Camelia, the Perl 6 bug

IRC log for #parrot, 2010-12-13

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:06 dalek parrot: dd7ff4e | NotFound++ | src/pmc/namespace.pmc:
00:06 dalek parrot: fix a segfaulting bug in method Namespace.export_to
00:06 dalek parrot: review: https://github.com/parrot/parrot/commit/dd7ff4e50f
00:20 dalek parrot: 6f2c9b1 | NotFound++ | t/pmc/namespace.t:
00:20 dalek parrot: regresion test for dd7ff4e
00:20 dalek parrot: review: https://github.com/parrot/parrot/commit/6f2c9b1ade
01:01 mikehh rakudo (d1e6636) - builds on parrot (2_10_1-764-g8d73990) - make test PASS, make spectest_smolder[(#1623) roast (744caf1)] FAIL - Kubuntu 10.10 amd64 (g++-4.5 with --optimize)
01:01 mikehh t/spec/integration/advent2009-day20.t - Failed test:  18
01:01 mikehh t/spec/S02-magicals/pid.rakudo - TODO passed:   1
01:01 mikehh 27,051 ok, 1 failed, 610 todo, 1,856 skipped and 1 unexpectedly succeeded
01:08 Kapace has anyone taken a look at https://github.com/kapace/parrot/commit/​d22eea0ccaf4a4876d1d011b2d6a291dbe68895d ?
01:11 kapace_ left #parrot
01:24 mikehh rakudo - reran the test on its own a few times and it did not fail
01:25 mikehh rakudo - I am getting untracked files (tempfile_filehandles_io.20461.856 and similar) with git status after a make realclean
01:25 NotFound Kapace: Why do you want to use registers instead of variables?
01:28 plobsing Kapace: what do you *mean* by "variables". AFAIK, PIR doesn't have variables distinct from registers.
01:29 NotFound Yeah, what plobsing says.
01:38 mikehh left #parrot
01:39 lucian joined #parrot
01:56 mikehh joined #parrot
02:01 Yuki`N plobsing, how do locals work then?
02:02 plobsing do you mean the '.local [pmc|string|num|int] var' syntax? that's just synatx sugar for registers.
02:03 plobsing it does the *exact* same thing
02:03 Yuki`N Ah.
02:03 kapace2 joined #parrot
02:03 Yuki`N How many registers are there then?
02:03 plobsing MAXINT
02:03 kapace2 NotFound: oh I don't know there are some variables that I only use like once
02:04 whiteknight joined #parrot
02:04 kapace2 using a register would get rid of one line and negative code is good
02:04 plobsing Yuki`N: try not to use them all at once ;)
02:06 kapace2 whiteknight: can you review https://github.com/kapace/parrot/commit/​d22eea0ccaf4a4876d1d011b2d6a291dbe68895d
02:06 whiteknight kapace2: sure, give me a minute
02:08 Yuki`N xD
02:09 Yuki`N Anyway as to why I almost 100% covered Iterator, base classes are important too. :<
02:09 whiteknight NotFound: ping
02:09 Yuki`N Just waiting on someone to close the pull request and the melange task.
02:09 whiteknight Yuki`N: what task?
02:09 Yuki`N http://socghop.appspot.com/gci/task/show/google​/gci2010/parrot_perl_foundations/t129193194564
02:11 whiteknight kapace2: Code looks decent so far. Are you ready to open a pull request?
02:12 mikehh left #parrot
02:14 dalek parrot/gci_iterator_coverage: 6b61c67 | Yuki`N++ | t/pmc/iterator.t:
02:14 dalek parrot/gci_iterator_coverage: Increase Iterator PMC test coverage.
02:14 dalek parrot/gci_iterator_coverage: review: https://github.com/parrot/parrot/commit/6b61c676a2
02:14 dalek parrot/gci_iterator_coverage: 532f1ce | Whiteknight++ | / (21 files):
02:14 dalek parrot/gci_iterator_coverage: Merge branch 'master' of git://github.com/nol888/parrot
02:14 dalek parrot/gci_iterator_coverage: review: https://github.com/parrot/parrot/commit/532f1ce469
02:15 kapace2 yeah, haven't tested 100% yet, but ill open a pull request
02:17 kapace2 oh yeah, I needed to ask, what to do about Connection: close|Keep-Alive
02:17 mikehh joined #parrot
02:17 whiteknight kapace2: okay. NotFound was the one who first wanted that task, so I was hoping he would be around to examine it
02:17 whiteknight I'll pull for test though if you open the pull request. I won't make you wait for him
02:18 kapace2 ok https://github.com/parrot/parrot/pull/32
02:18 kapace2 when he's around he can help me out maybe
02:19 whiteknight awesome. Give me a minute to get that pulled
02:19 kid51 joined #parrot
02:21 whiteknight Yuki`N: All tests appear to be passing. When my test run finishes I will merge and close your ticket
02:21 whiteknight thanks!
02:21 dalek parrot: 532f1ce | Whiteknight++ | / (21 files):
02:21 dalek parrot: Merge branch 'master' of git://github.com/nol888/parrot
02:21 dalek parrot: review: https://github.com/parrot/parrot/commit/532f1ce469
02:21 Yuki`N No problem. :P
02:21 dalek parrot: ee634da | Whiteknight++ | t/pmc/iterator.t:
02:21 dalek parrot: Merge branch 'gci_iterator_coverage'
02:21 dalek parrot: review: https://github.com/parrot/parrot/commit/ee634dab9e
02:21 whiteknight oops, I already merged
02:22 Yuki`N xD
02:25 whiteknight Yuki`N: Accepted and closed. Thanks!
