Camelia, the Perl 6 bug

IRC log for #parrot, 2011-07-22

Parrot | source cross referenced

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:17 rurban_ joined #parrot
00:17 lucian left #parrot
00:19 rurban left #parrot
00:19 rurban_ is now known as rurban
00:38 theory left #parrot
00:46 dalek parrot/m0-prototype: b4ac6ba | cotto++ | src/m0/perl5/m0_ (2 files):
00:46 dalek parrot/m0-prototype: start making m0 assembler and inter smarter about encoding strings
00:46 dalek parrot/m0-prototype:
00:46 dalek parrot/m0-prototype: still not smart enough to work though
00:46 dalek parrot/m0-prototype: review: https://github.com/parrot/parrot/commit/b4ac6baa1d
00:58 whiteknight joined #parrot
01:11 smash left #parrot
01:26 whiteknight left #parrot
01:28 cotto ~~
01:44 woosley joined #parrot
01:52 nbezzala joined #parrot
01:52 cotto left #parrot
01:53 cotto joined #parrot
02:03 woosley left #parrot
02:09 nbezzala_ joined #parrot
02:11 nbezzala left #parrot
02:11 nbezzala_ is now known as nbezzala
02:16 preflex joined #parrot
02:20 preflex_ left #parrot
02:54 plobsing joined #parrot
02:56 woosley joined #parrot
03:06 dalek parrot/soh-cah-toa/hbdb: 5984fcf | soh_cah_toa++ | / (3 files):
03:06 dalek parrot/soh-cah-toa/hbdb: Almost fixed "run" and "step" commands. The "run" command works fine but "step" only works right if you do at least one "run" before it.
03:06 dalek parrot/soh-cah-toa/hbdb: review: https://github.com/parrot/parrot/commit/5984fcfb72
03:34 jsut_ joined #parrot
03:34 jsut left #parrot
03:37 soh_cah_toa pff...readline_interactive() is so stupid. you can't even use the arrow keys to move the cursor
03:39 sorear next time, install readline before building Parrot
03:40 soh_cah_toa i have readline
03:47 cotto alias hbdb='rlwrap hbdb'
03:48 sorear soh_cah_toa: what about readline-dev?  and did you install that before configure?
03:51 soh_cah_toa looks like i had readline but not readline-devel. damn, now i gotta rebuild
03:52 cotto aren't you glad you installed ccache?
03:52 soh_cah_toa honestly, i can't tell a difference
03:52 cotto time it sometime
03:53 benabik I need fink back so I can have ccache again.
03:54 soh_cah_toa though compared to this other hbdb crap issues, sitting around waiting for parrot to build seems fun
03:58 benabik Although I wonder if clang will build parrot...
03:58 benabik Has anyone tried it?
04:19 theory joined #parrot
04:49 nbezzala left #parrot
05:02 soh_cah_toa left #parrot
05:09 nopaste "cotto" at 192.168.1.3 pasted "h8h8h8" (20 lines) at http://nopaste.snit.ch/63580
05:10 cotto ah.  There's a sneaky hidden character.
05:14 * cotto throws another hack on the pile
05:14 nbezzala joined #parrot
05:38 cotto What's the best way to get Perl to parse a float?
05:52 daniel-s joined #parrot
06:06 nbezzala left #parrot
06:08 particle left #parrot
06:09 mj41 joined #parrot
06:19 plobsing left #parrot
06:27 atrodo left #parrot
06:29 daniel-s left #parrot
06:29 atrodo joined #parrot
06:29 daniel-s joined #parrot
06:43 jsut joined #parrot
06:45 moritz Scalar::Util::looks_like_number ?
06:48 jsut_ left #parrot
07:02 theory left #parrot
07:11 dalek parrot/m0-prototype: 968b07d | cotto++ | src/m0/perl5/m0_ (2 files):
07:11 dalek parrot/m0-prototype: get string encoding (and m0b loading) to work
07:11 dalek parrot/m0-prototype: review: https://github.com/parrot/parrot/commit/968b07d70b
07:11 dalek parrot/m0-prototype: e0edb1d | cotto++ | src/m0/perl5/m0_interp.pl:
07:11 dalek parrot/m0-prototype: Don't print the terminal null from print_s.  It makes a test sad.
