|
Parrot 4.6.0 "Wild Parrots of Telegraph Hill" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC Set by moderator on 18 July 2012. |
|||
|
00:15
kid51 joined
02:12
kid51 joined
02:36
Psyche^ joined
05:38
fperrad joined
06:33
awwaiid joined
|
|||
| dalek | kudo/toqast: 633f401 | jnthn++ | src/ (2 files): Fix up handling of failed symbol lookups, thus eliminating a couple of reamining NPMCAs and fixing a couple more tests. |
10:48 | |
|
11:00
JimmyZ joined
|
|||
| dalek | p/toqast: c21542a | jnthn++ | src/QAST/ (2 files): Restore line number annotation emission. Fixes callframe.t. |
11:10 | |
|
11:11
brrt joined
|
|||
| dalek | kudo/toqast: d0dc2d3 | moritz++ | src/core/IO/Socket/INET.pm: explicitly set encoding on sockets |
11:16 | |
| kudo/toqast: e09fbe6 | moritz++ | src/core/IO/Socket/INET.pm: Merge remote branch 'remotes/origin/parrot-iocleanup1' into nom |
|||
| kudo/toqast: cd31947 | moritz++ | src/core/List.pm: fix List.tree it got broken by the last API change of MapIter.new |
|||
| kudo/toqast: 6db12d6 | moritz++ | src/core/IO/Socket/INET.pm: Revert "Merge remote branch 'remotes/origin/parrot-iocleanup1' into nom" This reverts commit e09fbe68707707004d1111dba7b0b6fb9e100541, reversing changes made to 5b56cfcf9286e70b9a77a73369b80ad98ad10786. |
|||
| kudo/toqast: 3ac8c77 | moritz++ | src/core/List.pm: fix another misuse of MapIter.new, felher++ |
|||
| kudo/toqast: acf9778 | jnthn++ | src/core/List.pm: Merge branch 'nom' into toqast |
|||
| kudo/toqast: 7f3a7b2 | jnthn++ | src/Perl6/Actions.pm: Fix detection of existing CATCH in a try, fixing one of the broken tests in catch.t. |
11:33 | ||
|
11:34
kid51 joined
11:38
JimmyZ_ joined
12:08
kid51 joined
|
|||
| dalek | kudo/toqast: e084b07 | jnthn++ | src/Perl6/Actions.pm: Fix last remaining failures in catch.t. |
12:20 | |
|
12:54
PacoAir joined
|
|||
| moritz | oh hai | 12:54 | |
| spawnw returns an int that you have to bit-shift 8 bits to get to the return value of the program | 12:55 | ||
| JimmyZ | hi | ||
| moritz | what's in those shifted-away bits? | ||
| and can they used to get more information about the error? | |||
| like, distinguishing a failed spawn from a command with non-zero return value? | |||
| Coke | moritz: I suspect it's using the same model as perl5. | 13:33 | |
| moritz | Coke: I suspect so too, but I don't want to base my code on suspicions | 13:35 | |
| Coke | the opcode just says "system dependant". | 13:50 | |
| the generic version seems to return the status value from waitpid | 13:51 | ||
| ... or the return value of execlp | 13:53 | ||
| Coke stops reading half the source and just points to "platform/generic/exec.c" | 13:54 | ||
|
13:56
bluescreen joined
|
|||
| moritz | ok, if forking worked, the result of waitpid is returned | 14:04 | |
|
14:30
JimmyZ_ joined
14:44
dmalcolm joined
15:53
lucian joined
16:00
darbelo joined,
jashwanth joined
16:48
whiteknight joined
|
|||
| whiteknight | good afternoon, #parrot | 16:48 | |
| benabik | o/ whiteknight | ||
| rurban | hi #parrot | ||
| benabik | o/ rurban, #parrot | ||
| whiteknight | hello benabik, rurban | ||
| dalek | p/toqast: f5d0884 | jnthn++ | src/QAST/Compiler.nqp: Don't emit variable lookup code when the variable is in void context. Cuts down on code generated for declarations, but also fixes implicit %_ handling. |
16:58 | |
