Camelia, the Perl 6 bug

IRC log for #parrot, 2010-12-29

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:00 NotFound Coke: if people insist on consting parameters disregarding warnings and common sense, I can't fight alone.
00:00 Coke NotFound: help me fix one of the broken files, I'll see what I can do about the rest.
00:00 NotFound Coke: wich one?
00:01 Coke rebuilding at HEAD...
00:02 Coke frontend/parrot_debugger/main.c
00:02 NotFound plumage installed from the new repo is able to install rakudo and it works fine, so I can say that plumage isn't broken.
00:02 Coke first is that Parrot_set_config_hash "was not declared in scope"
00:03 Coke that's declared in parrot/global_setup.h, which is included from parrot.h, which that file includes.
00:04 NotFound The functions in that file aren't PARROT_EXPORT'ed
00:04 whiteknight hmmm... I suspect parrot/global_setup.h should not contain Parrot_set_config_hash anymore
00:04 whiteknight let me take a look at it
00:05 NotFound I see Parrot_set_config_hash_pmc
00:06 Yuki`N cotto, Unfortunately since I have the text in pre tags, we can't use wiki formatting.
00:06 Yuki`N I don't see any better way to preserve the whitespace of the original source though.
00:06 Coke gah. make headerizer is failing also.
00:07 Coke NotFound: gah. good catch. :P
00:07 dukeleto ~~
00:07 NotFound Coke: if I make frontend/parrot_debugger/main.o in a C build I see a lot of warnings. No need for a C++ build to note that there are pitfalls.
00:08 cotto_work Yuki`N: you could use a fixed-width font and change spaces to   .  Generated code can be ugly and we have a codingstd test against tabs.
00:09 Yuki`N Oh I can insert raw HTML, problem solved.
00:10 cotto_work did you find {{{#html...}}}?
00:10 Yuki`N Yeah.
00:10 cotto_work ok
00:11 cotto_work that'll be fine
00:11 patspam joined #parrot
00:11 patspam left #parrot
00:15 dalek tracwiki: v1 | Yuki`N++ | ParrotHackList
00:15 dalek tracwiki: Initial script (first revision) output.
00:15 dalek tracwiki: http://trac.parrot.org/parrot/wiki/Par​rotHackList?version=1&action=diff
00:16 dukeleto cotto_work: hola
00:16 NotFound Parrot_set_config_hash now takes an iterpreter as argument? The we can hardly call it before creating an interpreter,
00:16 dukeleto cotto_work: i have some plumage tuits, if those are needed. I wrote the plumage test harness and a bunch of tests
00:17 nwellnhof the whole thing with the thawed config hash is completely braindead anyway.
00:18 NotFound This is fuckingly wrong. If we shoud use Parrot_set_config_hash from several places it *must* be in a header file. We have header files for something.
00:18 dukeleto nwellnhof: i agree
00:19 whiteknight NotFound: Parrot_set_config_hash is defined in src/parrot_config.c, which is a generated file and is not included in libparrot
00:19 NotFound nwellnhof: agreed, but while we are using it, it should be used is a reasonable way.
00:19 NotFound whiteknight: anyway. Is used, and is linked.
00:20 * dukeleto reads #ps from today
00:21 whiteknight NotFound: I'll fix this now, we can put together a header file for it if you want
00:23 silug left #parrot
00:24 NotFound The other problems seem to be fixable by including parrot/embed.h
00:25 NotFound whiteknight: I think that if embedders may need to call it, it should be in embed.h
00:25 Yuki`N to all who are interested, comments on format revision 2 are welcome. http://trac.parrot.org/parrot/wiki/ParrotHackList
00:26 NotFound (and if embedders will never need to use it, should be killed)
00:26 whiteknight NotFound: parrot/embed.h is the old API, we're going to deprecate it. parrot_debugger uses the old style and will get updated soon
00:26 dukeleto cotto_work: i plan to automate your GOAL 6
00:26 NotFound In api.h then
00:26 whiteknight I'm putting together a temporary fix now, I'll have to do the complete update this weekend
00:26 dukeleto cotto_work: manually doing it is setting ourselves up for failure
00:27 cotto_work dukeleto: that's exactly what I'm hoping for.
