dealt with this issue)
japhb Oh you've gotta be kidding me ...
japhb puts on a frowny face
plobsing japhb: dead serious. I hate it too.
:-| <-- extremely serious face 01:08
japhb Do I only have to worry about INTVAL casting, or do I need to FLOATVAL cast on float and double as well? 01:11
plobsing just cast everything 01:12
(can't hurt)
japhb Any special casts needed for char* and void*?
plobsing I don't know the spec on Cs varargs. probably not.
a pointer is a pointer right? 01:14
Whiteknight pointer == pointer, indeed
dalek rrot: r42497 | japhb++ | trunk/config/gen/opengl.pm:
[OpenGL] glut_callbacks.c: Now with Parrot_ext_call()
 01:15
japhb Whiteknight, god I hope so.
Although, aren't there some platforms where void*, char*, and (func*)() aren't the same size? 01:16
plobsing (func*)() and void*, possibly.
never heard of data pointers being different sizes though
japhb I should have said "ALL the same size" 01:18
01:28 kid51 joined
dalek TT #1263 closed by jkeenan++: Determine configuration value for 'platform' prior to gen::platform 01:39
japhb OK, As of 42498, OpenGL should be back as working as it ever was. 01:49
And now, pizza ...
Whiteknight chromatic: ping
dalek rrot: r42498 | japhb++ | trunk/config/gen/opengl.pm:
[OpenGL] glut_callbacks.c: Now with more fixed. (INTVAL casts for arguments to get proper sign extension)
 01:51
Whiteknight (more fixed)++ 01:57
01:58 sjohnson left
dalek rrot-linear-algebra: 09dc92e | Whiteknight++ | src/pmc/ (3 files):
change 'does matrix' to 'provides matrix'. The former doesn't work, the later does
 01:59
purl dalek: that doesn't look right
rrot-linear-algebra: 2f135a0 | Whiteknight++ | t/10-nummatrix.t:
add tests for NumMatrix2D that it does matrix properly
rrot-linear-algebra: af24681 | Whiteknight++ | t/1 (2 files):
add test files and initial tests for the PMCMatrix2D and CharMatrix2D PMC types
02:10 mikehh joined
dalek tracwiki: v9 | dukeleto++ | AllHackathons 02:11
tracwiki: trac.parrot.org/parrot/wiki/AllHac...ction=diff
mikehh messages 02:17
dalek tracwiki: v10 | dukeleto++ | AllHackathons
tracwiki: trac.parrot.org/parrot/wiki/AllHac...ction=diff
tracwiki: v11 | dukeleto++ | AllHackathons
tracwiki: trac.parrot.org/parrot/wiki/AllHac...ction=diff
rrot-linear-algebra: 83a8805 | Whiteknight++ | (2 files):
fix row-at-a-time string sets in CharMatrix2D. Add tests for it
 02:28
rrot-linear-algebra: 399d254 | Whiteknight++ | t/12-charmatrix.t:
add tests for a few more vtables for CharMatrix2D. I'm not happy with indexing. Needs rethinking
rrot: r42499 | jkeenan++ | branches/configtests/lib/Parrot/Configure/Options/Test/Prepare.pm:
Delete commented-out code.
 02:30
rrot: r42500 | jkeenan++ | branches/configtests/lib/Parrot/Configure/Step/Test.pm:
Improve documentation.
rrot: r42501 | jkeenan++ | branches/configtests/config/init/defaults.pm:
Comment out all set_p5 calls.
 02:37
nopaste "kid51" at 71.246.122.62 pasted "Very recent diff of 'configtests' branch with respect to trunk (branch point r42236)" (3910 lines) at nopaste.snit.ch/18681
kid51 If anyone is interested in reviewing the state of the 'configtests' branch (trac.parrot.org/parrot/ticket/1189), please see that paste. 02:38
02:38 bubaflub joined
dalek rrot-linear-algebra: dce3c00 | Whiteknight++ | src/pmc/charmatrix2d.pmc:
fix initialize_from_mixed_array to actually work. Not performant, but it works
 02:52
tracwiki: v42 | cotto++ | ParrotQuotes 03:00
tracwiki: Austin figures out the magic words.
tracwiki: trac.parrot.org/parrot/wiki/Parrot...ction=diff
03:35 janus joined 03:36 pdcawley_ joined
pmichaud mikehh: pong 04:38
04:42 tetragon_ joined 04:45 Eevee joined
dalek p-rx: 4199a71 | pmichaud++ | src/ (2 files):
More defactoring in preparation for <abc=def>.
 04:59
p-rx: 7aa09d7 | pmichaud++ | src/ (6 files):
Make final rules of alias sequences explicitly non-capturing.
p-rx: 690ec78 | pmichaud++ | src/Regex/P6Regex/Actions.pm:
Remove <aliased> flag.
p-rx: ffc125b | pmichaud++ | src/Regex/P6Regex/Grammar.pm:
Conversion #1 to new alias syntax.
p-rx: d78a6dc | pmichaud++ | src/HLL/Grammar.pm:
More <abc=.def> conversion.
p-rx: 42e5cce | pmichaud++ | src/NQP/Grammar.pm:
More <abc=.def> fixes.
p-rx: e5b5874 | pmichaud++ | src/NQP/Grammar.pm:
Last bit of subrule alias changes to align with new S05.
p-rx: cffd7d9 | pmichaud++ | src/stage0/ (3 files):
Update bootstrap.
p-rx: eb37eeb | pmichaud++ | :
Merge branch 'master' of git@github.com:perl6/nqp-rx

  \tsrc/NQP/Grammar.pm
p-rx: 0561bea | pmichaud++ | src/stage0/ (3 files):
Update bootstrap.
cotto looks like a good night for hackery 05:00
pmichaud long day of hackery, actually 05:03
japhb Is dalek no longer reporting nqp-rx to #perl6?
pmichaud it floods out
japhb oh ah
pmichaud 04:59 -!- dalek [i=dalek@feather.perl6.nl] has quit [Excess Flood]
japhb needs to be taught to slow down 05:04
pmichaud there was a semi-significant change to S05 this morning, so I've spent most of the day catching NQP up with the change (which was tricky because NQP is self-hosted)
japhb Looked like a good change, though.
pmichaud yeah, I'm in agreement with it. 05:05
cotto is there any pir code for running parrot as a separate process and checking its output a la Test::More 05:14
japhb cotto, you mean a 'qx'? 05:15
Plumage has code for that ... you'd just need to load parrot's standard Test::More emulation as well. 05:16
Look in plumage's src/lib/Glue.pir
cotto yes
great. I hate reinventing wheels. 05:17
Previously uninvented wheels are fair game, of course.