07:11 dalek parrot/m0-prototype: review: https://github.com/parrot/parrot/commit/e0edb1dcc5
07:11 dalek parrot/m0-prototype: 74c93bf | cotto++ | t/m0/ (2 files):
07:11 dalek parrot/m0-prototype: fix a couple of tests, update canonical hello world m0b
07:11 dalek parrot/m0-prototype: review: https://github.com/parrot/parrot/commit/74c93bf154
07:11 dalek parrot/m0-prototype: f68715a | cotto++ | src/m0/perl5/m0_assembler.pl:
07:11 dalek parrot/m0-prototype: add naive float parsing to the M0 assembler
07:11 dalek parrot/m0-prototype: review: https://github.com/parrot/parrot/commit/f68715a632
07:11 dalek parrot/m0-prototype: a6d210f | cotto++ | src/m0/perl5/m0_interp.pl:
07:11 dalek parrot/m0-prototype: fix M0 interp's N handling, get another batch of tests passing
07:11 dalek parrot/m0-prototype: review: https://github.com/parrot/parrot/commit/a6d210f64e
07:14 JimmyZ joined #parrot
07:18 dalek rakudo/nom: 05500ce | moritz++ | src/ (2 files):
07:18 dalek rakudo/nom: make $::($name) style indirect lookup work
07:18 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/05500ced74
07:43 JimmyZ left #parrot
07:55 he joined #parrot
08:16 fperrad joined #parrot
08:17 rurban__ joined #parrot
08:19 rurban left #parrot
08:19 rurban__ is now known as rurban
08:22 dalek rakudo/nom: afce340 | moritz++ | t/spectest.data:
08:22 dalek rakudo/nom: enable two (heavily fudged) test files, one of which was never passing in master
08:22 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/afce3407cd
08:54 dalek rakudo/nom: adcf1ed | moritz++ | / (2 files):
08:54 dalek rakudo/nom: Int.base
08:54 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/adcf1edf7e
09:22 dafrito joined #parrot
09:41 daniel-s left #parrot
09:50 cosimo_ left #parrot
10:05 cosimo joined #parrot
10:08 woosley left #parrot
10:45 preflex_ joined #parrot
10:47 preflex left #parrot
10:51 contingencyplan left #parrot
11:30 Felipe morning
11:30 ambs joined #parrot
11:34 dalek rakudo/nom: cdbbd7f | moritz++ | / (3 files):
11:34 dalek rakudo/nom: implement &reduce
11:34 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/cdbbd7f474
11:34 dalek rakudo/nom: 1b55dd4 | moritz++ | / (2 files):
11:34 dalek rakudo/nom: implement &rand and &srand, run tests
11:34 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/1b55dd4d02
11:39 dafrito left #parrot
11:42 Coke "let's get rid of c" "hey, now we need to reimplement ccache for parrot"
12:04 atrodo Coke> I would hope that getting rid of c means not needing ccache
12:11 whiteknight joined #parrot
12:18 whiteknight good morning, #parrot
12:24 bluescreen joined #parrot
12:27 Felipe good morning
12:28 dalek rakudo/nom: 6970e7e | moritz++ | / (4 files):
12:28 dalek rakudo/nom: values, pairs, kv
12:28 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6970e7e59b
12:40 whiteknight hello Felipe. How are you doing today?
12:49 plobsing joined #parrot
12:52 bubaflub joined #parrot
12:52 lucian joined #parrot
12:52 bubaflub ~
12:54 whiteknight hello bubaflub
12:54 bubaflub hello whiteknight.  how's life?
12:54 whiteknight bubaflub: Good! My sister had her baby two days ago, and we have a date picked for settlement
12:54 bubaflub whiteknight: that's great!
12:54 whiteknight for srsly
12:57 smash joined #parrot
12:57 smash hello everyone
12:58 whiteknight good morning smash!
12:58 whiteknight smash: for your thesis, you should write up a bunch of awesome DSL compilers for Parrot
12:59 smash whiteknight: writing DSLs is in the plan, let's try to target parrot for them
13:12 contingencyplan joined #parrot
13:20 particle joined #parrot
13:21 dalek rakudo/nom: ca7b148 | moritz++ | / (3 files):
13:21 dalek rakudo/nom: &sprintf, Enum.pairs, tests
13:21 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/ca7b148b47
13:28 Limbic_Region left #parrot
13:31 Coke smash: or, perhaps, perl6.
13:33 smash Coke: that is also an option, yes
13:45 dalek parrot/whiteknight/pbc_pbc: 287059a | Whiteknight++ | compilers/imcc/pbc.c:
13:45 dalek parrot/whiteknight/pbc_pbc: Uncomment the TT #1324 code from compilers/imcc/pbc.c. It was causing a test failure in code that expects to be able to HLL map Sub in an :immediate. This fixes t/dynpmc/subproxy.t
13:45 dalek parrot/whiteknight/pbc_pbc: review: https://github.com/parrot/parrot/commit/287059aad3
13:45 dalek parrot/whiteknight/pbc_pbc: 5acd4c4 | Whiteknight++ | t/compilers/opsc/06-opsfile.t:
13:45 dalek parrot/whiteknight/pbc_pbc: Bump the number of ops in t/compilers/opsc/06-opsfile.t. This fixes that test.
