Parrot 4.3.0 "In Which..." | parrot.org | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 25 April 2012.
benabik Huh. length(repeat(orig, 2)) != length(orig) * 2 00:02
00:12 lucian joined
dalek rrot: 6e617d2 | Whiteknight++ | / (2 files):
PackFile_pack_size should return size_t instead of opcode_t. This fixes #755. alester++ for finding this.
 00:12
benabik Oh. Because the file is different lengths for those operations. D'oh
dalek rrot: 1ed6d65 | Whiteknight++ | frontend/pbc_merge/main.c:
Add pointer modifiers to show which pointers are modified
 00:14
aloha (parrot/parrot) Issues closed : 755 (PackFile_pack_size returns the wrong type) by petdance : github.com/parrot/parrot/issues/755
dalek rrot: c193535 | benabik++ | t/dynpmc/gziphandle.t:
Clean up t/dynpmc/gziphandle.t stream tests

  * The two throws_substring tests had nothing to do with the rest of
   the tests in that function. Split them out into another function.
  * Make some test messages clearer.
  * Include comments for non-test portions.
  * Use variable names so new developers don't have to remember what
   registers are for.
  * Fudge by 1 byte for problematic zlib version on OS X		 00:24
rrot: b90faed | Whiteknight++ | api.yaml:
Add in a note about introspection methods for Key PMC in #717.
 00:29
aloha (parrot/parrot) Issues closed : 717 (Need a way to make register-ref Keys from PIR) by Whiteknight : github.com/parrot/parrot/issues/717 00:34
(parrot/parrot) Issues closed : 189 (Build failure in alarm.c on Win32) by jnthn : github.com/parrot/parrot/issues/189 00:39
dalek rrot: 55721f4 | Whiteknight++ | frontend/parrot_debugger/main.c:
Let's just call a spade a spade. parrot_debugger does not take any flags or commandline options, so test for them and fail if they are found. Also, the PIR compilation pathway does not work and it isn't worth the effort to fix it, so delete it. This fixes #540. The 'real' solution is to delete this file entirely and write a new debugger.
 00:57
aloha (parrot/parrot) Issues closed : 540 (parrot_debugger --help coredumps) by leto : github.com/parrot/parrot/issues/540 01:00
kid51 is glad that c193535 did not break that test file on Mac OS X 10.4 01:06
dalek rrot: f2b3047 | Whiteknight++ | src/pmc/lexpad.pmc:
Fix LexPad.get_iter to return an iterator of name->value paurs in the LexPad.

Previously this was redirecting to LexInfo.get_iter, which is a name->register_index hash. Creating the new, correct, iterator requires the creation of two temporary PMCs which is not performance-friendly. This is the best we can do with the current implementation of LexInfo/LexPad. This commit fixes #575.		 01:35
rrot: 34e5ad5 | Whiteknight++ | t/pmc/lexpad.t:
Add in a test for the new LexPad.get_iter implementation.

This test code is a modified version of the test code provided by pmichaud in #575.
whiteknight back below 500 open tickets
benabik kid51: I was writing it on OS X, with help from Coke who sees the gziphandle failures. I would have been very surprised it had broken OS X. 01:39
aloha (parrot/parrot) Issues closed : 575 (Bug in HashIteratorKey .value method when used with LexPad PMC) by pmichaud : github.com/parrot/parrot/issues/575 01:40
Coke benabik++ 01:51
whiteknight benabik: you're on OSX? As a quick question, can you tell me what config["osname"] is there? 01:54
benabik whiteknight: How do I get the config? 01:55
whiteknight ./parrot_config osname
benabik oh
d'oh
darwin 01:56
whiteknight okay, thanks
I'm heading to bed now. Talk to you later
benabik 'ngith
kid51 I wonder if it would be worthwhile to provide 'parrot-config' as an alias/symlink to 'parrot_config' 01:59
I first typed the hyphen rather than the underscore
benabik I seem to be having problems talking to parrot-users. Could someone else explain pir:: opcodes and Q:PIR to Zach S? 02:40
03:05 crab2313 joined 04:32 ingy joined
dalek rrot: 830e791 | petdance++ | / (2 files):
consting an arg and re-headerizing
 04:52
rrot: 094b1d5 | petdance++ | frontend/pbc_merge/main.c:
reheaderized, and removed an unused memory allocation
rrot: f6ed02d | petdance++ | frontend/pbc_merge/main.c:
remove unused var, and localize some loop variables
ttbot Parrot f6ed02d5 i386-freebsd-64int make error tt.taptinder.org/cmdinfo/82996 05:03
05:10 travis-ci joined
travis-ci [travis-ci] parrot/parrot#312 (master - f6ed02d : Andy Lester): The build was broken. 05:10
[travis-ci] Change view : github.com/parrot/parrot/compare/3......f6ed02d
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/1262798
05:10 travis-ci left 06:40 alvis joined 06:52 fperrad joined
dalek kudo/use-arglist: 6145974 | moritz++ | src/ (2 files):
allow importing of tags

