|
Parrot 3.10.0 "Apple Pi" | parrot.org | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC Set by moderator on 2 December 2011. |
|||
| dalek | nxed/version_1_5: fd7c18c | NotFound++ | winxedst (2 files): set version to 1.5.0 |
00:07 | |
| nxed/version_1_5: fcb0189 | NotFound++ | NEWS: update NEWS for 1.5.0 |
00:09 | ||
| nxed/version_1_5: a3d99c7 | NotFound++ | pir/winxed_compiler.pir: update installable compiler |
00:10 | ||
| nxed: 1e73802 | NotFound++ | NEWS: update NEWS for 1.5.0 |
00:15 | ||
| rrot: 3417b80 | NotFound++ | ext/winxed/compiler.pir: update winxed snapshot to release number 1.5.0 |
00:16 | ||
| NotFound | cotto: no external confirmation but I've built Rosella without problems, so it seems it's ready for release. | 00:18 | |
| dalek | nxed: eb72050 | NotFound++ | winxed.winxed: improve check for --version available in non installed driver |
00:34 | |
| nxed: 4cb7680 | NotFound++ | winxedst1.winxed: drop remains of --version from stage 1 |
|||
| nxed: 3bad01a | NotFound++ | winxedst2.winxed: jump version to 1.6 devel |
|||
|
00:36
jlaire joined
|
|||
| NotFound | msg cotto Winxed snapshot updated to 1.5.0, ready for release | 00:37 | |
| aloha | OK. I'll deliver the message. | ||
| cotto | NotFound: great | 00:40 | |
|
01:14
kid51 joined
|
|||
| nopaste | "kid51" at 192.168.1.3 pasted "t/examples/pir.t: One test failure" (46 lines) at nopaste.snit.ch/102496 | 01:16 | |
| dalek | rrot: 2691300 | jkeenan++ | src/pmc/filehandle.pmc: [codingstd] Remove trailing whitespace. |
01:17 | |
| cotto | how did I know it'd be make_hello_pbc | ||
| kid51 | I think that failure probably emerged within the last week. | 01:22 | |
| cotto | that's likely. There was a bytecode change. | ||
| benabik | Bytecode change? | 01:26 | |
| cotto | I thought so. I might be thinking of whiteknight++'s unmerged changes though. | 01:27 | |
| benabik | I haven't really been paying attention recently, so I have no idea. | ||
| cotto | It's probably an easy fix. | ||
| kid51 | This test did not fail in Smolder report 24850, so it's post Dec 09 | 01:30 | |
| cotto | kid51_at_dinner: thanks | ||
|
01:55
alvis joined
02:07
jsut joined
|
|||
| cotto | dukeleto: ping | 02:27 | |
| dukeleto: let's say I wanted to cut tonight's release on one of the beefy machines in the gcc compile farm. What would I need? | 02:28 | ||
| cotto will be back a bit later this evening | |||
|
02:29
Yuki`N joined
|
|||
| Yuki`N | What's good in the world of Parrot? | 02:30 | |
| kid51 | Assuming Smolder reports are correct, the failure in t/examples/pir.t emerged subsequent to 165e732 (Dec 18 2011). | 02:45 | |
| Wait, that's wrong. | 02:46 | ||
| The test was passing in smolder.parrot.org/app/projects/rep...ils/25061. RELEASE_3_10_0-163-g111b275. Dec 17. | 02:47 | ||
| So sometime since then. | |||
| benabik | 7 commits | 03:01 | |
| Error replicated on my machine, bisecting. | 03:09 | ||
| (Took a while to figure out how to make the linker happy without installing.) | |||
| Bisects to: 05a88d8 Remove FileHandle puts method and fix test and docs | 03:15 | ||
| kid51 | Given the error message, that sounds correct. | 03:18 | |
| cotto | ~~ | 03:21 | |
| benabik | kid51++ # testing and investigating | 03:24 | |
| dalek | rrot: 0a82f0a | benabik++ | examples/pir/make_hello_pbc.pir: Update example for change to FileHandle Updates make_hello_pbc.pir for the change in 05a88d8 Remove FileHandle puts method and fix test and docs |
||
| benabik | That should fix it. | 03:25 | |
| benabik knocks on wood. | |||
| benabik <3s git-bisect | 03:26 | ||
| kid51 | Yes, it works for me. | 03:28 | |
| linux/i386: make test PASS: with and without --optimize; all gcc and all g++ | 03:29 | ||
| That was at 2691300f884ff2b6db7081078b352e463610af56 | |||
| Testing now on darwin/ppc, but I think that will pass; smolder report in about an hour :-( | 03:30 | ||
|
03:36
Psyche^ joined
|
|||
| cotto | benabik++ | 03:41 | |
| seen dukeleto | 04:20 | ||
| aloha | dukeleto was last seen in #parrot 13 hours 21 mins ago saying "~~". | ||