japhb As far as build & test go, between fperrad, dukeleto, and I, we seem to be producing quite a few wheels free for the taking. 05:18
cotto japhb++ fperrad++ and dukeleto++ then 05:19
japhb We're still getting them into an "easy to reuse" state. fperrad++ seems best at this at the moment.
cotto Yeah. The code I want to write would most logically live in Parrot's repo. 05:20
It'd be a shame to have to reuse by copy/paste.
but reading from a pipe is a fairly small amount of code 05:21
japhb nodnod
dalek rrot-plumage: 7ff475c | japhb++ | :
[TESTS] sanity.t: Fix for Win32, I hope
 05:22
japhb We will get there (easy reuse without copy/paste), but right now Plumage has heavy churn on other tasks.
cotto This will involve many interesting problems.
sure
it's just at that point in its development
japhb nodnod
05:23 theory joined 05:26 eternaleye joined
dalek p-rx: e46c70d | pmichaud++ | src/NQP/Grammar.pm:
Convert statement_control:return to prefix:return .
 05:28
p-rx: 53fc7ad | pmichaud++ | src/NQP/ (2 files):
Convert statement_control:make to prefix:make .
p-rx: 710c598 | pmichaud++ | src/NQP/ (2 files):
Remove obsolete regexes and methods for 'return'/'make'.
p-rx: e574f6f | pmichaud++ | src/stage0/ (3 files):
Update bootstraps.
pmichaud I've been thinking about adding qx{} directly into nqp 05:32
so that it doesn't require a library 05:33
05:33 darbelo joined
japhb Won't hear any complaints from me. :-) 05:33
Now if someone would make it so parrot pipes weren't so insecure ... 05:34
darbelo I'm afraid to ask what are we using for pipes now... 05:35
japhb concatenate command and args into a string, open with mode string 'rp', which launches a shell to parse apart the concatenated command/args string. 05:36
darbelo But
ww 05:37
cotto pmichaud, +1
darbelo japhb: So, we do the most insecure thing available. Ugh. 05:38
japhb darbelo, yup.
Though to be fair, apparently it was difficult enough getting even that much to work that we should be grateful we can do it at all. :-/ 05:39
darbelo Ah, yes. The joys of portability. 05:40
japhb This has to have been solved by the Perl 5 guys ... I'd be surprised if they are intentionally doing the insecure thing on Windows. 05:41
Tene Pipes are on my TODO list. 05:42
darbelo Perl 5 is on the OpenBSD base system. It wouldn't be there if it did the isecure thing.
Tene My Parrot priorities were completely screwed up by getting sick.
I can look at pipes this weekend. 05:43
japhb darbelo, there's a lot of code in Perl 5 that is right on anything vaguely unixish, but ends up waving its hands REALLY HARD on Windows. The APIs just aren't there.
darbelo Tene: Good. All of my available hate is tied up in pmc2c and freeze_pmc.c
japhb ENOTENOUGHHATELEFT
dalek rrot: r42502 | darbelo++ | trunk/include/parrot (2 files):
Remove trailing comas from enumerations so that picky compilers shut up about them.
japhb nice commit message 05:44
Tene Oh. Um, I don't Windows.
I haven't Windows'd since 2003.
cotto How can I tell if a process opened via a open cmd, "rp" exited with non-zero status? 05:45
japhb cotto: I don't know that you can. 05:46
At least, it was not obvious to me. :-(
darbelo I just got my hands on a PC with icc installed. And we don't fail as much as I expected to. 05:47
japhb excellent
cotto I was expecting that. 05:48
darbelo We do get a whole lot of annoying warnings. But we fail only a few tests.
cotto Fix them warnings.
darbelo I just did (a few) 05:49
And others come from generated code which I'm just plain not touching.
cotto What would courage wolf say? 05:50
japhb CHEW IT
darbelo courage wolf? 05:51
cotto Somebody hasn't been reading whiteknight++'s blog. 05:52
darbelo googles it.
cotto www.lmgtfy.com/?q=courage+wolf 05:53
darbelo DON'T PROMISE
THREATEN
(courage wolf)++
cotto Is there any way to run a command and get its return value? 05:55
japhb cotto, as I currently understand it, you can either run it as a pipe xor get the exit status. 05:57
(in parrot) 05:58
06:01 chromatic joined
cotto Hmmm. Having both would be really nice. 06:01
dalek rrot: r42503 | pmichaud++ | failed to fetch changeset:
Update nqp-rx with latest S05 semantics for <a=b>.
nopaste "darbelo" at 190.3.155.112 pasted "Test results for Linux amd64 intel cc 11.1" (11 lines) at nopaste.snit.ch/18683 06:02
darbelo And the one t/op/sprintf.t failed test is for "-0" 06:03
chromatic Better than I expected. 06:04
darbelo Yeah. I was expecting it to go up in flames, but it just bitched about unused variables and unreachable code. 06:05
Ah, and "remark #2259: non-pointer conversion from "int" to "short" may lose significant bits" came up a *lot* 06:06
chromatic That's just snark.
darbelo I like the way it makes every warning report five lines long. two of them blank and one with only one non-whitespace character in it. 06:08
cotto That's because they're important. 06:09
darbelo So important they had to make them scroll out of sight five times as fast? 06:10
cotto chromatic, what would you expect to be a sane interface for capturing a child process' output and exit status? 06:11
darbelo cotto: Not implemented?
cotto you can do either or neither but not both 06:12
chromatic Why not both?
cotto There doesn't appear to be anything in Parrot that lets you.
(I'd be happy if i were missing something though) 06:13
chromatic Do we need to add waitpid() support then?
cotto Heh. src/io/unix.c uses waitpid and doesn't do anything with the status. 06:14
chromatic There's step one: wire it up.
Step two: cry, because it's going to be ugly anywhere non-POSIX. 06:15
cotto yeah
(my employer)--
chromatic, what should the pir code look like once everything's implemented? 06:16
would it be a method on a filehandle?
chromatic That depends how you get the results from the handle. 06:17
I'd return two results, the former the result string and the latter the return value.
... but I haven't checked how new PCC handles binding fewer positional returns than get returned. 06:18
darbelo Does anything like this shou up in whitekinght's IO tasklist?