the default is not imported when an explicit tag is provided		 08:55
kudo/use-arglist: 3c393c6 | moritz++ | src/Perl6/ (2 files):
refactor compile time evaluation into a World method
 09:19
kudo/use-arglist: 5b60299 | moritz++ | src/Perl6/Grammar.pm:
fix import + arglist
10:02 brambles joined 10:20 crab2313 joined
rurban I cannot understand what's going on with #766 perl6 -e''my @a=splice([], 1);' -=-=target=pir gives me push, but no splice. The error happens in splice though 10:27
perl6 -e'my @a=splice([], 1);' --target=pir
10:28 lucian joined
moritz there's a splice call in there 10:28
$ ./perl6 -e'my @a=splice([], 1);' --target=pir |ack splice $P105 = "&splice"($P103, $P104) 10:29
rurban that's a call? ok, thanks. I'll try do add that as parrot testcase
moritz the actual call to parrot's splice opcode happens in src/core/List.pm line 237 10:31
rurban nope, I'm giving up on this 10:37
10:44 JimmyZ joined 10:53 kid51 joined
nopaste "kid51" at 70.85.31.226 pasted "Parrot build is broken" (5 lines) at nopaste.snit.ch/138655 10:54
kid51 commit 830e79136b3ceb2f3c07efc1bc864a4871656e33 appears to be the culprit 11:02
JimmyZ cotto: ping 11:16
msg cotto I added comments to gist.github.com/2581153#comments 11:22
aloha OK. I'll deliver the message.
dalek rrot: 3ccd61d | jkeenan++ | / (2 files):
Revert 830e7913; it broke the build.