|
04:53
davidfetter joined
05:09
alvis joined
05:43
baest joined
|
|||
| cotto | cpanm++ | 05:52 | |
| d'oh. parrotbug doesn't know about github | 06:31 | ||
|
07:07
davidfetter joined
07:40
zby_home joined
07:51
krunen joined
07:54
krunen joined
08:01
baest joined
08:33
jlaire joined
08:43
mj41 joined
|
|||
| cotto | alvis++ | 08:54 | |
|
09:14
mj41 joined
09:24
lucian joined
10:02
mj41 joined
10:36
snearch joined
|
|||
| dalek | rrot: 2db4772 | cotto++ | ChangeLog: initial overly-general version of changelog for 3.11.0 |
10:46 | |
|
11:59
not_gerd joined
|
|||
| not_gerd | hello, #parrot | 11:59 | |
|
12:03
jsut_ joined
|
|||
| not_gerd | bye, parrot | 12:21 | |
|
13:33
bluescreen joined,
whiteknight joined
14:28
whiteknight joined
|
|||
| whiteknight | good morning, #parrot | 14:32 | |
| dukeleto | ~~ | 15:25 | |
| cotto: looking for me? | |||
|
16:13
mj41 joined
17:24
dmalcolm joined
17:36
fperrad joined
17:58
fperrad joined
18:01
not_gerd joined
|
|||
| dalek | p: bceccc5 | moritz++ | src/QRegex/Cursor.nqp: implement <at($pos)> |
18:11 | |
| p: 3ae3f57 | moritz++ | t/qregex/rx_subrules: regex tests for <at> |
|||
| kudo/nom: 106783c | moritz++ | t/spectest.data: run roles + stubs test file |
18:15 | ||
| kudo/nom: dcfd439 | moritz++ | lib/Test.pm: give eval_lives_ok some diagnostics in the case of test failure |
|||
| kudo/nom: 89f54dc | moritz++ | tools/build/NQP_REVISION: bump to a NQP_REVISION that supports <at> in regexes |
|||
|
18:18
robertle joined
|
|||
| robertle | hi folks, I am trying to understand subroutine calls in VMs from a performance point of view | 18:20 | |
| I am looking at parrot as one example, and I understand that the underlying assembly would give you different choices | 18:21 | ||
| but I am wondering what would a real language like perl6 do? saveall/restoreall on each sub call? | |||
| sorear | there is no saveall | 18:22 | |
| and there is not much flexibility in the call conventions | |||
| I suspect you may be looking at some *very* old documentaton | 18:23 | ||
| robertle | there isn't? then I am looking at soe broken/outdated docs... | ||
| sorear | also, Parrot's subroutine calls are the slowest in any VM I know of, so you may not want to focus your studies here | ||
| robertle | ok, got some better docs now | ||
| another one I did look at was lua | 18:24 | ||
| basically what I am wondering is this: | |||
| in a real machine, e.g. x86, you need to save the registers somewhere, so the safest thing is to push them all onto teh stack and restore them later | 18:25 | ||
| now in a VM the register file (if it is register based of course) isn't anything special, just another block of memory | |||
| so instead of transfering the registers, you could just create a new register file on each subroutine invocation | |||
| which is of course more or less the same | |||
| now I am wondering if anyone does that, if not why, and if: whether there are any further tricks to it | 18:26 | ||
| benabik | I think Parrot's registers are part of the call context, so we do essentially do that. | 18:27 | |
| cotto | dukeleto, still around? | 18:28 | |
| robertle | ah ok! excellent! two of the things I am wondering are: | ||
| a) do you still generate a new "call context"/register file even if it is a tail call? | |||
| b) how do you shovel the arguments/return values from one register file to the next | 18:29 | ||
| benabik | All the arguments are bound together and passed from one context to another by the call/return opcodes… I think. I'm not as familiar with the deep internals. | 18:32 | |
| cotto hopes dukeleto's new employeer doesn't mind irc | |||
| benabik | I think the new context is created for a tail call, but the old one isn't kept around. | 18:33 | |
| Parrot subs have information on the number of registers they use, which may be different for different subs. | |||
| (I'm not 100% sure of all this, by the way.) | 18:34 | ||
|
18:45
Aleksandar joined
|
|||