13:45 dalek parrot/whiteknight/pbc_pbc: review: https://github.com/parrot/parrot/commit/5acd4c4653
13:48 plobsing left #parrot
13:48 whiteknight I would really like to get this branch merged in to master soon, if I can manage it
13:59 mj41 left #parrot
14:13 whiteknight The weather outside is absolutely miserable today
14:14 whiteknight let me put it this way: I was drenched in sweat and had to change my shirt. On the way to work, at 8am in the morning
14:14 dalek parrot/whiteknight/pbc_pbc: b07fb58 | Whiteknight++ | / (101 files):
14:14 dalek parrot/whiteknight/pbc_pbc: Merge branch 'master' into whiteknight/pbc_pbc
14:14 dalek parrot/whiteknight/pbc_pbc: review: https://github.com/parrot/parrot/commit/b07fb58143
14:21 atrodo whiteknight> weather--
14:32 whiteknight msg NotFound Take a look at a little test program I wrote on the whiteknight/pbc_pbc branch: https://gist.github.com/1099564 We can create winxed include files from any compiled .pbc
14:32 aloha OK. I'll deliver the message.
14:32 whiteknight atrodo: tell me about it. It's rediculous out there
14:33 atrodo It happens every year, and I still hate it
14:33 NotFound whiteknight: nice
14:33 whiteknight I don't mind it usually, and I am normally a warm-weather person. However, it's extremely uncomfortable to sit in a puddle of your own sweat while you're working for 8 hours
14:34 whiteknight NotFound: I need to detect :method, and use that to declare classes
14:34 whiteknight but otherwise, we can do it all from .pbc files directly
14:34 whiteknight I'm surprised Sub doesn't have an .is_method method or attribute that I can easily access
14:35 SHODAN joined #parrot
14:36 NotFound whiteknight: that is with the new load_bytecode variant, isn't it?
14:38 whiteknight yeah. It returns a PackfileView
14:38 whiteknight and we can use that to read through the constants table
14:38 whiteknight pull out the subs, get the namespace from each sub, etc
14:38 NotFound Looks like it can be even more useful than we thought.
14:39 whiteknight I always thought it would be extremely useful
14:39 whiteknight that's why I'm working on it :)
14:39 NotFound Yes, but I didn't realize how easy to use it can become.
14:40 NotFound whiteknight++
14:50 whiteknight okay, I got class detection working: https://gist.github.com/1099564
14:50 whiteknight easy-peasy
14:50 whiteknight ideally, when we have classes being constructed at compile time and serialized into the constants table, we can just read them out of there directly
14:53 NotFound Good enough to be useful right now, even if only for checking.
14:55 whiteknight yeah. it obviously is letting in the :anon functions too, like Winxed_class_init. It would be trivial to filter those out
14:56 NotFound And it can be useful to add optional compile time checking of availability of classes or functions in external modules.
14:56 whiteknight yes!
15:01 whiteknight And testing can become much easier. I can open a file, extract all methods, and run them as tests, without any boilerplate code at all
15:02 dalek Rosella: 7785eb2 | Whiteknight++ | src/include/Core.winxed:
15:02 dalek Rosella: Update include/Core to include forward declarations of important namespaces, classes, and functions
15:02 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/7785eb2905
15:02 dalek Rosella: 2f90c7b | Whiteknight++ | / (5 files):
15:02 dalek Rosella: Update Action to make better use of the predeclarations of namespaces and extern functions in modern winxed. Add tests for various error conditions which weren't tested previously and make use of the new syntax features
15:02 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/2f90c7be36
15:02 dalek Rosella: 5a53a93 | Whiteknight++ | src/include/Action.winxed:
15:02 dalek Rosella: add forward declares for include/Action.winxed. Nothing uses it yet
15:02 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/5a53a93e76
15:05 NotFound whiteknight: How is the plan about that branch? Merge before next release?
15:05 whiteknight NotFound: I just sent out an email. Want to merge it tonight
15:06 whiteknight I have got lots of stuff I want to do that relies on it
15:06 NotFound Good. I must go now, we'll look at it later.
15:06 whiteknight later
15:21 contingencyplan left #parrot
15:26 plobsing joined #parrot
15:28 he left #parrot
15:33 JimmyZ joined #parrot
15:35 lucian whiteknight: i may have asked this before, but do you have an example of a rosella test harness in winxed?
