Parrot 4.5.0 "Buff-faced Pygmy Parrot" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 19 June 2012.
00:42 kid51 joined 01:06 seiferteric joined 02:29 bacek_at_work joined 05:29 kuku joined 05:43 aloha joined 07:08 bacek_at_work joined 07:17 brrt joined
dalek rrot/m0: 12b007e | jimmy++ | docs/pdds/draft/pdd32_m0.pod:
fix typo
07:21
rrot/whiteknight/io_cleanup1: 413ae84 | jimmy++ | src/ (3 files):
fixed builds
07:25 travis-ci joined
travis-ci [travis-ci] parrot/parrot#464 (m0 - 12b007e : Jimmy Zhuo): The build was broken. 07:25
[travis-ci] Change view : github.com/parrot/parrot/compare/0...b007e2787c
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/1698857
07:25 travis-ci left
dalek kudo/nom: 53fc0ae | pmichaud++ | src/core/Bool.pm:
Restore [^^] and [xor].
08:13
08:19 lucian joined
dalek Heuristic branch merge: pushed 17 commits to rakudo/ex-hardening by moritz 09:08
kudo/nom: b9a332b | moritz++ | src/core/Backtrace.pm:
improve robustness of internal error reporting
09:26
kudo/nom: b68a884 | moritz++ | src/core/Backtrace.pm:
be more tight about :D in invocant
kudo/nom: 57d251f | moritz++ | src/core/Backtrace.pm:
fix previous commit
kudo/nom: 2fdf4bc | moritz++ | / (11 files):
Merge branch 'nom' into ex-hardening
kudo/nom: 50774a3 | moritz++ | src/core/Backtrace.pm:
[Backtrace] it is fine to generate empty backtraces

this seems to happen only on the REPL. Fixes RT #113776
kudo/nom: 094e41b | moritz++ | src/core/Backtrace.pm:
Merge branch 'ex-hardening' into nom
09:42 brrt joined 10:11 kjs joined 11:19 mtk joined 11:36 JimmyZ joined
JimmyZ kjs: hello 11:37
kjs: failures/funcall.m1 wfm
12:55 PacoAir joined 12:57 kjs joined
JimmyZ hello: kjs 13:01
kjs hi JimmyZ 13:05
JimmyZ kjs: failures/funcall.m1 wfm
kjs JimmyZ: did you uncomment all calls/
JimmyZ kjs: nope 13:06
kjs try that;-)
JimmyZ failures/funcall.m1 have comments?
kjs it says choose 1 of 3 or something
all 3 calls wont work 13:07
JimmyZ kjs: that's not in funcall.m1
kjs ./run_m1.sh: line 14: 36711 Segmentation fault ./m0 $filename.m0b
Coke so the segfault is really in m0 ? 13:08
JimmyZ kjs: you meant struct.m1
kjs Coke: well if you generate wrong instruction sequences then it's not M0s fault 13:09
so it's M1 really
i suspect
JimmyZ: yes youre right. thats essentially same failure
13:10 Psyche^ joined
JimmyZ ;) 13:10
Coke kjs: m0 should never segfault. 13:17
so, 2 separate errors, maybe, but segfaultsbad. 13:18
I'd verify that with cotto/whiteknight/dukeleto, but that seems a reasonable goal.
kjs Coke: I would agree, long term. At mo it's very basic, and error handling like that isn't' implemented I think 13:19
but yes good point
thing is, it's a lot of bit twiddling that's going on.
1 small mistake and it blows up
atrodo Coke: You say that, but unless it's been fixed, parrot is pretty easy to segfault with a malformed bycode data
13:40 brrt joined
brrt hello #parrot 13:42
Coke atrodo: and that's a bug. 13:58
which should be fixed.
atrodo Coke> irclog.perlgeek.de/parrot/2011-07-28#i_4181699 14:04
dalek p: 5a8707b | pmichaud++ | src/HLL/Compiler.pm:
Add -V/--verbose-config option to HLL::Compiler (RT #62462).
14:13
Coke atrodo: OK. if there isn't a ticket open for that, there's a problem. 14:18
atrodo Coke> github.com/parrot/parrot/issues/498 ...
14:20 jsut joined
Coke atrodo: ... I get the feeling you're arguing with me, but that I'm winning. ;) 14:25
note that that's PASM, not PIR. there were notes in the IRC log that it was probably doable in PIR, but I don't see a PIR sample. 14:26
atrodo Coke: I wouldn't say arguing, just, informing. I think it's a very serious bug 14:27
Coke: But, besides that, it's bytecode. It can be easily hand crafted without PASM or PIR
dalek kudo/nom: 185f8fc | pmichaud++ | tools/build/NQP_REVISION:
Bump NQP_REVISION to get -V option (RT #62462).
Coke atrodo: then please do so and open a ticket. 14:28
I would argue that if you pass hand-crafted PBC to parrot and it segfaults, that it should still be fixed, but that's a much lower priority than if you can segfault it one of parrot's own tools.
JimmyZ well, I'm waiting m0b 14:31
and then change the name to pbc
14:33 JimmyZ joined 14:44 contingencyplan joined
dalek nxed: d84049d | NotFound++ | winxedst0.cpp:
a bunch of refactors and a few improvements in annotations in stage 0
15:05
15:13 Khisanth joined 15:25 dmalcolm joined 15:28 JimmyZ joined 15:45 ingy joined 16:02 brrt left
dukeleto kid51++ # congrats on the white camel! 16:29
16:52 jashwanth joined 17:26 darbelo joined
dalek : 348a52e | kjs++ | src/ (2 files):
clean up some code, and add a check for use of uninitialized variables.
