Parrot 3.9.0 "Archaeopteryx" | parrot.org | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 19 October 2011.
00:04 kid51 joined
kid51 Pre-release testing: make test and make fulltest PASS on Linux: reg, optimize, g++ and g++ optimize. 00:42
Darwin: t/pmc/nci.t fails test #45, as it has since (approx) the green threads merge 00:43
So: same results as last month 00:44
benabik Darwin PPC?
kid51 Yes
benabik Seems to work on my machine, so must be a difference between PPC and x86 00:46
(Or x86-64, since I'm on 10.7)
nopaste "kid51" at 192.168.1.3 pasted "t/pmc/nci.t failure on darwin/ppc" (59 lines) at nopaste.snit.ch/94904 01:01
kid51 (That's same error as last month; just re-pasting.) 01:02
And, yes, Darwin Intel is showing PASS on Smolder (my buddy Gary) 01:03
01:13 woosley joined
dalek rrot: 53c9e6a | bacek++ | runtime/parrot/library/Test/Builder.pir:
Prohibit to plan 0 tests. Part of #188
 01:34
04:26 rfw joined 05:16 jsut_ joined 06:30 contingencyplan joined 07:23 bacek joined 07:55 schmooster joined 08:02 PacoLinux joined 08:10 PacoLinux_ joined 08:24 mj41 joined 09:07 lucian joined
nine Coke: so what would I have to do to get the new .pmc into the MANIFEST? 09:56
10:03 woosley left
moritz nine: perl tools/dev/mk_manifest_and_skip.pl 10:07
(I think)
nine moritz: tried that but it didn't do anything 10:08
10:08 elmex joined
moritz nine: have you 'git add'ed it already? 10:09
nine no
moritz well, try that first
(and maybe you even need to commit)
10:19 snearch joined 11:01 lucian joined 11:17 jsut joined 12:11 Psyche^ joined 12:19 bacek joined
dalek rrot: cbe7cea | bacek++ | src/ (6 files):
Replace deprecated usage of 'pmc_new' with 'Parrot_pmc_new'
 12:37
rrot: dffbd93 | bacek++ | include/parrot/pmc.h:
Remove very old 'compatibilty macros'
12:46 bluescreen joined 13:34 whiteknight joined
whiteknight good morning, #parrot 13:35
very interesting rant about try/catch: groups.google.com/forum/#!topic/no...ESsssIxrUU
moritz I found the hn thread for that interesting too: news.ycombinator.com/item?id=3233526 13:36
especially the link to the lisp exceptions/conditions system: www.gigamonkeys.com/book/beyond-exc...tarts.html 13:37
whiteknight I didn't read that thread, I'll have to add it to my queue
dalek rrot/threads: 45d8c8b | nine++ | / (6 files):
Not yet functional implementation of proxy pmc
 13:38
rrot/threads: ddf2502 | nine++ | / (4 files):
Fix scheduling updates on data owning interpreters

Storing references to foreign interpreters and probably other objects as well screws up garbage collection and leads to all kinds of strange errors.
Accessing attributes or methods of the proxy pmc itself leads to a whole lot of confusion about for example which namespace to search.
Working solution is a new schedule_proxied method which keeps access to the proxy's attributes and the foreign interp in C space where it's safe.
moritz nine: want to know what the old "fix" for GC + threads was? :-) 13:40
nine disabling GC?
moritz right
nine I have to admit that it's tempting :) But as far as I can see it even works with GC now :) 13:41
moritz that's much better :-) 13:42
whiteknight yeah, this time we're not going to fall into the old traps 13:45
at least, not consciously
dalek p/bigint: f7b182b | moritz++ | / (3 files):
float -> bigint op

round-trips big FLOATVALs		 13:48
nxed: 20821eb | NotFound++ | examples/fly.winxed:
Shake a bit the point view in example fly automatic mode 2

Aren't trigonometrics fun?		 14:07
nxed: 343b220 | NotFound++ | winxedst1.winxed:
emit in void context the iteration expression of for
 14:10
14:21 whiteknight joined
dalek kudo/bigint: 9a01b11 | moritz++ | / (2 files):
use nqp::fromnum_I in Num.Int
 14:24
nxed: 45c0020 | NotFound++ | winxedst1.winxed:
check invalid uses of void bultins
NotFound Rants about exceptions? The neverending story. 14:29
moritz can't we and the OS just please stop making errors, so that we don't need exceptions? kthxbye 14:31
oh, and the internez also please stop erroring 14:32
whiteknight it's not a question of not having errors, it's a question of whether try/catch mechanisms are the best for handling them
NotFound But there is an interesting point in one comment: if some functions wants to throw a resumable exception, it can just set an appropiate continuation (or some other invokable). 14:33
whiteknight: the answer is easy: nothing is "the best".
whiteknight yes, true. Nothing is best. 14:38
NotFound Also, these discusions tend to generalize too much about what an error or an exceptional condition is. Is reaching EOF an error? That depends. On a text file is expected, reading a struct from a binary file usually is an error.
14:39 whiteknight joined 15:10 PacoLinux joined
dalek nxed: 836a8c4 | NotFound++ | winxedst1.winxed:
emit some comments only when compiling with --debug option