02:25 whiteknight msg cotto I closed the GCI task for improving iterator PMC test coverage. Hope you don't mind
02:25 whiteknight aloha?
02:25 whiteknight goddamn bot
02:26 plobsing aloha isn't here when you need her, purl was there even when you didn't want her, seems there is no perfect solution
02:27 Yuki`N lol.
02:28 whiteknight aloha has been pretty good, I thin
02:28 whiteknight think
02:30 dalek parrot/gci_chunked_receive: bbf140f | (David Czech)++ | / (4 files):
02:30 dalek parrot/gci_chunked_receive: Getting There on Chunked Encoding...
02:30 dalek parrot/gci_chunked_receive:  and the other md5 GCI task is in here too.
02:30 dalek parrot/gci_chunked_receive: review: https://github.com/parrot/parrot/commit/bbf140fc8b
02:30 dalek parrot/gci_chunked_receive: d22eea0 | (David Czech)++ | runtime/parrot/library/LWP/Protocol.pir:
02:30 dalek parrot/gci_chunked_receive: Chunked Encoding now working.
02:30 dalek parrot/gci_chunked_receive:  Need some more testing and review.
02:30 dalek parrot/gci_chunked_receive:  Perhaps a bit of cleanup, Use registers instead of variables.
02:30 dalek parrot/gci_chunked_receive: review: https://github.com/parrot/parrot/commit/d22eea0cca
02:34 Matt221 whiteknight: Is it possible to run a single test vs. all of them with `make test`?
02:34 whiteknight Matt221: yes, use the "prove" command
02:34 whiteknight prove t/foo/bar.t
02:35 Matt221 Thanks! will save so much time
02:37 whiteknight yeah.
02:37 dalek parrot: eaddf69 | Whiteknight++ | / (4 files):
02:37 dalek parrot: Merge remote branch 'origin/gci_chunked_receive'
02:37 dalek parrot: review: https://github.com/parrot/parrot/commit/eaddf691ce
02:39 whiteknight You guys are too good, and you're doing tasks too quickly
02:39 whiteknight we can hardly keep up
02:42 kapace2 whiteknight: you wouldn't know if we want to have chunked-encoding by default?
02:42 whiteknight kapace2: I have no idea. That's a question for NotFound
02:43 Matt221 whiteknight: Is the first parameter Parrot_api_string_export_ascii takes just a Parrot_Interp?
02:43 whiteknight Matt221: it's a ParrotInterpreter PMC, which contains a Parrot_Interp
02:43 whiteknight the API only really works on PMCs
02:43 Matt221 ok thanks!
02:45 whiteknight no problem
02:46 whiteknight the EMBED_API_CALLIN macro extracts a Parrot_Interp from the Parrot_PMC interp_pmc
02:50 kid51 whiteknight: Am reading your blog on Lorito
02:50 kid51 Where are the 'raw meeting notes' you refer to?
02:50 atrodo Ohhh, new whiteknight blog?
02:50 kid51 http://planet.parrotcode.org/
02:51 whiteknight kid51: cotto nopasted them a while back
02:51 kid51 do you recall on which day?
02:51 whiteknight I should have posted a copy, but I didn't want to put up something cotto wrote without permission
02:51 whiteknight the night after the meeting, I think
02:51 atrodo http://nopaste.snit.ch/26898 is cotto's
02:51 atrodo https://gist.github.com/735786 is dukeleto's
02:52 atrodo and https://gist.github.com/736285 is my response to cotto's
02:52 whiteknight yes
02:54 * atrodo reads whiteknights thoughts on lorito
02:55 kid51 No, dukeleton on lorito meeting is this:  https://gist.github.com/736628
02:55 atrodo haha, you're right
02:56 atrodo i should have verified
03:02 Yuki`N Lorito, that's the refactoring of some core parrot opcodes?
03:02 Yuki`N I think dukeleto was talking to me about that.
03:06 atrodo More or less, yes
03:07 atrodo The term is still kind of ambiguous, but it's getting better
03:07 whiteknight Yuki`N: That's what it was originally, it has grown into a massive architectural refactor
03:07 Yuki`N Ah.
03:07 Yuki`N I'm somewhat interested. :P
03:07 Yuki`N Especially since there's more C coding involved.
03:07 atrodo whiteknight> so in your lorito world, would each thread have it's own GC, or would everyone still share GC?
03:08 whiteknight atrodo: that's a great question, and I really don't know what the answer is or should be
03:08 lucian i'm sure there's plenty of research on the subject
03:08 atrodo i could go one of two directions with my prototype depending on the answer
03:08 whiteknight if we have one GC, it's either going to have to use a specifically-parallelizable algorithm (and such do exist) or we're going to have to add locks
03:09 whiteknight One GC per thread seems awfully heavy-weight, unless we limit the number of "real" OS threads and use a ton of Green-threads or Erlang-style actors
03:09 atrodo right
03:10 whiteknight so two processor cores = two parrot threads, each with a GC, but we have many tasks that run on them
03:10 whiteknight Anyway, I have to go to bed now. We can chat more about this tomorrow
03:10 atrodo but having a global gc would complicate your idea of manual object sync
03:10 atrodo night
03:10 whiteknight right, but like I said, there are GC algorithms specifically designed to be run in concurrent environmentments
03:10 whiteknight if we don't mind spending an entire OS thread on GC, we can do that
03:11 whiteknight goodnight
03:11 whiteknight left #parrot
03:15 kid51 left #parrot
03:20 Yuki`N left #parrot
03:37 plobsing NotFound: is it possible to have dynamic method calls in winxed? that is with the method name coming from a 'string' variable?
04:00 lucian left #parrot
04:12 bluescreen left #parrot
04:29 Coke joined #parrot
04:51 allison left #parrot
04:58 Matt221 left #parrot
05:11 kapace2 left #parrot
05:20 cotto seen aloha?