00:27 whiteknight NotFound: yes, api.h should include all functions an embedder needs
00:27 dukeleto cotto_work: we will use the GCC compile farm to run tests across all HLLs we care about
00:27 cotto_work do it manually as a stopgap until it can be automated through plumage
00:27 dukeleto cotto_work: post-commit hooks will run tests suites across various parrots with different configure options
00:28 dukeleto cotto_work: i don't need plumage to do it
00:28 dukeleto there is also parrothandler
00:28 cotto_work even better
00:28 dukeleto but i have a homebrew solution anyway
00:28 dukeleto i think parrothandler will make it even easier
00:29 NotFound whiteknight: but if the plan is to get rid of the need for the config object file, will be better to get api.h free from it.
00:29 whiteknight NotFound: I don't understand what you mean. The config object file is fine. We don't need to get rid of it
00:29 dukeleto cotto_work: what is the simplest useful config for running HLL tests? Just running a Rakudo spectest?
00:29 cotto_work Yuki`N: the last link doesn't work
00:30 cotto_work Yuki`N: it's looking at a generated file
00:30 dukeleto https://github.com/brianwisti/parrot-handler
00:31 Yuki`N cotto, oh.
00:31 Yuki`N My fault for not cleaning first then.
00:31 NotFound whiteknight: then, fine
00:32 dukeleto plobsing: ping
00:32 dukeleto plobsing: i would like to talk more about the Complex PMC
00:35 NotFound dukeleto: using plumage to load and test HLL is a way to keep plumage well tested.
00:35 cotto_work Yuki`N: you might add a warning to the inline POD (or just bail if you find Makefile)
00:36 Yuki`N Oh ok.
00:36 dukeleto NotFound: that is true
00:37 dukeleto NotFound: so we should make the compile farm run plumage tests and installs? That can be done.
00:38 NotFound dukeleto: and keeping its plumage updated, of course.
00:38 dalek tracwiki: v2 | Yuki`N++ | ParrotHackList
00:38 dalek tracwiki: Revision 2
00:38 dalek tracwiki: http://trac.parrot.org/parrot/wiki/Par​rotHackList?version=2&action=diff
00:38 dalek tracwiki: v3 | Yuki`N++ | ParrotHackList
00:38 dalek tracwiki: Revision 3, change fonts.
00:38 dalek tracwiki: http://trac.parrot.org/parrot/wiki/Par​rotHackList?version=3&action=diff
00:38 Yuki`N Ok, so I'm ready to submit a pull request.
00:39 Yuki`N http://socghop.appspot.com/gci/task/show/google​/gci2010/parrot_perl_foundations/t129315727394 here's the task
00:39 dukeleto Yuki`N: i will look
00:39 NotFound Talking about keeping updated...
00:39 Yuki`N Actually, my source tree is rather unclean for some reason.
00:39 Yuki`N I'll submit a patch.
00:40 dukeleto Yuki`N: ok
00:40 Yuki`N If you could accept the task, I will upload a tar.
00:41 cognominal joined #parrot
00:42 whiteknight I didn't realize that the runcore selection mechanism uses a string internally
00:42 dukeleto Yuki`N: no tar uploads
00:42 Yuki`N Really?
00:42 Yuki`N Hm.
00:42 whiteknight There's no reason why the embed API should take a string, turn it into a flag, then turn that back into a string, and pass that into the runcore selector
00:42 dukeleto Yuki`N: you have just a single file, no?
00:42 Yuki`N True.
00:42 dukeleto Yuki`N: i just think a tar is unnecessary
00:42 Yuki`N I was going to upload patches.