15:37 whiteknight lucian: http://whiteknight.github.com/​Rosella/libraries/harness.html
15:37 whiteknight It's a minimalist example, down towards the bottom of the page
15:38 lucian whiteknight: ah, very nice. thanks
15:38 whiteknight I think it should be up-to-date. I haven't tested those examples in a while
15:39 whiteknight hmm... no. broken. Give me a minute
15:40 lucian some confusion between factory objects and functions, i think
15:41 lucian it's doing "var harness = new Rosella.Harness();" , when i think it should be doing something like "var harness = construct(Rosella.Harness);"
15:41 dukeleto ~~
15:41 whiteknight no, Winxed doesn't need construct()
15:41 lucian oh, ok
15:41 whiteknight that's a compatibility thing for NQP
15:41 lucian i get "Method 'Harness' not found for invocant of class 'Rosella;Harness'"
15:43 dalek Rosella: ca509d5 | Whiteknight++ | src/harness/Harness.winxed:
15:43 dalek Rosella: Add empty constructor to Harness
15:43 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/ca509d5699
15:43 dalek Rosella/gh-pages: fcf4d6b | Whiteknight++ | libraries/harness.md:
15:43 dalek Rosella/gh-pages: fix winxed Harness examples
15:43 dalek Rosella/gh-pages: review: https://github.com/Whiteknig​ht/Rosella/commit/fcf4d6b616
15:44 whiteknight lucian: Update Rosella. I fixed the bug. Sorry for the confusion
15:44 lucian whiteknight: thanks. i'll try it
15:44 whiteknight The Rosella::construct() function looks for a constructor, and does nothing if it can't find one. The winxed "new Foo()" syntax calls the constructor without checking first
15:45 bubaflub lucian: when you're done, i'd like to see it - the only reason i used a NQP harness was because that was the example that was available.
15:45 dukeleto i finally got gcc64 (sparc64 openbsd) to submit a smolder report for coretest: http://smolder.parrot.org/app/​projects/report_details/18571
15:45 whiteknight so, it's a small mismatch in semantics
15:45 bubaflub dukeleto++
15:45 lucian bubaflub: ok. it's basically whiteknight's example
15:45 lucian whiteknight: i see
15:46 lucian hmm, i think i need new parrot & winxed
15:46 whiteknight Solution is either "new Foo" (without the parens), or add an empty constructor
15:46 lucian right. makes sense
15:48 lucian hmm, i got a parrot build error. hopefully i'm doing something wrong
15:48 whiteknight what's the error?
15:48 whiteknight if the last time you built was before the 3.6 release, you might need to reconfigure
15:49 lucian yeah, that's what i'm doing
15:54 whiteknight That's one of the things I hate most about the release, because some of the generated function names use the version number
15:54 whiteknight so when you build things look fine until the linker tries to resolve symbols
15:54 whiteknight And you have a Parrot_3_5_0_whatever instead of Parrot_3_6_0_whatever and bork
15:59 dukeleto lucian: what error are you getting?
15:59 lucian dukeleto: hmm, buried in the console log
15:59 lucian no error after Configure, though
16:03 dukeleto lucian: &> is your friend
16:04 lucian dukeleto: sure, but i can't reproduce it anymore
16:04 lucian since i re-configured
16:04 lucian whiteknight: does this make any sense to you https://gist.github.com/1099731?
16:04 dukeleto lucian: that is good. Just a partial build issue
16:05 dukeleto lucian: i usually get that error when I attempt to call a function with the namespace prefix
16:05 dukeleto lucian: instead of doing: using Namespace.foo; foo(...)
16:05 whiteknight lucian: Do a "winxed --version"
16:06 lucian whiteknight: hah https://gist.github.com/1099736
16:07 dukeleto lucian: that is an older winxed
16:07 dukeleto lucian: newer winxed's have a --version option
16:07 lucian dukeleto: i noticed
16:08 lucian whiteknight: yeah, it was just a very old winxed
16:08 whiteknight okay, awesome.
16:09 lucian i'm using the one that comes with parrot now, managed to build
16:10 whiteknight nice
16:10 whiteknight sorry this is turning into such a project!