Developers: Please don't commit to master if you haven't run 'make test'.		 11:40
12:06 mtk joined 12:14 crab23131 joined 12:23 crab2313 joined 12:50 PacoAir joined
Coke wonders if kid51 is using "git revert" or somehow manually reverting those patches. 12:56
JimmyZ or git --amend ? 12:57
13:04 bluescreen joined, PacoAir_ joined 13:14 bluescreen joined 13:24 bluescreen_ joined 13:25 bluescreen joined 13:28 crab2313 joined 14:21 jashwanth joined
jashwanth dukeleto:hello 14:22
14:31 Justin joined
Justin good morning everyone. I have returned! 14:33
14:51 dmalcolm joined 14:55 zby_home joined
JimmyZ Justin: welcome back! 15:04
15:21 jashwanth joined 15:39 whiteknight joined
whiteknight good morning, #parrot 15:40
cotto ~~ 15:46
whiteknight hello cotto 15:49
cotto 'morning, whiteknight
15:49 Psyche^ joined
Justin morning whiteknight and cotto 15:53
whiteknight hello Justin 15:56
www.infoq.com/news/2012/05/vs_c99_support 16:05
Another depressing reminder that Microsoft refuses to support a modern C standard
dukeleto ~~ 16:20
benabik ~~ 16:23
17:04 PacoAir joined
NotFound ~~ 17:05
whiteknight: ping
nine Which Windows versions does Parrot support? 17:26
benabik To my knowledge, most of them. 17:30
Coke benabik: that is overly generous. 17:32
I am pretty sure it works on win 7 and win xp. 17:33
benabik Coke: Probably true.
I've personally used it on both of those. Pre-XP may be a little too old. I would personally assume if you can get a recent version of Strawberry Perl on it, Parrot will work.
17:34 lucian joined
cotto If you're pre-xp, you've got other problems. 17:39
nine Let me rephrase: do we want to continue to support XP? 17:50
benabik nine: Need something from Vista/7 for threading? 17:51
NotFound Now that you talk about windows... Do we want to continue to support Visual C++ own C version? 17:52
Given the recent posts about it, I'd say "If you want to use Visual C++, honor its name and compile as C++" 17:53
And let the rest of the world use a dialect of C that makes some sense in the current millenium. 17:54
nine +1 on that
benabik: Vista onwards support condition variables with exactly the same semantics as pthreads. That'd be quite nice. 17:55
Coke main rakudo dev is using win 7 17:56
benabik I'm torn. On one hand, MS itself has finally dropped support for XP. OTOH, it's nice to support as much as we can.
Coke 7 and up is fine with me.
17:56 contingencyplan joined
tadzik worth noting: main dev, and the only dev that works on windows, afaik 17:56
benabik nine: Is it simple to continue to support non-threading on XP? 17:57
nine benabik: well we already have some implementation of pthread_cond_timed_wait for Windows but IIRC whiteknight mentioned some problems, so I'm not sure it works as it should. 17:59
benabik: from what I read it should be possible to emulate it well enough. Just thought that if we don't care for XP anyway, I'd just switch to the guaranteed to work stuff. 18:00
benabik Mm. Personally, I'd vote for "support as much as we can without effort" on XP. No reason to bend over backwards for an EOL OS.
whiteknight NotFound: pong
NotFound whiteknight: What's the use case for that command-line include? 18:02
whiteknight NotFound: I'm thinking about ways to include things like files with const and inlines in an existing file for compilation 18:05
or, to take multiple .winxed files and compile them together into a single output.pir, with separate annotations
NotFound That last one shouldn't be the job of pbc_merge? 18:07
18:07 Maddingue joined
whiteknight yes, but you can't compile and merge a file with all inline and const definitions 18:07
18:10 alester joined
NotFound whiteknight: I think there is too much potential for confusion, and you can always copy to other file adding the include, or meta-program. 18:12
whiteknight yeah, okay. I just worry about messing up annotations 18:17
they're only valuable if they point back to the original source, not some generated meta-program 18:18
NotFound whiteknight: yeah, but if they points that an original source that is not exactly the same you compiled, you can be fooled. 18:20
Point to an
whiteknight Okay, so the best thing is just to use $include everywhere explicitly? 18:21
NotFound I think so. For a now at least.
whiteknight I'm thinking about a way to specify a "standard library" of builtins, that's pluggable so we aren't only using the builtins provided by winxed compiler itself 18:22
that might be too ambitious
NotFound whiteknight: parrot identifies pbcs by name, so adding ways to compile different variants of the same file wich potentially can load lib files that load other variant can give us a debugging nightmare. 18:26
whiteknight true 18:27
NotFound For a now, if you want to plug a lib of bultins, you can just include it.
whiteknight yeah, okay 18:28
thanks
18:51 crab2313 left