I think he's been colectig ideas for improving IO there. 06:19
dalek rrot-plumage: 7fe3c74 | japhb++ | :
Make use of statement modifier if
cotto IOTasklist doesn't mention anything like this. 06:21
darbelo karma my employer 06:23
purl my employer has karma of -1
darbelo cotto: Then you get to do it all youself :)
cotto I'm happy to do it. I just don't want to make a mess of it. 06:24
chromatic Mark it as EXPERIMENTAL USE AT YOUR OWN RISK IF IT CHANGES YOU MAY HAVE TO UPDATE IT and it's all good.
Or port the Maybe monad to Parrot and we use that. 06:25
Hmm.
Asynchronous monadic lazy futures.
cotto s/i/o/ makes for some interesting typos when working with pipes 06:28
darbelo Hmm what's up with all of the "#skip perl5-specific extension " tests in t/op/sprintf.t ? 06:31
Are we going to implement them or did it get cargo culted in? 06:32
cotto This doesn't look like it'll be too bad. 06:40
anyone able to test a windows patch? 06:42
06:46 pdcawley_ joined
dukeleto 'ello 07:09
cotto hi 07:20
Is running "false" a reliable cross-platform to get a non-zero exit status? 07:21
i.e. is there such a beast on windows? 07:22
dukeleto cotto: why not exit 1 ? 07:23
cotto will that work?
dukeleto cotto: are you talking about /bin/false?
cotto yes
dukeleto cotto: which platforms?
purl which platforms are 8.3 an issue on, anyway?
cotto all
dukeleto cotto: it won't work on windoze 07:24
cotto I'd like to write just one test that'll work on window and *nix
figured
dukeleto cotto: what is the test?
purl testing 123 abc 4red 12blue 9green
cotto testing whether the exit status of a process is captured correctly
nopaste "cotto" at 74.61.2.46 pasted "store filehandle pipe exit status" (138 lines) at nopaste.snit.ch/18684 07:25
cotto I can always make a small pir program to run "exit 1", but that's excessive for something this simple. 07:28
dukeleto cotto: parrot has tests for that already, iirc 07:31
cotto: where are you adding this test?
cotto t/pmc/filehandle.t 07:32
purl t/pmc/filehandle.t is the new io impl tests
cotto It's a method on that pmc
dalek a: 3aedd24 | dukeleto++ | kea-compiler:
Basic example of a Kea grammar with the new NQP
 07:39
cotto how clever. The existing tests rely on Perl. 07:40
I suppose I could run ./parrot -o nonexistent_file 07:44
That'd definitely get me a non-zero exit status.
I'd just have to make sure the filename didn't actually exist. 07:45
As a counterpoint, if I'm thinking about doing that I probably need to go to bed. 07:50
07:52 TiMBuS joined
japhb cotto: the low level code is prone to throwing exceptions or even just exit(1) from parrot, if you try that sort of thing. 07:52
And with that, I'm off to bed, too.
cotto yeah 07:57
I'll just use perl for now. 07:58
nopaste "cotto" at 74.61.2.46 pasted "store filehandle pipe exit status, needs win32 testing" (173 lines) at nopaste.snit.ch/18685 08:00
cotto If any windows users are around, please build with that patch and run t/pmc/filehandle.t. 08:03
That's what I thought 08:13
night
08:18 fperrad joined 08:54 darbelo joined 08:55 iblechbot joined 09:17 MoC joined
dalek rrot: r42504 | barney++ | trunk/NEWS:
Added NEWS from the last two days.
 09:19
09:19 barney joined, MoC joined 09:20 MoC joined 09:46 davidfetter joined 10:33 Casper joined
mikehh pmichaud: missed your pong - in la-la land then 10:41
pmichaud: ping
dalek rrot: r42505 | mikehh++ | trunk/ext/nqp-rx/src/stage0 (2 files):
fix codetest failures - Pod syntax [unmatched =cut at the end]
 10:52
mikehh msg pmichaud: the Pod syntax failure is due to the unmatched =cut at the end - also a whole lot of trailing spaces that don't seem to be tested for 10:53
purl Message for pmichaud stored.
mikehh msg pmichaud: podchecker doesn't seem to care about that 10:54
purl Message for pmichaud stored.
11:04 davidfetter joined 11:17 davidfetter joined, bacek joined 11:22 mikehh joined 11:26 mokurai1 joined
bacek o hai 11:27
mikehh hoedy 11:28
that was supposed to be howdy :-} 11:29
bacek :) 11:35
darbelo hi everyone. 11:36
purl Howdy, darbelo, you fantastic person you.
darbelo It make me nervous to see purl being nice.
11:53 ilbot2 joined
moderator Parrot 1.7.0 "African Grey" is out! | Testing hackathon November 14 and 15 -- trac.parrot.org/parrot/wiki/Opcode...gHackathon | find out what's up with the slice opcode | Latest modified TT's: icanhaz.com/parrotbugs
11:57 integral joined 11:58 ttbot joined
fperrad darbelo, r42462 breaks the build of pirc, 12:00
problem with the prototype of emit_pbc_key() in bcgen.h
the headerizer doesn't handle "struct key", it handles only typedef
12:01 mikehh joined
fperrad ping bacek 12:04
purl I can't find bacek in the DNS.
bacek fperrad, pong 12:05
Did I broke win32 again?
fperrad bacek, could you help me with lists.parrot.org/pipermail/parrot-d...03314.html
WmlsBytecode.pmc needs a code review
bacek fperrad, looking 12:06
12:07 payload joined
bacek bacek@icering:~/src/wmlscript$ wmlsc hello.wmls 12:09
#hello.wmls:4#Error: Redefinition - main.
#hello.wmls:5#Error: block statement expected.
Oops. Double paste. 12:10
(gdb) p fp 12:13
$2 = 0xb67ba264 <Address 0xb67ba264 out of bounds>
hmmm 12:14
fperrad, nopaste.snit.ch/18687 looks veeery suspicious. 12:25
fperrad, looks like Parrot_Class_instantiate clobbering "fp" pointer somehow... 12:27
Bah! I'm an idiot. Are Parrot STRINGs in compacting pool? 12:28
darbelo bacek: yep.
(on the pool, not the idiot)
bacek fperrad, You are storing pointer to STRING guts. During Class.instantiate GC triggered, string moved, you've got stale pointer. 12:29
darbelo, I suspect both cases are true
darbelo bacek: I only know one answer. If I knwe them all I'd be rich by now :)
bacek darbelo, You don't need "all", "two" is enough :) 12:31
fperrad, check nopaste.snit.ch/18688, "fp" is "stale" because of string compacting.
fperrad bacek, Parrot_string_cstring() gives a pointer that moves
so I need to do my own copy of the C string
darbelo fperrad: You should probably switch to Parrot_str_to_cstring() 12:32
bacek fperrad, yes.