16:10 lucian not your fault
16:13 JimmyZ_ joined #parrot
16:15 lucian bah. i don't even get a line number
16:15 lucian too few positional arguments: 2 passed, 3 (or more) expected
16:16 whiteknight lucian: yeah, that was a bug. I updated the example on the webpage
16:16 rurban__ joined #parrot
16:17 lucian right, so not the harness runs
16:17 whiteknight this line: view.add_run(testrun, 0);
16:17 plobsing left #parrot
16:18 lucian s/not/now/
16:18 lucian whiteknight: yeah, i'd noticed
16:18 JimmyZ left #parrot
16:18 whiteknight oh, okay
16:18 whiteknight awesome
16:18 JimmyZ_ is now known as JimmyZ
16:18 lucian i still don't get any actual tests to run though :)
16:19 lucian ah, extension. i didn't notice it expects .t
16:19 rurban left #parrot
16:19 whiteknight Hmm.. I may not have an easy interface to change that
16:19 rurban__ is now known as rurban
16:20 lucian for not, i've renamed them
16:20 whiteknight still, that seems like a glaring omission now that I think about it
16:21 lucian for rosella.test's age, it's not exactly glaring
16:21 lucian but yeah, i'd prefer to keep the regular extensions. being in the t/ folder is plenty indication for me that it's a test
16:21 lucian and now all tests fail! awesome :)
16:22 whiteknight We need to differentiate tests from things like temporary files, hidden files, editor files, etc.
16:22 plobsing joined #parrot
16:22 lucian why would those be in t/ ?
16:24 whiteknight I don't ask questions. I close tickets
16:25 lucian :)
16:27 lucian ah, now i know why everything fails. because my built .winxed is empty
16:28 whiteknight heh, nice sanity check
16:32 lucian it's not on purpose. just that my setup.winxed build is broken :)
16:33 lucian bubaflub: here's my harness https://bitbucket.org/lucian1900/puff​in/src/c4f834b5392d/objects/t/harness
16:33 bubaflub lucian++
16:33 lucian it's really just whiteknight's example, pretty much verbatim
16:33 lucian make sure you change the cmd in setup
16:34 whiteknight lucian: You don't need to $load "rosella/core.pbc" separately. It's included in harness.pbc
16:34 lucian whiteknight: ok
16:34 whiteknight not that it matters, just trying to save some effort
16:35 lucian thanks
16:37 plobsing left #parrot
16:44 dalek parrot/whiteknight/pbc_pbc: 2ca406f | Whiteknight++ | src/packfile/api.c:
16:44 dalek parrot/whiteknight/pbc_pbc: Fix c_arg_assert.t
16:44 dalek parrot/whiteknight/pbc_pbc: review: https://github.com/parrot/parrot/commit/2ca406f8e2
16:44 dalek parrot/whiteknight/pbc_pbc: ed077ec | Whiteknight++ | src/p (2 files):
16:44 dalek parrot/whiteknight/pbc_pbc: fix c_functions_docs.t
16:44 dalek parrot/whiteknight/pbc_pbc: review: https://github.com/parrot/parrot/commit/ed077ecdcb
16:44 dalek parrot/whiteknight/pbc_pbc: 0838d53 | Whiteknight++ | src/platform/win32/dl.c:
16:44 dalek parrot/whiteknight/pbc_pbc: fix c_parens.t
16:44 dalek parrot/whiteknight/pbc_pbc: review: https://github.com/parrot/parrot/commit/0838d53048
16:44 dalek parrot/whiteknight/pbc_pbc: 5f99543 | Whiteknight++ | src/ops/core.ops:
16:44 dalek parrot/whiteknight/pbc_pbc: Fix opcode_doc.t, and expand some docs for the old-style load_bytecode
16:44 dalek parrot/whiteknight/pbc_pbc: review: https://github.com/parrot/parrot/commit/5f99543078
16:46 lucian yay! https://bitbucket.org/lucian190​0/puffin/changeset/98473f37ebd5 :)
16:52 dalek parrot/whiteknight/pbc_pbc: 6e5911f | Whiteknight++ | api.yaml:
16:52 dalek parrot/whiteknight/pbc_pbc: +experimental note about load_bytecode_p_s to api.yaml
16:52 dalek parrot/whiteknight/pbc_pbc: review: https://github.com/parrot/parrot/commit/6e5911f71a
16:55 theory joined #parrot
16:57 lucian whiteknight: thanks for all the help. not to plug the python distutils into parrot's
17:05 whiteknight holy crap it is hot outside
17:05 whiteknight like, working in an old-time steel mill hot
17:06 whiteknight it's 38.9 C
17:07 lucian nasty
17:09 dukeleto lucian: i have a puffin repo. the README is empty :(
17:10 lucian dukeleto: huh?
17:10 lucian dukeleto: ah, so it is
17:10 lucian don't worry, if there had been anything in it, it'd be useless by now :)
17:10 dukeleto lucian: how do I run the tests?
17:11 dukeleto lucian: python setup.py build works for me
17:11 lucian dukeleto: yeah, but it doesn't do anything
17:11 dukeleto lucian: stuff like how to build, test, clean and install is good stuff for the readme
17:11 lucian the object system has a working build/test
17:11 dukeleto lucian: how do I run those tests?