05:20 cotto aloha was last seen a while ago saying some stuff
05:25 simcop2387 left #parrot
05:28 rurban_ joined #parrot
05:31 simcop2387 joined #parrot
05:31 rurban left #parrot
05:31 rurban_ is now known as rurban
05:37 Kapace_ joined #parrot
06:26 simcop2387 left #parrot
06:28 jsut joined #parrot
06:33 jsut_ left #parrot
06:46 theory left #parrot
07:33 Khisanth left #parrot
07:54 Khisanth joined #parrot
08:09 sorear I seem to recall something about a LALR parser generator for Parrot either being created or desired by someone here recently, which was it and is there a reference?
08:15 mikehh sorear: we had pirc, which was moved out to parrot/pirc
08:17 sorear pirc is a LALR parser generator?
08:17 cotto it's an alternate pir compiler written in C
08:18 mikehh it was supposed to be a replacement for imcc
08:18 sorear I'm talking about LALR parser generators, not PIR compilers?  Or is there a connection I'm not seeing?
08:20 mikehh well it used bison, or something like that as an LALR parser
08:20 sorear imcc uses bison too
08:21 sorear what I'm talking about was a bisony thing that emits PIR
08:21 sorear something you could use instead of nqp-rx if you hate Perl
08:22 cotto sorear, pirc consumes pir, not emits it
08:23 sorear cotto: mikehh brought up pirc, not me
08:23 cotto ah
08:25 mikehh sorear: maybe http://whiteknight.github.com/2010/1​2/07/javascript_on_parrot_plan.html
08:26 fperrad joined #parrot
08:26 sorear I am thinking of something older
08:32 NotFound plobsing: you can use the predefined functions find_method and callmethodwithargs. They are remanants of early proof of concept tests and are ugly, but should do the job.
09:15 rfw left #parrot
09:50 he left #parrot
10:00 cosimo_ left #parrot
10:03 lucian joined #parrot
10:12 dip joined #parrot
10:27 lucian left #parrot
12:23 bluescreen joined #parrot
12:46 plobsing NotFound: I found those, thanks. Now I'm on to not being able to specify [method] on anonymous functions.
12:47 plobsing sorear: lucian/pynie don't want to work in nqp-rx
12:49 plobsing NotFound: is there a reason inner functions aren't allowed to have names (which might be handy for backtraces), or flags?
12:51 * moritz isn't NotFound, but thinks that named inner functions would need proper lexical scoping
12:52 moritz otherwise you could call it before the enclosing block has run, which makes closures weird
12:52 plobsing I don't want to be able to call it by the name. I just want to associate the name with it so backtraces look nicer.
12:53 plobsing functional style gets a little unweildy if you can't associate names with lambdas.
12:53 plobsing because backtraces look like "In function '' called from function '' called from function '' ..."
13:07 mtk joined #parrot
13:11 cosimo joined #parrot
13:29 Coke aloha, 1+1?
13:31 rurban_ joined #parrot
13:33 rurban left #parrot
13:33 rurban_ is now known as rurban
13:41 jsut_ joined #parrot
13:45 jsut left #parrot
13:49 whiteknight joined #parrot
14:06 whiteknight good morning #parrot
14:09 plobsing o/
14:09 dalek TT #1890 created by coke++: installed parrot requires build dir.
14:09 dalek TT #1890: http://trac.parrot.org/parrot/ticket/1890
14:09 dalek parrot: b3ec6cf | mikehh++ | MANIFEST:
14:09 dalek parrot: re-generate MANIFEST
14:09 dalek parrot: review: https://github.com/parrot/parrot/commit/b3ec6cf554
14:12 whiteknight good morning plobsing
14:14 Coke msg whiteknight I just stole a ticket from you as I didn't realize it was already assigned.
14:16 whiteknight msg Coke: okay, you can have it
14:16 whiteknight which ticket?
14:18 atrodo ESTILLNOALOHA
14:21 Coke spaces in build dirs
14:32 dalek parrot: 7959525 | mikehh++ | md5task/main_m.pir:
14:32 dalek parrot: fix codetest failures
14:32 dalek parrot:
14:32 dalek parrot: add copyright line
14:32 dalek parrot: add pir coda
14:32 dalek parrot: remove trailing spaces
14:32 dalek parrot: some documentati8on would be welcome!
14:32 dalek parrot: review: https://github.com/parrot/parrot/commit/79595251a9
14:32 dalek parrot: 55c5515 | mikehh++ | md5task/main_nfio.pir:
14:32 dalek parrot: fix codetest failures
14:32 dalek parrot:
14:32 dalek parrot: add copyright line
14:32 dalek parrot: add pir coda
14:32 dalek parrot: remove trailing spaces
14:32 dalek parrot: some documentati8on would be welcome!
14:32 dalek parrot: review: https://github.com/parrot/parrot/commit/55c5515f66
14:32 dalek parrot: 4e18c27 | mikehh++ | runtime/parrot/library/LWP/Protocol.pir:
14:32 dalek parrot: fix codetest failures - trailing spaces
14:33 dalek parrot: review: https://github.com/parrot/parrot/commit/4e18c2770d
14:36 simcop2387 joined #parrot
14:40 mikehh we reaqlly need to get aloha on one of our servers, haven't seen bacek for a bit
14:45 mikehh bah, new keyboard and it is slightly different in size, seem to be hitting adjacent keys (aq i8)
14:47 particle perhaps osuosl has a box we can host aloha on
14:47 particle don't know if we want to run it on the parrot.org vm
14:47 mikehh where was purl hosted?
14:57 particle i forget. feather, maybe?