| benabik | I love it when I look at a failing test and go "how did this code ever work"? | 16:59 | |
| Even better is when I can make it work somewhere else. | 17:00 | ||
| Hm. Editing commit messages in vim with subjects of "vim: did stuff" leads to interesting errors. | 17:06 | ||
| Oh! I know why this wasn't triggered by my manual testing. I had dead code at the beginning of the PBC. | 17:07 | ||
| dalek | CT: 24ba6e9 | benabik++ | t/ (2 files): t/common: Load Decompile For maximum DRY (or is that minimum RY?) |
17:09 | |
| CT: c3d555e | benabik++ | src/PACT/Packfile/Decompile.winxed: Packfile.Decompile: use $load directive It's clearer than writing my own init function, and matches the style I'm using for tests. |
|||
| CT: c725ca9 | benabik++ | t/ (2 files): t/common: Central compile function Takes in PIR, outputs a Packfile. Going to be doing this a lot so would rather not write those three lines over and over again. |
|||
| CT: d05e70f | benabik++ | src/disasm.winxed: disasm: Use Winxed $directives |
|||
| CT: 5377db3 | benabik++ | src/PACT/Packfile/Decompile.winxed: Decompile: Don't try to pop things off a null sub It doesn't work very well. Managed to avoid this one in manual testing by having an :immediate sub in there so there was dead code before the first sub. First bug caught by writing tests. How exciting. |
17:10 | ||
| whiteknight | wowsers! | 17:11 | |
| benabik | ? | ||
| whiteknight | lots of commits | ||
| I like commits | |||
| I CRAVE THEM | |||
| our preciouses | |||
| benabik | whiteknight: I've been updating the vim highlighting for winxed too, don't know if you saw that one. (It seems not to be connected to dalek) | 17:12 | |
| whiteknight | I haven't seen that. I wonder how we can connect it to dalek | ||
| benabik | Huh. The most basic PIR file has 4 strings and 2 PMCs | 17:14 | |
|
17:16
kjs joined
|
|||
| whiteknight | that few? | 17:20 | |
| benabik | I'm not sure why there's an empty string, although I follow the other three. | 17:21 | |
| dalek | kudo/toqast: eb13c5a | jnthn++ | src/Perl6/Actions.pm: Fix an accidental violation of the QAST::Stmt usage rules when a block becomes a hash. |
17:22 | |
| benabik | Huh. Not empty string, null string? | 17:29 | |
| whiteknight | whatev | 17:34 | |
| benabik | Oh. Maybe that's the UUID. | 17:38 | |
| whiteknight | the UUID shoudl be part of the header, I think | 17:39 | |
| not one of the string constants | |||
| benabik | Oh. | ||
| whiteknight | although, we don't really use it anyway | 17:42 | |
| benabik | Yeah. | ||
| whiteknight | That's something we're definitely going to want at some point, I think | ||
| but, low on the priority list until we get security | 17:44 | ||
| benabik | huh. There's no way to get the name of an oplib out of its PMC? | ||
| whiteknight | there isn | 17:45 | |
| 't? | |||
| add it | |||
| benabik | Oh. I get a hash, not an RPA of them. | 17:46 | |
| whiteknight | okay | ||
| benabik | Yay tests exposing bad planning. | 17:49 | |
| alvis++ | 18:21 | ||
| whiteknight | alvis++ indeed! (but wait, why are you giving him karma, specifically?) | 18:22 | |
| not like we need a reason, of course :( | |||
| ...:) | |||
| benabik | Reporting on #ps. Remembering that today is #ps. Doing useful things to report on #ps. :-D | 18:23 | |
| whiteknight | ....damnit! | ||
| thanks for the reminder | |||
| benabik | #ps in 67m | ||