17:35
17:35 kjs joined
dalek : 11e86f0 | kjs++ | src/ (3 files):
refactor warning() and reuse it in gencode.
17:42
: 18c025c | kjs++ | src/ (3 files):
object node now also has a line number for more precise error handling.
17:47
jashwanth hello dukeleto 17:50
17:51 darbelo joined
dalek : cf11e41 | kjs++ | / (2 files):
add status to TODO.minor cleanup/reformatting.
17:52
17:58 birdwindupbird joined
dalek nxed: 2527d53 | NotFound++ | winxedst (2 files):
return token instead of string in Builtin.getname
18:52
nxed: 192aba6 | NotFound++ | winxedst0.cpp:
missing annotation
nxed: 83e2c8f | NotFound++ | winxedst0.cpp:
annotate return in stage 0
nxed: 2a7802c | NotFound++ | winxedst0.cpp:
annotate argument list preparation in stage 0
benabik ~~ 18:59
19:03 brrt joined
brrt good evening #parrot 19:03
19:13 tuxit joined 19:24 bluescreen joined
dalek p/toqast: 4c86f84 | jnthn++ | src/QAST/Node.nqp:
$!name can't be str if we put it here, it turns out. May be better to move it into various other nodes where it can be, since it's just Regex with specialer needs.
19:33
p/toqast: 20a37a9 | jnthn++ | src/Q (3 files):
Various fixes so we can use native types on QAST::Regex attributes.
19:45 alvis joined
dalek kudo/toqast: 841e783 | jnthn++ | src/QPerl6/ (2 files):
Translate method calls to QAST. Again, a few complicated ones left over (please don't 'help' by changing the leftovers!)
19:52
brrt lazily, i ask if there is a simple way - a builtin, an opcode perhaps - to find the presence of a single character in a string in winxed 19:59
i can probably figure this out by myself, but it is hard :-)
benabik brrt: indexof(haystack, needle) 20:02
Coke parrot$ perldoc src/ops/string.ops - look for "search" 20:03
brrt thanks 20:12
handy stuff, i never realised you could run perldoc on that 20:14
20:29 tuxit joined 20:31 autark joined
brrt another question, winxed and shift? 20:45
dalek kudo/toqast: 27f9453 | jnthn++ | src/QPerl6/ (2 files):
Translate a bunch of calls to QAST.
20:46
brrt ${ shift, foo, bar }; works
but, ugly
benabik brrt: shift_<type>(array)
e.g. shift_int, shift_var
NotFound brrt: there are some ugly builtins. 20:47
Yeah, that.
brrt oh, that works
thanks
if it'd be welcome I'd like to write up my experiences with winxed 20:48
NotFound brrt: more than welcome!
brrt good, will do then 20:50
NotFound But I'll need someone to write ugly things. They say a language is not popular if it lacks a big bunch of haters.
(paradigmatic examples: cobol, pascal, C, java...) 20:51
brrt good point 20:55
there are plenty ugly things
but - as far as i'm concerned - most of them have to do with the hilarious mess that is vtable versus method
versus opcode
21:14 brambles joined, perlite joined
dalek p/toqast: 702c1b7 | jnthn++ | / (2 files):
Add a QAST::CompUnit, which will be involved in SC handling and a few other things we have one of per compilation unit.
21:45
21:55 brambles joined 22:23 autark joined 22:33 whiteknight joined
whiteknight Good afternoon, #parrot 22:35
Apparently a new C standard, C11 is in the works
one more modernizing development that Microsoft isn't going to support and Parrot won't be able to use
22:49 brambles joined
dalek rrot/whiteknight/io_cleanup1: b4b92b9 | Whiteknight++ | / (3 files):
Add in a function Parrot_io_get_standard_piohandle to replace the older Parrot_io_get_std_handle
23:28
rrot/whiteknight/io_cleanup1: 5eee9ba | Whiteknight++ | src/ (10 files):
Several build fixes for g++. kid51++ for the error report
rrot/whiteknight/io_cleanup1: dc11d0d | Whiteknight++ | src/ (4 files):
Merge branch 'whiteknight/io_cleanup1' of github.com:parrot/parrot into whiteknight/io_cleanup1
whiteknight Thanks to commits from kid51++ and jimmyZ++, we're only failing the codetest about function docs now 23:31
and it appears several people have been able to build NQP and Rakudo on the branch 23:32
We need a tool to automatically insert documentation stubs for functions 23:40
23:41 kjs joined, kid51 joined
whiteknight unless there already is such a thing and I just don't know about it 23:42
dalek rrot/whiteknight/io_cleanup1: e85091e | Whiteknight++ | src/platform/ (2 files):
Re-format some function-level docs so that the test recognizes that they exist
23:45
rrot/whiteknight/io_cleanup1: d06de9b | Whiteknight++ | src/io/userhandle.c:
Small doc fix for userhandle.c
whiteknight I wonder how hard it would be to update headerizer to automatically jam in empty documentation stubs with TODO notes? 23:47
23:49 jlaire joined
kid51 I would support that if the creators of those functions were sent an email once a day reminding the authors to write the docs. :-) 23:49
whiteknight headerizer already spits out warnings when it finds a function without docs. 23:50
it could just spit out a bigger warning
we could update the c_function_docs.t test to find these auto-generated TODO notes and fail 23:51
23:58 kid51_ joined