(you have to free it after)
darbelo Use Parrot_free_cstring() when you are done with the string. 12:33
It'll cost you a memcpy(), but the string will stay put this time ;) 12:34
12:34 cosimo joined
darbelo fperrad: The only problem with the headerizer is the (lack of) typedef-ing? 12:36
bacek fperrad, or you can try to switch to string's iterators (I'm not sure about GC properly handling static GCable object)
12:41 davidfetter joined
fperrad darbelo, bcgen.h contains 2 prototypes of emit_pbc_key() 12:43
- the generated is wrong
- the hand written is good
darbelo Ah. That makes more sense. I'm unable to build pirc right now (damm picky linker) but I'll try to fix it. 12:45
fperrad bacek, I try nopaste.snit.ch/18689 without success 12:53
bacek fperrad, free fp, not bytecode 12:56
bah... I'm definitely an idiot... Ignore me
12:57 he_ joined
bacek Hmm... Works for me it I remove Parrot_free_cstring 12:57
12:58 jan joined, preflex joined
bacek pmc2c-- 12:58
12:58 he joined 12:59 cognominal joined, joeri joined
darbelo Shouldn't "RETURN(PMC *NULL);" be returning PMCNULL instead of a NULL pointer? 13:00
bacek fperrad, name clash... In WmlsBytecode.load "bytecode" is arg. But you want to free "static bytecode"
fperrad bacek, in my patch, I rename argument 'bytecode' to 'str' 13:02
bacek fperrad, ah... sorry.
fperrad, than.... It passed on my box. 13:03
nopaste.snit.ch/18690
fperrad bacek, the next step is 13:05
$ parrot wmlsi.pir hello.wmlsc main
bacek bacek@icering:~/src/wmlscript$ parrot wmlsi.pir hello.wmlsc main
Hello World!
still works :)
fperrad bacek, the whole test suite 13:06
$ make test
bacek Can't locate Parrot/Test/Harness.pm... 13:08
Is it required to be inside parrot/languages?
fperrad bacek, yes
bacek moving 13:09
13:09 nopaste joined 13:10 PacoLinux joined
bacek t/libstring.t (Wstat: 256 Tests: 24 Failed: 1) 13:11
Failed test: 19
Non-zero exit status: 1
everything else passed
nopaste "bacek" at 114.73.159.244 pasted "t/libstring.t ........ 15/24 #" (13 lines) at nopaste.snit.ch/18691
bacek Ho! nopaste is back! 13:12
fperrad, nopaste.snit.ch/18691 is only one (strange) failure I've got
fperrad bacek, it's ok 13:13
bacek fperrad, looks like main problem is solved :) 13:15
fperrad bacek, yes, I apply my patch on linux, it works. But not on Windows. 13:16
bacek fperrad, "Windows"? Never heard about this :) (What's the problem on windows?) 13:18
13:19 kid51 joined
darbelo The problem with "Windows" is that people keep pretending it's an operating system... 13:19
fperrad bacek, I have in a Windows in a VMware (my real box run with Ubuntu) only because some people breaks the build 13:21
darbelo Those pesky buildbreakers! 13:22
bacek darbelo, actually windows kernel isn't so bad. But whole userland is crap (including compilers, etc)
fperrad so I use mingw/gcc & gmake (not MSVC & nmake) 13:23
bacek fperrad, you can ask mj41++ to extend TapTinder to cover HLLs as well.
fperrad what is TapTinder ? 13:24
purl hmmm... TapTinder is software development tool - taptinder.org . For Parrot project running on tt.perl6.cz/ and reporting build failures to #parrot channel as ttbot.
bacek wow... useful purl 13:25
darbelo We should all enshine this moment, I'm sure we won't be seeing anything like it anytime soon. 13:27
bacek useful purl is quite purl 13:33
bacek falling asleep. 13:34
G'Night all
darbelo Good morning! Sleep well. 13:35
nopaste "darbelo" at 190.192.220.13 pasted "fperrad: Does this work fix pirc for you?" (15 lines) at nopaste.snit.ch/18692
13:36 Whiteknight joined
darbelo fperrad: Is nopaste.snit.ch/18692 enough to fix pirc? 13:37
fperrad darbelo, yes
darbelo Excelent. Now I need a way to make that permanent.
Whiteknight good morning 13:38
darbelo Morning Whiteknight 13:39
Whiteknight hello darbelo 13:42
darbelo Whiteknight: Do you know of any way to tell the headerizer to please stay away from a prticular funtion? 13:44
Whiteknight no 13:45
darbelo headerizer-- 13:55
dalek rrot: r42506 | darbelo++ | trunk/compilers/pirc/src/bcgen.h:
Remove incorrectly generated prototype for a function.
darbelo headerizer--
dalek p-rx: 5b31d0a | pmichaud++ | src/HLL/ (2 files):
Replace decimal and binary digit strings with <decint> and <binint>.
 14:29
p-rx: 453c3b5 | pmichaud++ | src/stage0/ (3 files):
Update bootstrap with new decint/binint.
rrot: r42507 | pmichaud++ | failed to fetch changeset:
[nqp]: New decint/binint, update tests.
 14:38
pmichaud I'm still seeing a failure in t/op/number.t on my platform. 14:39
t/op/number.t (Wstat: 0 Tests: 173 Failed: 1)
Failed test: 172
dalek p-rx: 7b584c0 | pmichaud++ | src/ (2 files):
Remove spurious "=cut" lines to make Parrot's pod checker happier (mikehh++).