instead of when compiling with it		 15:19
nxed: b8e779e | NotFound++ | winxedst1.winxed:
delete commented out debugging statments
 15:22
15:22 alester joined
alester github.com/petdance/scratch/issues now properly quotes the Trac tickets coming in 15:44
but I still don't have it saving history.
whiteknight alester++
you
re doing a great job 15:45
alester thanks 15:49
dalek nxed: 5779af8 | NotFound++ | winxedst1.winxed:
optimize code generated for case selection in switch
 16:03
rrot-libgit2: 7a34e6a | dukeleto++ | t/winxed/001_load.t:
Add a test for git_repository_is_bare
 16:04
16:08 lateau joined
dukeleto alester: looks like you are making progress. Also, you mailbombed me :) 16:22
alester OH? 16:28
How so? 16:29
dukeleto alester: i get an email every time you put @leto in a github issue
alester oh my
dukeleto alester: which is fine usually, but I have gotten >100 emails in the last 24 hours :)
alester You won'tb be the only one.
whiteknight yeah, I got a pile too. I had 100 messages from github this morning
dukeleto alester: nope :(
alester SORRY!
whiteknight not that I mind. A little mailbombing in the name of progress is no big deal
dukeleto alester: it is ok, i understand it is part of testing
alester I didn't realize that would happen.
dukeleto alester: i am fine with it, if the conversions happen soon :) No worries. bacek++ might come to your door with a pitchfork, though 16:30
alester I just fired off mail to Github about ticket ownerhisp.
dukeleto alester++
alester My converter can't say "This was created by @coke"
dalek nxed: 5952fb1 | NotFound++ | / (4 files):
update NEWS and generated files
alester It does in the body of the ticket, but I can't make the ticket actually look like it was submitted by him.
OK I commented out the part that includes @github ID in the ticket. 16:32
sorry guys! :-)
dukeleto alester: great for the final conversion, but yeah, leave it commented out for testing :) 16:33
alester Yes, mom. :-)
That's only a stopgap in case I can't show the real ticket creator anyway.
dalek rrot: 5911a95 | NotFound++ | ext/winxed/ (2 files):
update winxed snapshot to 5952fb1701