14:58 particle maybe not.  hachi (in #perl5) fed and cared for purl.  may have been hosted on a server under his control.
14:58 NotFound plobsing: pong
15:00 plobsing NotFound: IWBNI inner functions were a little less restricted in winxed (not sure if this is possible)
15:01 NotFound plobsing: the main design principle for the current implementation was simplicity. Now that is starting to be well proven, we can talk about adding features,
15:02 plobsing I got around the method issue though. Apparently non-methods when used as methods get the invocant as the first arg.
15:02 NotFound plobsing: what usages you have in main for inner methods?
15:03 plobsing when you start using higher order functions on methods, you eventually want anonymous methods
15:03 * mikehh got to go out for a couple of hours - bbl
15:04 plobsing callmethodwithargs() is the start of that path
15:06 NotFound plobsing: yes, but I'm not sure if using a method declaration is a blessed way for that type of things, or only works by accident
15:07 Andy joined #parrot
15:07 plobsing NotFound: I don't follow
15:09 plobsing are you saying you don't think the :method flag is for that kind of thing?
15:09 NotFound plobsing: :method is supposed to be a method for the class related to the namespace where it lives, ot derived from that class. Don't know if using it for any other objects is desired.
15:11 NotFound In fact I think that all that :method does is to provide 'self'.
15:11 moritz I think it also hides it from the namespace, unless :nsentry is present
15:11 plobsing That's exactly all it should do. It should set the signature up to take an invocant and put it in 'self'.
15:11 NotFound moritz: yes, but for anon functions makes no difference ;)
15:12 moritz NotFound: point taken :-)
15:12 plobsing The fact that the PBC loader looks at the sub flags of every sub and loads them into the appropriate namespaces/classes/vtables is a horrid accident of history.
15:13 NotFound plobsing: that is my second concern. If you declare an anonymous method inside of a method, you lose access to the outer 'self'.
15:13 whiteknight yes
15:13 whiteknight plobsing: it sounds like what you want is :invocant
15:13 plobsing NotFound: I am aware of variable aliasing. Its perfectly acceptable.
15:14 fbrito joined #parrot
15:14 NotFound plobsing: well, I'll try a temptative implementation to play with it, we'll see were it gives us.
15:15 plobsing whiteknight: we have neither tests nor libraries that make use of ':invocant'. I highly suspect it will not work.
15:16 NotFound plobsing: about Issue 7, I thought that problem was already fixed, mut maybe fooled myself. I'll look at it as soon as I can.
15:17 plobsing no worries. I've worked around the issues encountered
15:17 plobsing with any luck, soon, I will unleash my horribly named library on the world. muahahahaha
15:18 moritz what's the name?
15:19 plobsing Ωη;)XD
15:19 NotFound WWD advances!
15:19 NotFound (Winxed World Domination)
15:20 fbrito1 joined #parrot
15:21 fbrito left #parrot
15:21 fbrito1 is now known as fbrito
15:21 Matt221 joined #parrot
15:21 fbrito Kovensky: ping
15:24 whiteknight NotFound: ping
15:25 whiteknight plobsing: no, you're right. :invocant was never implemented. It's still a wishlist item
15:25 whiteknight a good hacker who knows PCC should be able to put it in without too much hassle, if you're blocking on it
15:26 plobsing I'm not
15:26 fbrito does anyone here speak portuguese? :)
15:30 NotFound whiteknight: pong
15:31 NotFound fbrito: I'm from Galicia, I can underst portuguese more or less.
15:31 fbrito oh :o
15:31 fbrito I guess I need someone to review my work on this task: http://www.google-melange.com/gci/task/show/goog​le/gci2010/parrot_perl_foundations/t129186823954
15:32 whiteknight NotFound: a GCI student submitted work for the chunked receive thing in LWP and I merged it yesterday
15:32 whiteknight I wanted you to take a look at it to make sure it's sane
15:33 whiteknight fbrito: I think darbelo might
15:33 NotFound whiteknight: I've seen it, and do some test. The only problem is that I don't understand why he ser "Connection: Keep-alive" unconditionally.
15:33 NotFound s/ser/set
15:33 whiteknight NotFound: oh, okay
15:34 NotFound In my ajax example I explicity set "Connection: close", and is overwritten.
15:36 NotFound And I don't see much point on keeping the connection when the socket is created and owned by the UA just for one request, BTW.
15:38 NotFound But anyway, if the user provides a value it shouldn't be overrided except for security concerns, IMO.
15:39 NotFound fbrito: my level in portuguese technical jargon is null, I can't help.
15:40 Patterner left #parrot
15:41 fbrito NotFound: ahhh. Thank you anyway
15:43 fbrito where is aloha? :O
15:45 plobsing she's taking a vacation in hawaii
15:45 NotFound aloha: menos samba e mais trabalhar.
15:46 NotFound fbrito: that's my brazilian level ;)
15:47 fbrito ahahhaa, but in Brazil we say "menos trabalho e mais samba"
15:47 fbrito (less work and more samba)
15:48 NotFound fbrito: here is usually quoted that way, don't know if is a portuguese variant or just a Galician deformation.
15:48 NotFound Ah, you mean the order ;)
15:48 NotFound I was thinking about trabalho/trabalhar
15:50 Psyche^ joined #parrot
15:50 fbrito but you said "less samba and more working", but brazilian people don't like to work :(, so I think "less work and more samba" would be more appropriate :P
15:50 Psyche^ is now known as Patterner
15:50 NotFound But aloha isn't brazillian, I guess.
15:51 fbrito my sister (who taught english classes) reviewed my README translation and said that it was ok
15:52 NotFound Good
15:54 fbrito but maybe cotto wants someone from Parrot to review it
15:54 fbrito anyway, I have to leave now... see you guys
15:55 whiteknight later fbrito
15:55 NotFound fbrito: keep the good work, and the good samba.