 14:50
15:13 Whiteknight joined 15:17 Psyche^ joined 15:20 kid51 joined 15:38 cconstantine joined
dalek lscript: d8eb135 | fperrad++ | dynext/pmc/wmlsbytecode.pmc:
fix segfaults, thanks to bacek & darbelo
 15:40
cconstantine I just did an svn update (to r42507) and i"m getting a compilation failure in src/pmc/continuation.c 15:49
15:49 Essobi_ joined
cconstantine hmm, a fresh svn checkout built 15:58
darbelo cconstantine: there's been some heavy makefile hacking lately, you probably needed a 'make realclean' 15:59
cconstantine ah, I only did a realclean
I think the last time I updated was in the 1.6 timeframe
bah
I only did a 'clean' 16:00
darbelo Jeez. 1.6 is, like, the jurassic, dude. You need to update more often. 16:05
Whiteknight I think 1.6 is the version of Parrot my grandfather used 16:09
and his "computer" was a pile of rocks and leaves. I doubt parrot will even compile there anymore 16:10
dalek a: a2a604f | fperrad++ | t/metatable.t:
untodo tests (metamethod __call)
 16:13
darbelo Whiteknight: Sort of what happens with windows now an then. ;)
cconstantine yeah yeah... work started acting up, and I got the bad flue 16:15
16:16 sjn joined 16:17 payload joined
darbelo cconstantine: Anyway, good thing you dropped by. We are having testing hackathon today. 16:19
cconstantine I'm not sure Reason is up for testing quite yet... ;) 16:20
darbelo We'll break your build in all sorts of ways and the demand diagnostig information before fixing it.
cconstantine hehe
darbelo It'll be *fun*.
cconstantine I bet i can thwart many attempts to break my build by refusing to update 16:21
cotto Speaking of *fun*, can someone test a windows patch?
moderator Parrot 1.7.0 "African Grey" is out! | Testing hackathon November 14 and 15 -- trac.parrot.org/parrot/wiki/Opcode...gHackathon | Latest modified TT's: icanhaz.com/parrotbugs 16:21
cotto nopaste.snit.ch/18685 16:22
16:23 theory joined
dalek a: 6cb69ad | fperrad++ | (2 files):
fix segfault in bytecode translator
 16:24
16:24 iblechbot joined
darbelo I can test that it causes ho harm on OpenBSD, but you probably knew that already. 16:26
cconstantine OSX here if someone needs remote hands
cotto no harm in trying
but it's got some windows-specific code that I haven't been able to test 16:27
if t/pmc/filehandle.t passes, the patch is good
darbelo cconstantine: I've been bitten before by confusing OS X with an actual UNIX. You should probably test there as well. 16:28
cconstantine which patch?
cotto's? 16:29
darbelo cconstantine: Yeah, it touches src/io/unix.c in ways that should work on all systems pretending to be unix. But you can never be sure with OS X. 16:30
cconstantine running, nothing looks bad so far 16:31
t/pmc/filehandle............................ok \\n 1/18 skipped: no asynch calls yet
cotto looks good
thanks
cconstantine np 16:32
cotto cconstantine++
cconstantine uhh, I think I have a windows VM sitting around somewhere
what's the dev env like for windows?
darbelo Like hell, but with an uglier color scheme. 16:33
cotto cconstantine, you can install Strawberry Perl to get the tools to build the mingw version pretty easily.
cconstantine yeah, hell-like..
I'll see what I can do
cotto great 16:35
cconstantine strawberry perl installed, now what? 16:37
cotto get Parrot, open the nasty dos prompt, run perl Configure.pl, mingw32-make, prove t/pmc/filehandle.t 16:38
cconstantine so, install mingw?
cotto Strawberry comes with it. 16:39
cconstantine ah, useful
cotto alias++
cconstantine the dos prompt... it burns
cotto powershell also works fine if that's less painful
cconstantine I find doing windows things the windows way ends up better for me 16:40
cotto powershell is very windows
cconstantine so, it comes with mingw but not svn :( 16:41
cotto Yeah. And TortoiseSVN requires a restart.
cconstantine hates tortoisesvn
cotto You could probably just download a release.
cconstantine yeah, working on it 16:42
tortoise would be ok except for the background cache process that looks for updates
cotto yeah. It applies fine against the 1.7.0 release 16:44
feel free to use that
cconstantine svn checkout is in-progress 16:45
cotto ftp://ftp.parrot.org/pub/parrot/releases/devel/1.7.0/
ok
cconstantine as a sanity check I'm gonna build/test my checkout before doing your patch... make sure I didn't bork something up
cotto also, jk about it working. It depends on the pmc header refactor
sure 16:46
cconstantine this might take awhile...
cotto no worries
cconstantine this strawberry perl thing looks compelling
cotto It makes windows much less painful. 16:47
kid51 cotto: You've made the front page of the New York Times: www.nytimes.com/
16:48 allison joined
cotto There's no way I'd ever be confused with that guy. 16:48
kid51 Testing your patch on Darwin/PPC. 16:50
cconstantine is there anyway to just run that testsutie?
kid51 cconstantine what do you mean? 16:51
cconstantine instead of going through all of t/ just do t/pmc/filehandle
kid51 cotto's patch touches both configuration and source code files; so you have to reconfigure and rebuild 16:52
at that point, you can say: 'prove t/pmc/filehandle.t'
... but what if the code caused tests in other files to fail?
darbelo cotto: Say, how much pain does $dayjobos
kid51 u could get a good subset: make coretest
darbelo cause you, now you work at the evil empire.
;) 16:53
cotto cconstantine, you can run prove t/pmc/filehandle.t
cconstantine I'm getting a fair amount of ugly output running just the tests (no patch)
cotto darbelo, it's an interesting job. I'm in the Open Source Technology Center. 16:54
cconstantine my cats... they are trying to kill me
16:57 mikehh joined
cconstantine cotto: building your patch in windows 16:57
I've got a failure 16:58
darbelo cotto: Oh, I've heard about that. I think I even read an interview with the guy that started it. Sounded like a really interesting place. 16:59
nopaste "cconstantine" at 71.98.77.93 pasted "cotto's patch test results in winxp" (22 lines) at nopaste.snit.ch/18693 17:00
cotto At least it's consistent. jonathan++ got the same thing in #perl6 17:01
lemme dig and try to make a less broken patch
darbelo cotto: And while I'm being annoying, how would irclog.perlgeek.de/parrot/2009-11-13#i_1723717 affect your patch? 17:02
If we cared about VMS at all, of course. 17:03
cotto Heh. I can pretty easily test for non-zero instead of 1. 17:04
dalek rrot: r42508 | fperrad++ | trunk/tools/install/smoke.pl:
[install] add nqp-rx
 17:08
rrot: r42509 | fperrad++ | trunk/runtime/parrot/library/Configure.pir:
[library] s/output/spew/
 17:14
cotto cconstantine, can you try changing WAIT_FAILED to WAIT_OBJECT_0 in src/io/win32.c ? 17:15
cconstantine I have been removed from the box runing windows...
the local sports team is playing and my attendance was required at the TV
dukeleto cconstantine: sounds like a personal problem ... 17:16
cconstantine sorry :(
yeah
cotto ok
cconstantine is someone in #perl6 still able to help?