| whiteknight | I used to have some sort of calendar reminder set up for #ps, but that went away and I can never remember it on my own | 18:24 | |
| Coke | there's a google calendar, "comp.lang.parrot" | 18:29 | |
| dalek | CT: a74de0c | benabik++ | src/PACT/Packfile.winxed: Packfile: dump oplibs |
18:42 | |
| CT: 048e1d2 | benabik++ | src/PACT/Packfile.winxed: Packfile: oplibs are stored in hashes, not RPAs |
|||
| CT: acc68f9 | benabik++ | t/ (2 files): t/common: Add decompile function |
|||
| CT: 7d6e739 | benabik++ | t/common.winxed: t/common: PACT uses call_bits frequently |
|||
| CT: c1f27b1 | benabik++ | / (2 files): Test Decompiling a (basically) empty sub This is in t/02-decompile because I realized I probably want some tests for the basic packfile data structures in 01-packfile/ |
|||
|
18:56
schmooster joined
|
|||
| benabik | msg whiteknight Oh. My winxed highlighting has been in my own fork. Sending pull request.. | 18:59 | |
| aloha | OK. I'll deliver the message. | ||
| dalek | kudo/nom: a4eacdf | moritz++ | t/spectest.data: remove ucfirst/lcfirst tests they are gone or going |
19:14 | |
| benabik | #ps in 15? | 19:15 | |
| cotto | yup | 19:22 | |
| benabik | o/ Coke | ||
| Errrr... | 19:23 | ||
| cotto too | |||
| cotto | I get myself confused with him all the time. | 19:25 | |
| Coke | Just the other year I worked for Microsoft. | 20:04 | |
| benabik | Coke: That sounds like the setup to a joke, but I don't know the straight line. | 20:05 | |
| Coke | just that cotto used to work there. | 20:07 | |
| I myself am now a GE drone. | |||
| rurban: asn? | 20:12 | ||
| er, asan? | 20:13 | ||
| rurban | address sanitizer: code.google.com/p/address-sanitizer/ | ||
| Coke | aaaaah, thanks. | 20:14 | |
| rurban | perldoc perlhacktips | 20:15 | |
|
20:21
tobym joined
20:22
sivoais joined
|
|||
| dalek | p: cff22c2 | moritz++ | src/PAST/NQP.pir: add nqp::tclc op |
20:39 | |
| kudo/nom: 4fe23e7 | moritz++ | / (5 files): implement tclc |
20:40 | ||
| kudo/toqast: a9af262 | jnthn++ | src/ops/perl6.ops: A couple of native type handling tweaks. |
20:51 | ||
| kudo/toqast: 3cc9d76 | jnthn++ | src/Perl6/Optimizer.pm: Re-instate compile time resolution of multiple dispatches, where we have sufficient type info. |
|||
| p/toqast: 0555208 | jnthn++ | src/QAST/Operations.nqp: Be sure that invocants are boxed objects. |
|||
| kudo/toqast: 03dfa84 | jnthn++ | src/Perl6/Optimizer.pm: Unbust line number reporting in errors generated by l'optimizer. |
21:00 | ||
| kudo/toqast: ee26e27 | jnthn++ | src/Perl6/Actions.pm: Fix the busted macro stuff. Long story short. :lexical(0) has gone away, and this is how its replaced. |
21:30 | ||
| kudo/toqast: 5a3d1f9 | jnthn++ | src/Perl6/World.pm: This is probably needed for serializing macro stuff (though maybe not sufficient); adding it since I spotted it. |
|||
| Heuristic branch merge: pushed 21 commits to nqp/toqast by jnthn | 21:47 | ||
| kudo/toqast: a4eacdf | moritz++ | t/spectest.data: remove ucfirst/lcfirst tests they are gone or going |
22:04 | ||
| kudo/toqast: 4fe23e7 | moritz++ | / (5 files): implement tclc |
|||
| kudo/toqast: c23523d | jnthn++ | / (4 files): Merge branch 'nom' into toqast |
|||
| kudo/toqast: a9dbdef | jnthn++ | src/Perl6/Ops.pm: Add tclc op (could go in NQP also; feel free to move). |
|||
| p/toqast: 8a9b631 | jnthn++ | src/QAST/Block.nqp: Toss unsued attribute on QAST::Block. |
22:05 | ||