15:55 fbrito left #parrot
15:58 NotFound plobsing: issue 7 acknoledged. It was fixed... but only in stage 0.
16:00 plobsing NotFound: closures are read-only. is this known, or a bug?
16:01 NotFound plobsing: read only? What do you mean?
16:02 nopaste "plobsing" at 192.168.1.3 pasted "[winxed] writing to closure variable test (fails)" (13 lines) at http://nopaste.snit.ch/27034
16:03 NotFound plobsing: ah, yes. This is intentional to avoid the need to reload lexicals in any usage. Use the =: operator.
16:04 plobsing OK, thanks. I suspected that guarding against closure writes would give winxed much of the horrible PIR output that NQP has.
16:07 NotFound plobsing: the same happens with attributes. I choose trade offs towards speed.
16:08 plobsing I agree with those tradeoffs. they also tend towards more sane PIR.
16:08 plobsing I just wish they were documented somewhere (maybe they are and I'm just being stupid)
16:09 whiteknight plobsing: what project are you working on?
16:10 NotFound plobsing: Did you know the 'Loch Ness monster' documentation system?
16:10 plobsing whiteknight: Ωη;)XD => Ohm - eta - wink - kzd
16:11 NotFound "All people believe it exists, some even claim have seen it personally. But if you want to see it, you never can"
16:12 whiteknight plobsing: brillian
16:12 whiteknight brilliant
16:16 plobsing NotFound: now that you jog my memory, I appear to have a copy of it here. whoever printed it out must have spent a fortune on all that white toner.
16:19 atrodo My last job was in the print industry, and I seriously had a salesman that had charged a customer for two color printing.  "We don't do two color printing." to which he retorted "It's black and white, that's two colors, right?"
16:28 whiteknight that's like charging a person for "one color printing" for a blank page
16:29 plobsing paper ain't free
16:29 NotFound Don't give them ideas...
16:30 NotFound Better go patent it.
16:34 Coke http://rt.perl.org/rt3//Publ​ic/Bug/Display.html?id=80668 - looks like a parrot bug
16:36 NotFound I fixed a heisenbug in library searching some days ago, maybe is related.
16:36 plobsing I suspect the different drive letters
16:37 Coke so, in our brave new world, should be we getting "merge" commits to master?
16:37 moritz when merging a branch, sure
16:37 Coke is the answer "no, but we tolerate them"?
16:37 Coke moritz: like https://github.com/parrot/parrot/commit/​532f1ce4695812892db4bf6a68b89173fa3b98ec
16:37 plobsing when merging a branch, we *want* merge commits. when merging regular commits, we don't.
16:38 Coke hurm. that's a bad example, as that is a merge of nol188
16:38 plobsing it all starts to seem really silly after you drink the git coolaid and realize there is no priviledged branch
16:39 whiteknight plobsing: not enforced, no. But there are by convention
16:39 Coke this one: https://github.com/parrot/parrot/commit/​39b8b702957a0ad783450933a0e60008eff18c80
16:39 Coke where we have a merge of master into master from parrot/parrot into parrot/parrot
16:39 whiteknight Coke: yeah, those cases could probably be avoided
16:39 moritz those should be avoided
16:40 whiteknight if it's a pain point, I can avoid those in the future
16:40 Coke just makes it harder to figure out what the hell is going on for those of us trying to review commits.
16:41 moritz do merge commits actually contain a diff?
16:41 NotFound I think a good way is: when your branch is ready, start a local branch from master, merge the branch into it, solve conflcts. Then start yet another local branch, merge the first into it. Then merge that last branch into master.
16:41 moritz I kinda thought they just bring in more commits from somewhere else (and those commits then have separate commit messages)
16:42 NotFound This is less noisy and error prone than merging master into the branch,
16:42 moritz but it doesn't sound easy :-)
16:42 Coke https://github.com/parrot/parrot/commit/​39b8b702957a0ad783450933a0e60008eff18c80 does seem to have a diff, yes,
16:43 moritz Coke: 'git show 39b8b702957a0ad783450933a0e60008eff18c80' doesn't show one
16:44 Coke moritz: ok. why does github & git show disagree, and should I have to care? ;)
16:44 moritz Coke: if you want to read the commit messages that describe the changes, you should care
16:45 Coke moritz: git diff does.
16:45 Coke moritz: this is what I'm saying. I see a diff for something that has no informative commit message.
16:45 moritz I guess gitk will show you the commits involved
16:47 Coke ok. I can't run gitk most of the time. Will check it out when I get home.
16:49 dalek winxed: r716 | NotFound++ | trunk/winxedst1.winxed:
16:49 dalek winxed: fix postincrement and postdecrement in void context, Issue 7, plobsing++
16:49 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=716
16:51 dmalcolm joined #parrot
16:53 nopaste "coke" at 192.168.1.3 pasted "build failure of embed_api2 for dichi" (1018 lines) at http://nopaste.snit.ch/27037
16:56 plobsing ... fun
16:57 Matt221 whiteknight: Can you do a quick review? https://github.com/mattrajca/parrot/commit​/e1d979dca972bd7ba370343f7b2a47910340a00a
16:59 NotFound Coke: there is a warning about an installed parrot in the Configure output.
16:59 Coke NotFound: yes. I get that every single time.
17:00 Coke for years now.
17:02 Coke (and I actually have the reverse problem - the installed version is linked against the build dir.)
17:02 Coke (remove the build dir? boom, installed parrot fails)
17:05 cotto ~~
17:07 whiteknight Matt221: sure. give me a minute
17:08 Matt221 P.S. - I can't seem to create a pull request for something thats not in the master branch?