17:11 lucian cd object, and there you can do setup.winxed build/test
17:11 lucian i've just made i work
17:11 whiteknight ./dwim "test some stuff"
17:13 dukeleto lucian: https://gist.github.com/1099882
17:13 mj41 joined #parrot
17:13 whiteknight yay! the whiteknight/pbc_pbc branch is fulltest-clean on this machine
17:13 lucian dukeleto: you have to say build or something
17:15 lucian dukeleto: try pulling again
17:15 * Coke wonders if it's time for a smolder restart.
17:18 dukeleto lucian: still the same error after pulling and updating: "load_bytecode" couldn't find file 'src/winxed/Distutils.bootstrap.pir'
17:18 dukeleto lucian: do you have some special env vars set ?
17:18 lucian dukeleto: might've not commited that file
17:19 whiteknight where the heck are you getting that from?
17:19 whiteknight that looks like a Rosella temporary file
17:19 dukeleto whiteknight: it is right after: Bootstrapping build with existing distutils library
17:19 lucian whiteknight: i hacked that to get my boostrapping
17:19 Coke how do I diff master vs. whiteknight/pbc_pbC ?
17:19 lucian dukeleto: pushed fix
17:19 whiteknight Coke: git diff master..whiteknight/pbc_pbc
17:20 dukeleto Coke: git diff master...whiteknight/pbc_pbc
17:20 whiteknight lucian: ah, okay
17:20 lucian whiteknight: it works, so atm i'm not too bothered
17:21 whiteknight lucian: Okay, Rosella has a bootstrappy process similar to that as well
17:21 dukeleto Coke: .. is for a straight diff, ... is to use the common ancestor of the two branches
17:22 Coke ah. I need to have a local branch for that work, aye?
17:22 dukeleto Coke: no, just use the origin/whiteknight/pbc_pbc syntax
17:22 dukeleto Coke: your git index knows about every branch on the remote, even if you don't have a local branch
17:22 Coke if I do that, does it go over the network?
17:23 dukeleto Coke: nope
17:23 Coke k.
17:23 mj41 left #parrot
17:23 dukeleto Coke: git branch -a to see all branches
17:24 dalek rakudo/nom: 8c0ca27 | moritz++ | src/core/Enum.pm:
17:24 dalek rakudo/nom: return a List from Enum.{keyes,values,pairs}
17:24 dalek rakudo/nom:
17:24 Coke ugh. of course, this diff is useless because it's swamped by the generated files. ;)
17:24 dalek rakudo/nom: Based on feedback by pmichaud++
17:24 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/8c0ca2715d
17:24 dukeleto Coke: are you using the ... form ?
17:24 dukeleto Coke: that is usually what you want.
17:25 Coke dukeleto: yes. there are still generated files in the branch that differ from master.
17:26 cosimo left #parrot
17:26 Coke hurm. I had created a branch for whiteknight/pbc_pbc, realized I didn't need it, and then did a branch -d to remove it and got a warning... ignorable?
17:26 whiteknight should be fine, yes
17:26 Coke just want to make sure I wasn't going to kill the remote one. ;)
17:27 whiteknight I've got a pristine copy here. I can replace it if you do
17:27 lucian do parrot/rosella.test output any particular character encoding?
17:27 whiteknight lucian: I'm sure it's ascii by default, or whatever the platform default is
17:28 whiteknight heh, our first piece of trac spam since we opened up ticket creation again
17:28 lucian ok, i'll use platform default. although utf-8 seems to work ok too
17:30 dalek rakudo/nom: a29189c | moritz++ | src/core/operators.pm:
17:30 dalek rakudo/nom: allow assignment to indirectly looked up variables
17:30 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a29189c72a
17:31 cosimo joined #parrot
17:32 dalek TT #2162 created by maro++: The rock plays casino
17:32 dalek TT #2162: http://trac.parrot.org/parrot/ticket/2162
17:34 dukeleto spam--
17:34 dukeleto Coke: git branch never touches the network
17:34 Coke Yes, I figured if I screwed anything up it'd require a push to be noticed.
17:37 preflex joined #parrot
17:37 whiteknight deleted the ticket and locked down ticket creation again
17:38 moritz whiteknight++
17:38 lucian dukeleto: have you managed to get the tests running?