whiteknight devopsangle.com/2012/05/07/dell-ann...evelopers/ 19:17
dukeleto oooh shiny 19:18
benabik Is it just me or is "devops" the buzzword for "sharing those scripts every sysadmin was using anyway"? 19:19
dukeleto benabik: not really, it goes deeper than that
benabik: there is a whole new world of automation with Chef, Puppet and their competitors
benabik Well, once you share code, you get tools. 19:20
dukeleto benabik: portland even has a devops group: groups.google.com/group/pdxdevops?pli=1 (because Puppet is here)
benabik: i agree with you that "devops" is very buzzwordy and way too much marketing is going on about it now
benabik: but it also is a new "niche", where you have to understand how the code works as well as how to deploy that code to the cloud or where-ever and how to keep the deployments happy and healthy 19:21
benabik Ah, yes. EC2 is much more useful if you know how to build an image that sets itself up without assistance. 19:22
dukeleto benabik: devops is sometimes called "Site Reliability Engineer" or SRE, a term started by Google
but now lots of other places use the term "SRE"
tadzik hehe, SRE
dukeleto they are the people who push to production and deal with the consequences 19:23
tadzik SRE sounds a bit funny in Polish
dukeleto tadzik: did I miss a joke? Do tell!
tadzik well, "sra" is "(he) defecates" in a not really polite way
NotFound So "configuration as code" is the next big thing? Funny that my simple text editor did that the past millenium.
benabik Ars had a fascinating article about Facebook's release engineering.
dukeleto: thank you for being informative, but I must run. 19:24
Or, as I should have spelled that: dukeleto++
class &
dukeleto tadzik: good to know :) 19:30
NotFound: configuration as code seems trivial to us parrot devs, but many huge organizations are still living in COBOL-land. Configuration-as-code is like flying cars to them :)
NotFound dukeleto: also if "stored and managed in a git repo" means "being code" to them, we live in different universes. 19:33
19:35 lucian joined
NotFound whiteknight: issue 26 was already fixed, 2bd58f9c0914f4a35bfd6832278ea3f0fc8b37aa 19:36
19:36 GodFather joined
NotFound Note that the "THIS IS A GENERATED FILE" comment in the header was added because of parrot codingstd. If we want a mention of the sources everywhere we should add a codingstd test for it. 19:38
Every time I use a tablet or a smartphone near the desktop I miss a remote clipboard feature. 19:42
In fact, I'm thinking about writing it. 19:43
dalek nxed: ad68234 | NotFound++ | NEWS:
some more news
 19:56
kudo/use-arglist: 6fa9034 | jnthn++ | src/ (3 files):
Refactor handling of traits taking named arguments. The argument is now passed along with the name. Updated various of the built-in traits along with this.
 20:50
kudo/use-arglist: f9a94d8 | jnthn++ | src/core/traits.pm:
Handle export tags.
22:20 kid51 joined 22:26 whiteknight joined
whiteknight good evening, #parrot 22:27
kid51 yo 22:31
whiteknight hello kid51 22:36
what's crack-a-lackin'? 22:37
dukeleto kid51: hey there 22:40
dalek rrot: 95a7413 | Whiteknight++ | frontend/p (2 files):
pbc_disassemble, with no arguments, should print the help message. This fixes #505
 22:41
kid51 msg Coke I was rushing out the door to $job, so that patch was just applied with patch -R, not git revert 22:44
aloha OK. I'll deliver the message.
(parrot/parrot) Issues closed : 505 (pbc_disassemble with no arguments is not useful) by leto : github.com/parrot/parrot/issues/505 22:45
dalek rrot: 17287fe | Whiteknight++ | / (15 files):
Remove all uses of the unused, undocumented, meaningless :base_core modifier from ops. This fixes #275
 22:51
kid51 OT: Is anyone on channel using a Macbook Pro with OS X 10.7 (Lion)? 22:54
aloha (parrot/parrot) Issues closed : 275 (Remove :base_core modifier from *.ops files) by soh-cah-toa : github.com/parrot/parrot/issues/275 22:55
23:01 travis-ci joined
travis-ci [travis-ci] parrot/parrot#314 (master - 95a7413 : Whiteknight): The build was fixed. 23:01
[travis-ci] Change view : github.com/parrot/parrot/compare/3......95a7413
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/1271247
23:01 travis-ci left
whiteknight t/src/extend_vtable.t takes forever, comparatively 23:18
kid51_at_dinner yup
whiteknight I wonder if that should really be in coretest 23:19
or if we can move it further up in the sequence so we can parallelize it better 23:20
dalek rrot/whiteknight/gh_675: 4fd7560 | Whiteknight++ | / (4 files):
Remove vtable can from most places where it's used. Fix a test in namespace.t which is (perhaps unintelligently) using hard-coded vtable indices
rrot/whiteknight/gh_675: 34b729c | Whiteknight++ | / (3 files):
Remove two VTABLE can implementations, and fix a test in extend_vtable.t

Powered by Cro and the Raku® Programming Language.

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

Thank you!