| Aleksandar | Hey does someone can tell me which function is used in sub test for load_bytecode tests are in t/pmc/sub.t for example test starting on line 180 because i have some problems with some test from sub.t especially when needed to compile pasm file but cant find which i s the exact function which is called in this tests | 18:48 | |
| cotto | #ps in 40 | 18:50 | |
| dalek | rrot: c27c041 | cotto++ | ChangeLog: polish the changelog a bit |
18:52 | |
| cotto | Aleksandar, that test will call the load_bytecode_s op | ||
| that will call Parrot_load_bytecode | 18:53 | ||
| Aleksandar | Ok thanks i find it :) | 19:04 | |
| cotto | you're welcome | ||
| #ps time | 19:30 | ||
|
19:45
Coke joined
|
|||
| Coke | "add coket to active connections" anyone know what that comment means in "runtime/parrot/library/HTTP/Daemon.pir" ? | 19:45 | |
| tadzik | I assume it's "socket" | 19:46 | |
| dalek | website: cotto++ | Parrot 3.11.0 "Duct Tape" Released | 20:12 | |
| website: www.parrot.org/news/parrot-3.11.0-d...e-released | |||
| cotto | and the release is done | 20:20 | |
| dalek | rrot: a6c5dd2 | cotto++ | / (8 files): cut the 3.11.0 release "Duct Tape" |
||
| benabik | Stealth release. Didn't know it was coming. :-D | 20:21 | |
| cotto | benabik, funny you should say that. I only realized it on Saturday. | ||
| dalek | rrot: 618ded2 | cotto++ | docs/project/release_manager_guide.pod: add alvis++ as the 4.1.0 release manager |
||
| cotto | and docs.parrot.org is done | 20:26 | |
| sorear | What happened to naming parrot releases after psittacines? | 20:29 | |
| cotto | you haven't heard of the the rare Duct Tape Parrot? | 20:30 | |
| sorear | nope | ||
| benabik | duckbrand.com/Duck%20Tape%20Club/du...arrot.aspx | ||
| NotFound | It quacts like a duct. | 20:31 | |
| cotto | I thought I just made it up, but I guess it's a thing. | ||
| NotFound++ | |||
|
20:32
whiteknight joined
|
|||
| sorear | I also haven't heard of the Apple Pi parrot, the Archaeopteryx parrot, the Magrathea parrot, or the Wanda parrot | 20:32 | |
| whiteknight | Parrot releases are so common that no individual one is particularly important | ||
|
20:50
marcel_r joined
|
|||
| cotto | ~~ | 21:00 | |
| whiteknight | hello cotto | 21:05 | |
| cotto++ on the release. I forgot we were having a release today | 21:06 | ||
| cotto | whiteknight: you, me and benabik | 21:08 | |
| 4.0 should be harder to forget | |||
| btw, do you have the tuits to take that one? | |||
| whiteknight | yeah, I think I can | ||
| pencil me in | 21:09 | ||
| dalek | itor: ce2360e | NotFound++ | examples/pokedit.winxed: keep focus and set extended WM type in alert window |
21:16 | |
|
21:36
bluescreen joined
|
|||
| not_gerd | Parrot build system 101: Who can spot the off-by-one error: | 21:41 | |
| $conf->{hash_of_steps}->{ $new_steps[$i] } = $i + 1; | |||
| NotFound | I'm going to invent the off-by-zero error. | 21:43 | |
| And patent it. | 21:44 | ||
|
21:44
mj41 joined
21:56
zby_home joined
|
|||
| cotto | NotFound: good luck getting the uspto to ignore prior art. | 21:56 | |
| The first bug ever was an off-by-one error. | 21:57 | ||
| They expected zero moths to be in the computer, but there was actually one. | |||
|
22:01
zby_home joined
|
|||
| sorear | cotto has a shocking amount of faith in the USPTO | 22:02 | |
| NotFound | Next project will be off-by-NaN error. | 22:05 | |
|
22:21
snearch joined
22:36
jsut joined
|
|||
| aloha | (parrot/parrot) Issues opened : 236 (config/auto/memalign.pm is broken) by gerdr : github.com/parrot/parrot/issues/236 | 22:57 | |
| cotto | whatever happened to the github issue migration? | 23:04 | |
| dalek | rrot: 80e365d | cotto++ | docs/project/release_manager_guide.pod: add some release managers for 2012, should the world not end |
23:09 | |
|
23:11
contingencyplan joined
|
|||
| dalek | rrot/memalign-fix: 2443cc6 | cotto++ | config/auto/memalign.pm: initial patch from gerdr++ for issue #236 This shouldn't go into master without some discussion and iteration. |
23:16 | |
|
23:25
Hunger joined
23:42
kid51 joined
|
|||
| cotto | kid51: ping | 23:59 | |