17:40 dukeleto lucian: i think my rosella is borked, possibly because I recompiled parrot but not rosella. will fiddle with it
17:41 * dukeleto thinks he has gcc64 (openbsd sparc64) smoking parrot again
17:41 preflex_ left #parrot
17:41 bluescreen left #parrot
17:41 dukeleto poor little machine takes 10 minutes to comile the core ops
17:41 dukeleto compile, rather
17:41 whiteknight ouch
17:45 dukeleto yeah, it is a 1 ghz processor running openvpn and test suites for various projects like ffmpeg and whatnot. It is "the little sparc that could"
17:45 dalek nqp: 5134a59 | moritz++ | tools/build/PARROT_REVISION:
17:45 dalek nqp: bump PARROT_REVISION to 3.6.0
17:45 dalek nqp: review: https://github.com/perl6/nqp/commit/5134a598c2
17:45 dukeleto whiteknight: ooh, it just found a compile error on master
17:46 dukeleto whiteknight: https://gist.github.com/1099956
17:47 dalek rakudo/nom: 21756e5 | moritz++ | tools/build/NQP_REVISION:
17:47 dalek rakudo/nom: bump nqp revision to one that brings us parrot 3.6.0
17:47 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/21756e5971
17:49 whiteknight dukeleto: string API? I don't know who's been touching that
17:49 whiteknight also, (g++)--
17:50 dukeleto whiteknight: the file hasn't changed for a while, but this smoker hasn't been running for longer. It seems like some wonkiness in the macros
17:50 whiteknight hmm
17:50 dukeleto whiteknight: i just did a git clean -fdx and I am trying to again to make sure it wasn't a fluke
17:51 dukeleto whiteknight: nope, same error
17:52 dukeleto whiteknight: this is gcc 3.3.5 !
17:53 whiteknight ......
17:53 whiteknight *facepalm*
17:54 Coke That's seems to be old enough to not bother supporting.
17:54 cotto_work ~~
17:56 bluescreen joined #parrot
17:57 whiteknight I think we absolutely have to draw the line at 3.4.5, if we can't demand more modern than that
17:57 whiteknight anything older is absurd
18:00 dukeleto whiteknight: well, we should tell the user at configure time that they are absurdly behind the times, then, right?
18:01 Coke I would setup the gcc config step to warn.
18:01 Coke (but not die)
18:01 dukeleto funny thing is that i think gcc 3.3.5 will compile parrot, but not g++ 3.3.5
18:02 mj41 joined #parrot
18:02 dukeleto Coke: why should it warn but not die?
18:04 Coke because if someone wants to make it work, now they have /2/ problems.
18:04 dukeleto Coke: but the point is, we don't *want* to support such old versions
18:07 whiteknight dukeleto: Can you upgrade GCC on that box?
18:09 Coke dukeleto: no, but are you saying you'll reject patches to make them work?
18:10 Coke no point in going out of our way to make it hard for folks to get parrot working on their system (if it doesn't otherwise hurt us)
18:10 cotto_work Coke: +1
18:10 atrodo gcc 3.3.3 compiles parrot
18:11 dukeleto atrodo: what about g++ 3.3.3 ?
18:11 * Coke is loving git-svn this time around.
18:11 atrodo dukeleto> no idea, i can try
18:12 dukeleto whiteknight: i think i have access to all gcc versions ever on the farm machines, that is just the default for openbsd 4.6, i guess
18:13 dukeleto whiteknight: i have used the default gcc on every other box, this one happens to have the oldest i've found
18:14 whiteknight dukeleto: okay. Using the default gcc is not bad practice, until we run into one of these landmines
18:15 dukeleto i think our support policy should state the minimum versions of compilers that we support
18:16 dukeleto supporting every compiler ever written because we can is not a good use of resources
18:19 atrodo dukeleto> Looks like no.  Configure doesn't even work with g++ 3.3.3
18:20 Coke dukeleto: I think we're covered by the "2 year old" rule.
18:23 whiteknight We're coming up on the 2-decade rule with that one
18:25 whiteknight no, that's an exaggeration. It was 2004
18:29 JimmyZ left #parrot
18:39 atrodo 2004 is correct.  That's when I built my server that's running 3.3.3
18:48 mj41 left #parrot
19:16 dalek parrot: a233112 | dukeleto++ | t/src/extend_vtable.t:
19:16 dalek parrot: [t] Fix Parrot_PMC_get_pointer and Parrot_PMC_get_pointer_keyed tests, which assumed that memory locations were always positive integers. It just ain't so.
19:16 dalek parrot: review: https://github.com/parrot/parrot/commit/a233112d65
19:21 Coke tortoise-git has git-svn support. whee.
19:27 whiteknight say it ain't so!
19:28 benabik "it ain't so!"
19:30 whiteknight yay!
19:42 bubaflub left #parrot
19:44 benabik …  Apple depreciated all of OpenSSL.  How weird.
19:45 cotto_work nobody uses that
19:45 benabik It threw me for a loop when I tried to build git...
19:47 cotto_work odd indeed.  What's the replacement?