17:08 Matt221 (embed_api2 in this case)
17:15 lucian joined #parrot
17:17 kapace_school joined #parrot
17:17 whiteknight hmmm, that's weird
17:19 whiteknight Matt221: The tests for the new functions need to go in t/src/embed/strings.t, not t/src/string.t
17:20 whiteknight if that file doesn't exist, create it.
17:24 whiteknight Matt221: And don't worry about a pull request. Just tell me when you are ready
17:26 Matt221 whiteknight: Here: https://github.com/mattrajc​a/parrot/commits/embed_api2
17:26 TimToady /e
17:30 Coke /i
17:32 dalek parrot/embed_api_wchar: e1d979d | (Matt Rajca)++ | / (3 files):
17:32 dalek parrot/embed_api_wchar: Add wchar_t variants for Parrot embedding API string interface
17:32 dalek parrot/embed_api_wchar: review: https://github.com/parrot/parrot/commit/e1d979dca9
17:32 dalek parrot/embed_api_wchar: cb7d717 | (Matt Rajca)++ | src/embed/strings.c:
17:32 dalek parrot/embed_api_wchar: Code formatting fixes
17:32 dalek parrot/embed_api_wchar: review: https://github.com/parrot/parrot/commit/cb7d7179e4
17:32 dalek parrot/embed_api_wchar: e3104aa | (Matt Rajca)++ | t/src/ (2 files):
17:32 dalek parrot/embed_api_wchar: Fixed folder structure
17:32 dalek parrot/embed_api_wchar: review: https://github.com/parrot/parrot/commit/e3104aad1c
17:33 whiteknight Matt221: okay, pulled. Give me a minute to review/test it and I'll close the Task
17:38 whiteknight I fixed the comments in the test file to show the right filename and the current year for copyright
17:38 whiteknight but that's not a big deal
17:44 dalek parrot/embed_api2: e1d979d | (Matt Rajca)++ | / (3 files):
17:44 dalek parrot/embed_api2: Add wchar_t variants for Parrot embedding API string interface
17:44 dalek parrot/embed_api2: review: https://github.com/parrot/parrot/commit/e1d979dca9
17:45 dalek parrot/embed_api2: cb7d717 | (Matt Rajca)++ | src/embed/strings.c:
17:45 dalek parrot/embed_api2: Code formatting fixes
17:45 dalek parrot/embed_api2: review: https://github.com/parrot/parrot/commit/cb7d7179e4
17:45 dalek parrot/embed_api2: e3104aa | (Matt Rajca)++ | t/src/ (2 files):
17:45 dalek parrot/embed_api2: Fixed folder structure
17:45 dalek parrot/embed_api2: review: https://github.com/parrot/parrot/commit/e3104aad1c
17:45 dalek parrot/embed_api2: e403c24 | Whiteknight++ | t/src/embed/strings.t:
17:45 dalek parrot/embed_api2: some small fixes to test file
17:45 dalek parrot/embed_api2: review: https://github.com/parrot/parrot/commit/e403c240cd
17:45 hudnix left #parrot
17:46 whiteknight Matt221: change the task to NeedsReview
17:47 Matt221 whiteknight: Done - sorry about delay
17:49 kapace_school its going to be more diifcult to catch up to you Matt221 :P
17:51 kapace_school left #parrot
17:52 bluescreen is there any way to get an OSX VM somewhere? want to test parrot build there and maybe gdb it
17:52 fbrito joined #parrot
17:53 NotFound bluescreen: ask Steve Jobs
17:53 bluescreen lol
17:53 fbrito Kovensky: ping
17:53 bluescreen hackintosh?
17:53 whiteknight Matt221: Closed
17:54 theory joined #parrot
17:54 whiteknight bluescreen: If you find a way, let me know. I would love to be able to test OSX locally
17:55 fbrito cotto ping
17:55 bluescreen there is a distro called hackintosh - 99.9% illegal - I'll try to install in a virtual box image
17:57 hudnix joined #parrot
18:02 Matt221 I know you can virtualize OS X Server (legally) but only in VMWare Fusion
18:04 cotto fbrito, pong
18:05 fbrito cotto have you found anyone to review my README translation? I asked my sister (who lived 1 year in USA and used to teach english classes) and she said it was OK
18:06 cotto I haven't.
18:06 fbrito should we wait for Kovensky?
18:07 Kovensky ?
18:07 fbrito oh, you are there :O
18:08 cotto Kovensky, do you speak Portuguese?
18:08 Kovensky yes
18:08 cotto can you verify that https://github.com/fernandobrito/parrot/comm​it/0e4a95c6c1b7487a7ab81e47c1ab375a20d9aee3 looks good
18:09 * Kovensky waits on firefox
18:10 cotto I'm not sure what we'll do about the Romanian translation.
18:13 whiteknight why did we offer a translation task if we don't have a dev to check it?
18:13 Kovensky PLATAFORMS <-- typo
18:13 fbrito ops :)
18:14 Kovensky otherwise it looks OKish, but I haven't read computer stuff in portuguese in a long time...
18:14 * Kovensky avoids them like the plague ;)
18:14 Kovensky so I dunno if your choice of wording for the technical terms is good or not
18:17 bluescreen left #parrot
18:17 cotto I'm starting to question the value of README translations.
18:17 whiteknight cotto: Maybe we shouldn't make any more tasks for it until we know if it's worthwhile
18:18 whiteknight if we had i18n support, making translations for that would be very beneficial
18:19 rurban left #parrot
18:19 cotto whiteknight, that might be a good idea.