wait, you work at MS and don't have access to windows? 17:17
17:17 mikehh_ joined
cotto not at home 17:17
Technically I could go to work, but that's what the week is for.
cconstantine ah 17:18
17:20 Zak joined
dalek tracwiki: v5 | allison++ | AllisonTasklist 17:25
tracwiki: trac.parrot.org/parrot/wiki/Alliso...ction=diff
nopaste "cotto" at 74.61.2.46 pasted "updated patch: store filehandle pipe exit status, needs win32 testing" (185 lines) at nopaste.snit.ch/18694
cotto cconstantine, when you have the time, try that one 17:26
cconstantine k
cotto thanks
nopaste "cotto" at 74.61.2.46 pasted "updated patch v2: store filehandle pipe exit status, needs win32 testing" (185 lines) at nopaste.snit.ch/18695 17:28
cotto actually, that one. it only tests for non-zero exit status
dalek rrot: r42510 | fperrad++ | failed to fetch changeset:
[distutils] build dynpmc before dynops
 17:32
cotto I'm so glad I don't have to do win32 programming more often.
kid51 cotto: 'make test' all PASS with your patch applied at r 42507 17:37
That was the 18685 paste, not any later one 17:39
cotto kid51, on windows? 17:42
dalek rrot: r42511 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
[distutils] quote prove_exec when contains space
 17:52
17:58 payload joined
dalek rrot: r42512 | fperrad++ | trunk/tools/dev/mk_language_shell.pl:
[languages] add another way to test with prove
 17:59
18:06 mikehh joined 18:09 Whiteknight joined
kid51 cotto: No darwin/ppc 18:12
18:47 redbrain_ joined
dukeleto is attempting to recruit plparrot hackers at opensqlcamp 19:03
Whiteknight nice 19:05
get a lot of them
kid51 plparrot?
Whiteknight it's the postgres+parrot integration project 19:06
darbelo Yeah, we are pulling Postgres into parrot in order to capture their userbase. 19:07
kid51 purl, plparrot is the postgres+parrot integration project
purl OK, kid51.
kid51 purl, plparrot?
purl it has been said that plparrot is the postgres+parrot integration project
19:17 chromatic joined
Whiteknight anybody else hackathoning today on t/ops? 19:31
I just started, working on t/op/time.t
dukeleto purl, plparrot is also github.com/leto/plparrot
purl okay, dukeleto.
dukeleto how is the hackathon going? 19:34
mikehh I get a failure with t/op/comp.t with i386 g++ (4.4.1) with optimize but passes without --optimize and gcc and amd64 all flavours 19:35
cotto It'd be great if I could get a windows tester. 19:36
;)
dalek kudo: 7347ec0 | masak++ | src/parser/actions.pm:
[parser/actions.pm] aligned documentation with code

more sense.		 19:37
rrot: r42513 | whiteknight++ | trunk (3 files):
Convert most of t/op/time.t to PIR. Some functions rely on perl behavior (or some other external mechanism to verify current time) and could not be updated
 19:38
dukeleto mikehh: that is interesting. can you make a ticket for that? 19:39
dalek rrot: r42514 | whiteknight++ | trunk/t/op/time.t:
add test coverage for sleep_n opcode
 19:41
Whiteknight decodetime and decodelocaltime ops return an Array, not a FixedIntegerArray like one would expect 19:52
or even a ResizableIntegerArray
mikehh dukeleto: om it 19:53
on
interesting slip there :-} 19:54
dalek TT #1270 created by whiteknight++: decodetime and decodelocaltime return Array 19:56
TT #1187 closed by mikehh++: t/src/warnings fails test 2 on optimized build with Ubuntu 9.10 amd64 19:59
Whiteknight ...and now we have at least some coverage for all time-related ops 20:05
dalek rrot: r42515 | whiteknight++ | trunk/t/op/time.t:
Add much more test coverage for time-related ops. Up to 29 tests now from 7 this morning
 20:08
rrot: r42516 | whiteknight++ | trunk/src/utils.c:
add quick note to the code about TT #1270
rrot: r42517 | whiteknight++ | trunk/t/op/spawnw.t:
add note about TT #1271 to the t/op/spawnw.t file
 20:15
TT #1271 created by whiteknight++: t/op/spawnw.t relies heavily on perl5 20:16
Whiteknight the pushaction/pushmark/popmark opcodes are really pretty useless 20:18
and it would be almost trivial to create an alternative to those with more features
20:19 Zak joined
dalek rkdown: 260f2ad | fperrad++ | (4 files):
conversion nqp to nqp-rx
 20:27
fperrad ping pmichaud 20:32
purl I can't find pmichaud in the DNS.
dukeleto which ops have the worst coverage?
Whiteknight I have no idea 20:33
cotto What about creating a small exit_with_status.pir program to make it trivially easy to test that exit status detection works? 20:36
dalek rrot: r42518 | mikehh++ | trunk/t/op/comp.t:
convert t/op/comp.t to pir - patch from bubaflub++ (TT #1269)
 20:38
purl I don't know how to convert t/op/comp.t to pir - patch from bubaflub++ (TT #1269).
mikehh I still can't build with g++ - src/glut_callbacks.c - errors (18 x) - error: ‘Parrot_ext_call’ was not declared in this scope 20:39
dukeleto cotto: that sounds reasonable to me
cotto tools/build perhaps? 20:41
I'll do that. 20:44
dalek rrot: r42519 | mikehh++ | trunk/MANIFEST:
re-generate MANIFEST
 20:45
20:46 mokurai joined
dalek rrot: r42520 | mikehh++ | trunk/t/op/time_old.t:
set svn properties
 20:48
pmichaud fperrad: pong 20:50
dalek rkdown: a589a18 | fperrad++ | Makefile.in:
clean up
purl i heard clean up was a breeze
fperrad pmichaud, in language Markdown (github.com/fperrad/markdown)
I convert with success nqp to nqp-rx
But I fail to convert pge to nqp-rx
cconstantine I'm back, kind of 20:51
20:51 kid51 joined
pmichaud nqp-rx has its own regex engine, separate from pge 20:51
nopaste "fperrad" at 79.81.86.137 pasted "Markdown (from pmichaud)" (24 lines) at nopaste.snit.ch/18698
pmichaud pge is on its way to obsolesence, unless someone else wants to maintain it
"Confused at ..." generally means a syntax error or something the compiler doesn't quite understand yet 20:53
20:53 bacek joined
pmichaud fperrad: I think that nqp-rx might not like the comment prior to the | in Title 20:54
mikehh pmichaud: I keep getting codetest failures when you change ext/nqp-rx
pmichaud # TitleSingle
mikehh: imo, we shouldn't be running codetest on things in ext/
mikehh: if you're editing the files in ext/nqp-rx, it won't do much good -- they get overwritten by newer versions
cconstantine cotto: still need a test of nopaste.snit.ch/18694 ? 20:55
pmichaud (and most of that code is "generated" code)
I did fix the sources so that we won't end up with the solo "=cut" lines, though.
fperrad pmichaud, without comment => error at same line
mikehh you need to remove the redundant =cut at the end - podchecker don't mind but t/codingstd/pod_syntax.t does 20:56
pmichaud mikehh: I did that.