19:48 benabik I can't find a notice of it being depreciated anywhere…  It just throws warnings on build
19:48 whiteknight no replacement, it's a hardware module built into macs that can't be user-serviced
19:48 benabik I think the replacement is "provide your own SSL"?
19:52 whiteknight poppycock
19:53 benabik Huh.  Parrot 3.3.0 is in homebrew.
19:59 whiteknight msg NotFound can winxed add a __CLASS__ builtin, like __FUNCTION__, to give the name of the current class? __NAMESPACE__ would be nice too
19:59 aloha OK. I'll deliver the message.
20:00 fperrad left #parrot
20:11 rblackwe left #parrot
20:14 lucian dukeleto: i improved the readme, and the build/test process too. now you can just use setup.py
20:15 lucian dukeleto: if you want to try, it should work now
20:22 dalek Rosella: c88bc7f | Whiteknight++ | src/container/ (7 files):
20:22 dalek Rosella: Update Container-related files to use the new forward-declarations for Core and Action.
20:22 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/c88bc7ff6b
20:22 dalek Rosella: ca5a31c | Whiteknight++ | / (3 files):
20:22 dalek Rosella: Fix some tests in Container, and fix some bugs I found while fixing tests
20:22 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/ca5a31c41a
20:27 whiteknight left #parrot
20:42 dukeleto bugs bug BUGS!
20:45 Coke Stop steaming up my tail.
20:48 contingencyplan joined #parrot
20:49 M_o_C joined #parrot
20:58 M_o_C left #parrot
21:04 zpmorgan joined #parrot
21:05 zpmorgan left #parrot
21:18 dalek rakudo/nom: a52ea8f | moritz++ | src/Perl6/Metamodel/MROBased (2 files):
21:18 dalek rakudo/nom: fix newlines in two MOP files; no functional changes
21:18 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a52ea8fcb6
21:21 rblackwe joined #parrot
21:24 ambs left #parrot
21:26 lucian i'm having trouble figuring out an error https://gist.github.com/1100462
21:26 lucian code https://bitbucket.org/lucian1900/puffin/​src/0d6245934d55/objects/src/boot.winxed and test https://bitbucket.org/lucian1900/puff​in/src/0d6245934d55/objects/t/boot.t
21:30 dmalcolm joined #parrot
21:43 Psyche^ joined #parrot
21:45 SHODAN left #parrot
21:47 bluescreen left #parrot
21:48 Patterner left #parrot
21:48 Psyche^ is now known as Patterner
21:58 whiteknight joined #parrot
21:59 pyrimidine left #parrot
22:05 whiteknight would anybody mind a whiteknight/pbc_pbc merger soonish?
22:11 whiteknight I hate messing with some of the big files like core_ops.c, or imcparser.c in a branch. The diffs are always huge and contain zero relevant information
22:11 whiteknight I wish git had a tool to "don't freaking show this damn file in a diff"
22:23 plobsing joined #parrot
22:23 kid51 joined #parrot
22:23 * kid51 is in PHL for FOSSCON
22:24 * kid51 goes out to dinner with pmichaud shortly
22:24 * kid51 is testing pbc_pbc branch on linux/i386
22:25 kid51 whiteknight: Would you expect that branch's behavior to vary between gcc/g++, optimize/non-optimize, or across OSes?
22:25 whiteknight kid51: no. I'm doing some testing with compilers right now
22:25 whiteknight I've been skimpy on my g++ testing so far
22:26 kid51 pbc_pbc branch PASS linux/i386 gcc non-opt (my regular linux build)
22:26 kid51 dinner
22:26 kid51 left #parrot
22:38 cotto_work whiteknight: only if you promise to break something
23:01 dmalcolm left #parrot
23:15 * lucian raises fist at imcc
23:15 lucian (i think, anyway)
23:26 Felipe :)
23:43 whiteknight cotto_work: I'm trying my best!
23:43 whiteknight no objections. So....incoming
23:43 dalek parrot: d6cb4d8 | Whiteknight++ | / (20 files):
23:43 dalek parrot: Merge branch 'whiteknight/pbc_pbc'
23:43 dalek parrot: review: https://github.com/parrot/parrot/commit/d6cb4d899e
23:46 whiteknight now, we can start playing with our packfiles from PIR code like real adults
23:46 bluescreen joined #parrot
23:49 whiteknight I couldn't remove do_sub_pragmas entirely, because it's still used in IMCC and intertwined in a few other things
23:49 whiteknight TT #1324 was a big part of why I didn't give it the axe
23:50 whiteknight pmichaud: ping
23:51 contingencyplan left #parrot
23:52 whiteknight plobsing: ping
23:54 whiteknight You know, I can't be held responsible for what happens if I'm not properly supervised with all this code

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

Parrot | source cross referenced