18:19 rurban joined #parrot
18:21 whiteknight cotto: Yes, I would definitely like to have i18n support in place before next CGI
18:21 whiteknight if for no other reason than we get a huge influx of multilingual programmers for that period
18:21 whiteknight or, partially-multilingual
18:23 NotFound If you internationalize exception messages, be ready for lot of failing tests.
18:25 moritz well, testing against the text of an exception is a design fallacy anyway
18:25 moritz in fact I did some planning for a nicer exception system for perl 6
18:25 cotto moritz, unfortunately that doesn't mean we don't do it.
18:25 moritz when I get around to writing it down, maybe you want to steal some ideas :-)
18:25 cotto I love stealing ideas.
18:26 cotto dukeleto, I'm questioning the value of translating README.  It seems that many foreign language speakers prefer the English versions.  Unless you know that some contributors like them, I think we should focus on other tasks.
18:27 cotto fbrito, task approved
18:27 NotFound I dislike the spanish ones, can't talk for other languages.
18:28 cotto There's a request for Russian.  I know someone who can proofread that one.
18:28 whiteknight 90% of the test failures I've been fixing in the embed_api branch are for exact matches in error messages
18:29 whiteknight and when I removed a newline, *BAM*. like 50 tests were failing
18:29 cotto fbrito, http://socghop.appspot.com/gci/task/show/google​/gci2010/parrot_perl_foundations/t129202922193 would be really useful if you have the tuits.
18:32 fbrito cotto thank you :). hm, this GitHub hook task seems interesting, but I also liked the "Increasing 10% of Complex PMC code coverage" task
18:32 NotFound I'm thinking about adding a sub to Test;More that takes a exception number, a sub and a message as arguments, and report failure if the sub doesn't throw or throws any other type. Opinions?
18:32 bluescreen joined #parrot
18:32 NotFound A exception type number, I mean.
18:38 whiteknight NotFound: seems reasonable. In the future I would like to get rid of exception type numbers though
18:38 ligne joined #parrot
18:39 NotFound whiteknight: I was under the idea that the tests test the present ;)
18:42 whiteknight NotFound: right. I said it was reasonable
18:51 whiteknight To fix Coke's build problems in embed_api2 I'm probably going to need to get access to a Mac
18:51 rfw joined #parrot
18:58 Coke I am happy to provide realtime OSX build info.
18:59 Coke making the build warnings clean is an excellent start.
19:00 whiteknight I have a few warnings fixed locally now. Will push tonight when I get home
19:00 whiteknight I have a few more warnings I can fix easily, from looking at your make output
19:01 Coke hokay. I'll try again then.
19:01 Coke if we're desperate I can see if I can setup a generic "build parrot" account on my desktop at home, but I'd like to avoid that if possible.
19:02 NotFound Better tell Jobs to send a few free Macs to PaFo ;)
19:03 NotFound I promise stop ranting about iPhones for about a month if he does X-)
19:05 whiteknight What's getting me is that we haven't really touched pbc_merge in this branch yet
19:05 whiteknight I was planning to update it, but hadn't yet. It's still using all the old src/embed.c functions, none of which have been touched in this branch
19:07 whiteknight well, it looks like there was one change to it
19:08 whiteknight Coke: are you at your build machine, able to make and test a quick change?
19:08 Coke rebuilding with --cflags="-g" so I can get a better backtrace.
19:09 Coke whiteknight: I IRC from that machine, so if I'm chatting here, yes. Just fyi. what do you need tested?
19:11 nopaste "coke" at 192.168.1.3 pasted "better backtrace" (3 lines) at http://nopaste.snit.ch/27044
19:11 Coke was the GC interface changed at all?
19:11 Coke looks like GC might not have been initialized.
19:12 Coke seems like a simple enough thing to avoid that sf. no clue if it'll DTRT, though.
19:13 whiteknight Coke in src/pbc_merge.c, add #include "parrot/embed.h" where the rest of the #includes are
19:15 whiteknight the embed_api2 branch doesn't touch the GC at all, so this failure is extremely weird
19:15 Coke ... that fixed it.
19:16 whiteknight okay. That was supposed to be there and I deleted it
19:16 whiteknight and without that, the compiler auto-defined the function with the wrong type semantics, which probably caused a pointer to get garbled
19:17 whiteknight Coke++ for sending in the build log with the warning messages in it
19:18 whiteknight you want to commit that, or you want me to?
19:19 mtk left #parrot
19:19 mtk joined #parrot
19:21 ok22_ joined #parrot
19:24 ok22 left #parrot
19:24 ok22_ is now known as ok22
19:27 Coke ... is that .h overriding an existing definition or something?
19:27 Coke Just wondering how it's a segfault and not a build error.
19:28 Coke yah, I'll commit it (needs root.in fix)
19:32 whiteknight Coke: it's a build warning. Parrot_new is not being defined. The compiler assumes that it returns an int, not a pointer, and then generates the wrong machine code based on that wrong assumption
19:32 dalek parrot/embed_api2: 5f5c181 | Coke++ | / (2 files):
19:32 dalek parrot/embed_api2: Fix build on OS X.
19:32 dalek parrot/embed_api2: review: https://github.com/parrot/parrot/commit/5f5c18131b
19:32 whiteknight you're on PPC, right?
19:32 Coke nope. x86
19:32 whiteknight oh. Not as obvious but still the problem
19:40 Coke I'll give you another run after your push later tonight to see if there's any dangling warnings.
19:43 bluescreen whiteknight++ you save me from installing an OSX... tough i can keep on research mode
19:48 whiteknight Coke++
19:49 whiteknight bluescreen: I have wanted to get a copy of OSX for a while now, but I don't want to spend a bunch of money on it
19:49 whiteknight so it's always a good thing to keep in mind
19:50 Coke we have an ex parrot developer at apple, but ISTR they don't have any kind of build farm access.