I haven't updated anything in parrot yet for that, though -- I'm waiting for a more significant change before doing that
mikehh you also have a lot of trailing spaces - it doesen't test for that though - but my editor removes them 20:57
pmichaud the trailing spaces are in generated code, I think
cotto cconstantine, I'd love one
pmichaud if there's trailing spaces in anything outside of the src/stage0 dir, I'll fix those, though.
but the files in src/stage0 are generated files
and it's a pain to get the generator to avoid the trailing spaces
mikehh run it through a script to remove them :-} 20:58
pmichaud mikehh: if someone wants to do that, they're welcome to do that too
I just don't see the trailing space issue as being worth dealing with, and especially not for generated code 20:59
cotto yeah
20:59 Eevee joined
kid51 Could someone who opposes 'make codetest' on /ext open a TT to that effect? 21:00
mikehh we probably shouldn't be doing codetest on /ext - will look into that
I'll open a ticket on it 21:02
pmichaud fperrad: looks like nqp-rx doesn't like '\\'' yet 21:03
fperrad: maybe \\x27 would work
(yes, it's an nqp-rx bug, I'll look into fixing it today) 21:04
cconstantine cotto: bad news, failed 16/18 21:06
wait a minute... I might be an idiot 21:07
yup, pretty sure 'perl Configure.pl' isn't 'make'
dalek TT #1269 closed by mikehh++: [PATCH] convert t/op/comp.t to PIR 21:09
moritz is it new that parrot build requires perldoc to be installed? 21:12
21:13 AndyA joined
nopaste "cconstantine" at 71.98.77.93 pasted "cotto's patch on winxp" (24 lines) at nopaste.snit.ch/18699 21:13
cotto That makes me sad. 21:15
cconstantine sorry
fperrad pmichaud, \\x27 works, comments are correctly handled,
my grammar.pg is parsed, but now I have a new error
nopaste "fperrad" at 79.81.86.137 pasted "Markdown (from pmichaud)" (35 lines) at nopaste.snit.ch/18700
moritz oh, and is the parallel build known to be broken again? 21:16
cotto cconstantine, thanks. I really hate fumbling around in the dark with the windows api like this.
cconstantine I feel your pain
21:16 kid51 joined
cconstantine do you need more output? 21:18
21:21 mikehh_ joined
cotto cconstantine, no 21:23
how long will you be around?
cconstantine not sure, hoping it'll be measured in hours
bacek morning, good morning
cotto good morning bacek
kid51 amuses $self by installing Catalyst::Devel -- which, like Catalyst, requires half of CPAN! 21:26
bacek kid51, next version of Catalyst will be Moose-based. And will require second half of CPAN :) 21:28
kid51 But I believe I'm using Moose-based version already.
Am working my way thru MST's book -- and Chap 1 was all about Moose. 21:29
21:29 japhb joined
kid51 Thing is .. I thought I had most of these modules installed already. 21:29
bacek kid51, 5.8?
kid51 5.10.1 21:30
cconstantine cotto: I'm behind a dsl line running winxp in a virtualbox vm... anyway I can make this easier for you?
bacek kid51, sorry, Catalyst version is 5.8xx?
kid51 But, yeah, maybe I installed them when I was at 5.8.
bacek, don't know about catalyst version.
how do I check that?
bacek kid51, it is 5.8... I didn't know that it was already released. 21:31
kid51, latest version on CPAN is 5.80013 21:32
kid51 our $VERSION = '5.80011';
What I meant earlier was, that perhaps I installed these modules back when I was running Perl 5.8. 21:33
I installed Catalyst 2 months ago ... but apparently need to install Catalyst::Devel as well to do anything useful.
For example, this is requiring Module::Build -- which I have had installed for years. 21:34
21:36 Austin joined
cotto cconstantine, being willing to test is a huge help. 21:36
cconstantine okie 21:37
nopaste "cotto" at 74.61.2.46 pasted "updated patch v3: store filehandle pipe exit status, needs win32 testing" (183 lines) at nopaste.snit.ch/18701
cotto actually, I can simplify that some 21:39
(but if you're already testing, don't stop
)
cconstantine passes 21:41
nopaste "cotto" at 74.61.2.46 pasted "updated patch v4: store filehandle pipe exit status, needs win32 testing" (180 lines) at nopaste.snit.ch/18702
cotto That should pass to, but it won't hurt to test.
HAPPY 21:42
VERY HAPPY
Austin JOY
cotto gives hugs to everyone, making some uncomfortable
cconstantine w00t
cotto cconstantine++ for the tests 21:43
cconstantine w00t w00t
cotto if that patch works, I'll be applying it presently
cconstantine v4 is currently compiling
Austin Is there any way to have assertion failures in Parrot dump the PIR stack?
21:44 mikehh joined
cconstantine passes 21:44
cotto Austin, modify Parrot_confess. That'd be a nice feature. 21:45
cconstantine, thanks a bunch
cconstantine np, glad I could help
Austin Cotto: I take it that means "no" ?
cotto Austin, there's a PIR op that makes a backtrace. You could probably call whatever that calls in Parrot_confess to get the desired behavior. 21:47
but you'd have to modify the code to get that 21:48
Austin :)
And I'm already looking forward to the huge tax I'm going to have to pay next week.
cconstantine cotto: are we done with windows?