  - Better checking of void expressions use		 16:34
16:37 mj41 joined
dalek nxed/version_1_4: 72ef5b8 | NotFound++ | NEWS:
put build 1.4.0 in NEWS
 16:41
nxed/version_1_4: 713720c | NotFound++ | pir/winxed_compiler.pir:
jump version to 1.4.0
nxed/version_1_4: d5354a1 | NotFound++ | winxedst1.winxed:
Ooops. Forgot to add one file
 16:48
rrot: 85ad83d | NotFound++ | ext/winxed/compiler.pir:
update winxed snapshot to RELEASE_1_4_0
 16:51
16:52 PacoLinux_ joined
dalek rrot: ad4b9de | NotFound++ | ChangeLog:
put winxed news in ChangeLog
 16:54
nxed: 72ef5b8 | NotFound++ | NEWS:
put build 1.4.0 in NEWS
 16:55
whiteknight holy crap, I just realized the release is tomorrow 16:56
...and that I am the release manager for it
Coke alester++ # ticket migration specialist.
dukeleto whiteknight: good thing it keeps getting more automatede 16:57
dalek nxed: 99b9b01 | NotFound++ | / (2 files):
start 1.5 devel
Coke whiteknight: The best part is when you realize that the day OF. ;)
whiteknight Coke: I was wondering why NotFound was updating the snapshot and updating NEWS 16:59
like, why would anybody touch NEWS unless it was right before the release ?
dukeleto sighs 17:00
whiteknight: i seem to be the only person that attempts to iteratively update it
NotFound Hey, I can release a new winxed if I want to, no need to wait for a parrot one ;) 17:01
I'm happy with the recent changes in stage 1, is much more robust and testable. 17:03
whiteknight I've been seeing some of them, does look better 17:04
NotFound The limitations of parrot iterators are starting to hurt me, with better iterators we'll be able to write nice algorithms, binders and functors C++std-alike. 17:08
whiteknight what's wrong with our iterators?
NotFound Reversed ones, for a start.
whiteknight our iterators do allow reversing
most of them, anway
NotFound Yeah, if you rewrite the algorithm for reverse mode. 17:09
Also, no way to advance without reading. 17:10
No write
whiteknight ok 17:11
NotFound For example, to write an equivalent of transform, the destination must be a container, not an iterator. 17:14
whiteknight Rosella has some stuff like that. Streams that you can prepend data into 17:23
NotFound Yes, but for general efficiency such algorithms should be closer to the virtual metal. 17:24
Doing a bunch of procesing with just an algorithm, an iterator and a closure is nice. 17:25
alester well, phoeey 17:27
I have to create all the Github tickets as JSON
no more API fun. 17:28
Coke was there an API and now its goone? 17:30
*gone?
or are you trying to do things the API can't?
NotFound Do you have an APIphany? 17:33
17:34 fperrad joined 17:40 ambs joined 17:57 mtk joined
Coke alester: for bonus points, if the trac ticket is really an RT, you could pull the original opener from THAT ticket. ;) 18:16
(I got 171 [scratch] emails today, no doubt because I manually transferred a ton of the original tickets.) 18:17
dalek nxed: f1c678e | NotFound++ | winxedst1.winxed:
shorten a bit syntax error emision
 18:57
alester Coke: my goal is to have this done by the end of the week for GCI 19:13
I think that's what dukeleto was asking for.
sorear alester: hi 19:17
Coke alester: \\o/
sorear alester: are you just migrating tickets or are you commenting on thme? 19:18
alester just migrating
dalek nxed: 26f7c56 | NotFound++ | winxedst1.winxed:
builtins unshift and pop/shift _ var/int/float/string
 19:21
19:33 benabik joined, alester joined 19:42 alester joined 20:19 ambs_ joined 20:35 schmooster joined
dalek p/bigint: f41528b | jnthn++ | src/ops/nqp_bigint.ops:
Work around MSVC++ being too clever and catching a compile time divide by zero.
 21:03
Coke is reminded to try to build parrot on windows. 21:13
21:16 mj41 joined
whiteknight Coke: yes, please. I'm going to try and block off some time tomorrow morning to do the same 21:21
nopaste "NotFound" at 192.168.1.3 pasted "Experiment on constructor calls" (46 lines) at nopaste.snit.ch/95049 21:26
NotFound This experiment shows a possible way to add support for constructor calls. 21:27
benabik I have a working Win7 install now, although no dev environment on it. What would I need to build Parrot?
cotto benabik: trac.parrot.org/parrot/wiki/Platforms/Windows 21:31
benabik cotto++ 21:33
benabik loads up Parallels.
cotto That's one of the wiki pages we shouldn't lose.
benabik Maybe clean up a bit… Is the Strawberry 5.10/Parrot 0.9 info still needed? 21:36
cotto that part might be dated 21:44
21:45 rfw joined
dalek p/bigint: beb69a6 | moritz++ | src/ops/nqp_bigint.ops:
unbust conversion from float to bigint (hopefully)
 22:04
benabik PowerShell is vaugely useful. 22:08
This is far less painful than I expected. 22:10
It also appears that we need TAP::Harness::Archive instead of Test::TAP::HTMLMatrix? 22:18
cotto It's not bad.
benabik Ah. i != install in CPAN. That makes a difference. 22:19
benabik updates wiki. 22:22
Looks like t/pmc/alarm.t hangs. 22:24
(Win7 32b) 22:25
aloha: msg whiteknight Looks like t/pmc/alarm.t hangs on 32b Win7. 22:26
aloha benabik: OK. I'll deliver the message.
Coke that windows platform page is all wrong. ;) 22:30
I have switched to using msvc 2010 express and activestate perl.
(that and command line git makes it very easy to build rakudo on windows.) 22:31
I'll write up an alternate to that page with the MS tools links tonight. 22:33
dalek tracwiki: v9 | benabik++ | Platforms/Windows 22:37
tracwiki: Removing stale Windows environment information.
tracwiki: trac.parrot.org/parrot/wiki/Platfor...ction=diff
kudo/bigint: 296212e | jnthn++ | src/core/Num.pm:
fail for Inf.Int and Inf.Rat. Maybe not the perfect solution, but it's far better than hanging.
 22:53
kudo/bigint: e4cd7a4 | jnthn++ | src/core/Num.pm:
fail for Inf.Int and Inf.Rat. Maybe not the perfect solution, but it's far better than hanging.
 23:06
nopaste "NotFound" at 192.168.1.3 pasted "Unpolished experimental patch to add support for constructors in Class" (121 lines) at nopaste.snit.ch/95083 23:09
23:47 PacoLinux_ joined 23:49 kid51 joined
bacek_at_work msg NotFound I think it's better to check "does(invokable)" instead of "isa(CallContext)" and just invoke passed Sub. Or I totally misunderstand your idea? 23:50
aloha OK. I'll deliver the message.
23:51 whiteknight joined
whiteknight good evening, #parrot 23:55
benabik: ping
dalek rrot-libgit2/oid: 8090e6f | dukeleto++ | / (3 files):
Start of a Git2.Oid class
 23:56
rrot-libgit2: 608f65b | dukeleto++ | README.md:
Update readme
rrot-libgit2: 9023894 | dukeleto++ | README.md:
Add a link to winxed.org in the readme
 23:58
23:59 benabik joined
cotto hio whiteknight 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!