19:51 whiteknight I know a few Mac people. Next time one of them upgrades I'll try to "borrow without intent to return" their old mac
19:52 bluescreen can we add any HW to our existent build farm?
19:53 bluescreen maybe we can ask some founding to buy a Mac and put it in smolder
19:53 Coke we don't really have a build farm.
19:54 whiteknight bluescreen: you *are* the build farm
19:54 bluescreen I know we borrowing one.. but maybe the allow us to put something in there
19:54 Coke has taptinder been resurrected yet?
19:54 bluescreen asking too much?
19:54 whiteknight I don't even know what taptinder is
19:54 Coke bluescreen: if we have access to a build farm, it's not something that we have access to add machines to.
19:54 Coke whiteknight: the buildbots that use to carp in here when someone broke the build on windows.
19:55 Coke they went silent around git-switch time.
19:55 whiteknight We do have a certain amount of access to the GCC compile farm, dukeleto can probably explain more about that
19:55 whiteknight Coke: ah, okay.
19:55 whiteknight a good build-bot would be very nice to have in here
19:55 Coke mj41 was the taptinder guy.
19:56 Coke IIRC>
19:56 Coke I'd ask aloha...
19:56 Coke when it comes back, I can put my osx box in as a client.
20:00 whiteknight We really need to put together a list of software and services that we rely on as a community
20:38 fbrito left #parrot
20:48 whiteknight some days I really hate ASP.NET.
20:48 whiteknight ...I've never experienced the other days
20:53 bluescreen left #parrot
20:55 nwellnhof joined #parrot
21:03 lucian left #parrot
21:03 bluescreen joined #parrot
21:22 bluescreen left #parrot
21:26 linge joined #parrot
21:29 rurban_ joined #parrot
21:30 whiteknight left #parrot
21:32 rurban left #parrot
21:32 rurban_ is now known as rurban
21:38 linge left #parrot
21:59 nwellnhof left #parrot
22:04 Andy ping kid51
22:07 mtk left #parrot
22:08 Andy I'm so frustrated with trac. :-(
22:09 Andy Once again it doesn't recognize my login.
22:09 Andy I've cleared cookies.
22:10 Andy It's like it's punishing me for only being an ocassional user.
22:18 Coke I wonder if your account was never really fixed the first time this happened.
22:18 Coke victim of a 2 year old upgrade.
22:26 NotFound Maybe should develop a game, "Join the build farm" with graphics in the style of farm games.
22:27 NotFound And give the players good harvest depending on the reports obtained.
22:31 gg411 joined #parrot
22:45 gg411 left #parrot
22:46 Coke left #parrot
22:46 Coke joined #parrot
22:47 AndChat| joined #parrot
22:47 AndChat| left #parrot
22:53 Eclesia joined #parrot
22:56 bluescreen joined #parrot
22:56 Andy left #parrot
23:04 Kapace_ NotFound: I didn't want to set Connection: Keep-Alive by default, but I wasn't sure where to give the user the option.. so I just set it for testing
23:04 Kapace_ (because you need alive connections for chunked encoding)
23:04 Eclesia left #parrot
23:13 dukeleto ~~
23:14 dalek parrot/lorito: 9a53192 | dukeleto++ | src/pmc/loritocontext.pmc:
23:14 dalek parrot/lorito: Create a skeleton init VTABLE for LoritoContext
23:14 dalek parrot/lorito: review: https://github.com/parrot/parrot/commit/9a53192458
23:19 fperrad left #parrot
23:28 kid51 joined #parrot
23:33 cosimo left #parrot
23:37 kid51 I've done a git pull on master, and I see that this file has been added to the top-level of our repository:
23:37 kid51 $ cat test_data
23:37 kid51 blah
23:37 kid51 Is David Czech on channel?
23:38 kid51 This file does not appear to be referenced in any other file in the distro (other than MANIFEST), so it appears to be superfluous.
23:40 Kapace_ kid51: thats me
23:41 Kapace_ test_data was for the md5task, how did that get there?
23:41 kid51 Hunch:  At some point, you typed: 'git commit -a'
23:41 kid51 I gather we can remove it, correct?
23:42 Kapace_ yes,
23:42 Kapace_ i can't believe i didn't clean that up before commiting! sorry
23:42 kid51 np, that's what us cage-cleaners are for ;-)
23:42 Kapace_ ok thank you :)
23:43 Kapace_ oh and md5task also shouldn't exist
23:43 kid51 Kapace_:  Could I ask:  Where are you located, terrestrially speaking?
23:44 linge joined #parrot
23:44 Kapace_ kid51: sure, uh Canada
23:44 Kapace_ British Columbia
23:45 kid51 Aha!  I figured Canada because of 'telus.net'.  I've spoken on Parrot in Toronto several times and was wondering if you were around there.
23:45 kid51 So, in kilometers, you're not all that far away from cotto and particle (Seattle-area)
23:45 Kapace_ ah, cool
23:45 Kapace_ i went to seattle once.. couldn't find parking
23:45 kid51 Can I remove the entire md5task/ directory?
23:46 kid51 I went to Seattle once, was cited for jaywalking!
23:46 kid51 I skipped out on Traffic Safety School.
23:46 Kapace_ kid51: yes
23:46 Kapace_ heh
23:48 dalek parrot: 65b4794 | jkeenan++ | / (4 files):
23:48 dalek parrot: Remove some files that were accidentally committed.
23:48 dalek parrot: review: https://github.com/parrot/parrot/commit/65b4794692
23:52 whiteknight joined #parrot
23:53 linge left #parrot

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

Parrot | source cross referenced