21:49 Essobi joined
cotto cconstantine, actually, could you change 1 to 5 in the t/pmc/filehandle.t tests for exit status and see if they still pass? 21:51
cconstantine 1 to 5? 21:52
cotto nm'
Yes, we're done
cconstantine okie
dukeleto so, what are people working on? 21:58
bacek world domination? 21:59
purl somebody said world domination was evillemur.blacklightmedia.com/world-dom.jpg or at www.lexie.pp.se/ or at pes.eunet.cz/billy.jpg or Seig heil Microsoft
cotto What I finished was a building block for enabling profiling runcore testing.
no, world domination is the eventual goal
purl okay, cotto.
dalek rrot: r42521 | cotto++ | trunk (6 files):
[pmc] make the FileHandle PMC keep track and expose the exit code of a child process run as a pipe
 22:00
bacek cotto: what do you think about resurrecting opsc/pmcc? Based on new shiny nqp-rx? 22:01
cotto very yes 22:03
I'd be hacking on that if I didn't feel that the profiler needed more work.
you might want to wait until pmichaud adds mmd support to nqp-rx, though. 22:05
(he said that it was part of his plan once he figured out the syntax iirc) 22:06
bacek ah, yes. MMD... 22:07
cotto or someone else could add it 22:18
22:20 petdance joined
dukeleto i am going to do a lightning talk about plparrot at opensqlcamp, sometime soon. 22:20
cotto Lightning talks are great. 22:23
chromatic Sanity check: someone tell me what's the deal with t/op/lexicals_32.pir 22:25
Specifically: why is it okay to call a sub by name if that sub has :anon. 22:26
Austin If it has a name, it isn't really anonymous, is it? 22:27
chromatic That's my thinking.
Austin But that's pretty irritating, since it implies that the sub can't appear in a backtrace. 22:28
:anon should probably be changed.
GeJ Good morning everyone. 22:29
chromatic Changed how, Austin?
cotto is :anon necessary there?
Austin chromatic: By figuring out what the ultimate objective of :anon is/was, and renaming it to reflect that.
dalek TT #1272 created by mikehh++: Problems with codetest on files in ext/ directory (should ext/ be included ... 22:35
TT #1273 created by Austin_Hastings++: Provide a parrot 'backtrace' with assertion failures. 22:45
dukeleto leto.net/dukeleto.pl/2009/11/parrot...abase.html
chromatic I *think* :subid is a better approach than :anon-only-after-compilation. 22:46
22:58 kthakore joined
kthakore chromatic: hi! 22:58
chromatic: check out sdl.perl.org 22:59
chromatic: new site!
purl new site is unreadable
kthakore chromatic: still in the works
Austin That's amazing, dude. 23:00
kthakore Austin: the site?
purl the site is probably done enough for Allison to do announcing
Austin But I have no idea what SDL is.
kthakore Austin: Ah ... maybe a better blurb on the Site is needed
Austin: thanks
Austin I gather it's "Simple Directmedia Library", instead of some of the many other SDL acronyms. But what is that?
kthakore Austin: SDL Perl is bindings to that library. Which allows lowlevel access to Video, Audio and other systems 23:01
Austin Ahh.
Okay.
You should probably put that in the intro.
kthakore Austin: used mostly for games but PDL is taking an interest
Austin: I will 23:02
Austin: thanks
Austin Not to get repetitious, but what's PDL?
etc.
kthakore Perl Data Language 23:03
Austin: I don't know the details but
Austin: its Matlab in Perl
Austin okay.
kthakore Austin: sourceforge.net/apps/mediawiki/pdl/...n_with_SDL
pmichaud (more) 23:05
my understanding of :anon is that it simply suppresses adding a sub into the namespace
Coke 23:19
23:19 Coke joined
Coke purl, anything for me? 23:19
purl i don't know, coke
kthakore ah he is gone 23:20
...
Coke he how? 23:21
er, who?
kthakore Coke: Austin. 23:22
purl it has been said that austin is nice. or a city in Texas.
kthakore Coke: I was showing him sdl.perl.org
Coke: what do you think? 23:25
purl I think kthakore should try flossing more often!
kthakore ??
Coke purl, quiet down
purl Coke: i'm not following you...
kthakore Coke: oh ... sdl.perl.org 23:26
Coke kthakore: I default to http these days. =-)
kthakore: sdl is outside my area of expertise, sorry. 23:27
kthakore Coke: oh I meant the site? 23:28
Coke: how does it look? It's new.
calculus any body with OS X, that can test a patch for the build system? 10.5 needs some adjustments to the cflags (like 10.6) as well as support staged installs better, pastebin.com/m32336486
and hopefully apply it to svn 23:29
allison mentioned testing it on 10.6 but needed someone with 10.4 and what better way than during the hackathon 23:30
:)
kthakore calculus: mac users are so hard to find ... >.>
calculus: I like your name btw
calculus kthakore: thanks
kthakore calculus: you are my favorite subject 23:31
calculus++ just for the name
calculus kthakore: it is mine as well :)
calculus blushes
kthakore yay
kthakore approaches carefully 23:32
kthakore hugs calculus and runs
yay!
lol
geek love is so hard to find :p
Coke calculus: I have 10.4; is there a ticket? 23:41
23:41 bacek joined
kid51 calculus: Please open a TT for this. Since in the past we've had a lot of problems with this section of the code, we'll have to track review of this patch in a ticket. 23:43
Also, config/gen/makefiles/root.in has changed since you pulled that patch, so it applies with an offset. Consider preparing new patch after svn up
Also, please do: perl Configure.pl --test 23:44
... because that's giving me:
t/steps/auto/snprintf-01.t .................. 1/33 Use of uninitialized value in concatenation (.) or string at config/init/hints/darwin.pm line 70.
... which I did not have previously.
(That may not be strictly due to your patch, but it does need investigation; hence: TT) 23:45
dukeleto languages? 23:49
purl languages is trac.parrot.org/parrot/wiki/Languages
cconstantine calculus: you still need a 10.5 tester?
is there a parrot equivalant of python's pprint? 23:50
dukeleto cconstantine: _dumper from PGE ? 23:52
cconstantine I want to see what exactly I have when I'm executing actions from my grammar 23:53
calculus Coke, kid51: no ticket, let me update to make a current patch with no offsets 23:55
oh kid51 left 23:56
well, Use of uninitialized value in concatenation (.) or string at config/init/hints/darwin.pm line 70. seems very odd to me 23:57
I get it too with the patch, without the patch applied it errors out much earlier because it fails to build
what is wrong with line 70, pastebin.com/m9bfd049 (with patch applied) 23:58
cconstantine: more testing wouldn't hurt 23:59
cconstantine it wouldn't apply for me, I think I did it wrong

Powered by Cro and the Raku® Programming Language.

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

Thank you!