00:42 dukeleto Yuki`N: just give a link to the new file
00:43 Yuki`N Ok.
00:43 dukeleto Yuki`N: it will probably live in tools/dev or something
00:43 dukeleto Yuki`N: make sure it looks like other stuff in tools/dev, with POD and an example of how to use it
00:43 Yuki`N Way ahead of you.
00:43 Yuki`N https://github.com/nol888/parrot/b​lob/master/tools/dev/findhacks.pl
00:43 dukeleto Yuki`N: glad to hear :)
00:44 dukeleto Yuki`N: accepted
00:45 Yuki`N Submitted.
00:45 dalek winxed: r725 | NotFound++ | trunk/ (3 files):
00:45 dalek winxed: update installable files
00:45 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=725
00:46 Yuki`N dukeleto, can you take care of removing http://socghop.appspot.com/gci/task/show/google​/gci2010/parrot_perl_foundations/t129315789586 since it's already done by someone.
00:46 NotFound winxed installed from plumage works fine again
00:47 Yuki`N >So he's going to get another 60 points.
00:47 Yuki`N Fuck jumpyshoes.
00:47 Yuki`N I give up.
00:47 cotto_work ?
00:47 Yuki`N He's apparently done 60 points worth of tasks
00:47 Yuki`N And is going to get those 60 points in the next 30 or so minutes.
00:48 whiteknight Yuki`N: how do you know that?
00:48 Yuki`N https://groups.google.com/d/msg/gc​i-discuss/uPktT0t-gqU/GHdJamlNSFsJ
00:48 Yuki`N That's ridiculous.
00:49 dukeleto yep, i just read it too
00:49 NotFound www.parrot.org and trac.parrot.org looks nice in a Toshiba Folio 100 :)
00:49 dukeleto they are batching tasks.
00:49 Yuki`N I do not believe sincerely that 15 "difficult" tasks can be completed in about 3 days.
00:49 dukeleto Yuki`N: the definition of "difficult" needs to be more well-defined, i agree
00:49 Yuki`N And at this point he is just mining points.
00:50 dukeleto Yuki`N: but i also doubt that contest rules will be changed mid-way. Things will be different next year, though.
00:50 Yuki`N While the rest of us have to actually work to find that many tasks.
00:50 Yuki`N Never mind complete them.
00:50 dukeleto Yuki`N: as long as you stay in the top 10, i wouldn't worry too much. Ignore those shenanigans and just continue to do good work (as you have done)
00:51 Kapace_ I don't think jumpyshoes is the only one
00:51 * dukeleto plans to do a task dump soon. Needs to make some jet fuel
00:51 Yuki`N I just think it's unjust for rfw to get first stolen from him.
00:51 Kapace_ I know someone who has done like 16 dutch translation tasks
00:51 Kapace_ I can't keep up
00:51 Yuki`N Yes, but see translation tasks are usually rated at 1 or 2 points.
00:51 Kapace_ no but these are all 4 points
00:51 Yuki`N Well that's the same concept, only jumpyshoes has done it on a much larger scale.
00:51 Kapace_ http://www.google-melange.com/gci/student_ran​king/show_details/google/gci2010/aureianimus
00:52 Kapace_ yeah
00:52 rfw oh god apertium
00:52 Yuki`N People are comparing asm functions in lines of code to other hll's.
00:52 NotFound "I would just like to point out two one-liners http://socghop.appspot.com/gci/task/s​how/google/gci2010/ssf/t129152150430 and http://socghop.appspot.com/gci/task/s​how/google/gci2010/ssf/t129152118642 , against a 879-line"
00:52 rfw NotFound: it's fun to note that the 879-line diff was for multiple functions
00:52 Yuki`N Just those arguments alone are just completely bs.
00:52 whiteknight the problem is that porting a C function to assembly is hard, if you don't know how to tell a compiler to do it for you
00:52 NotFound This is wrong. The number of lines has nothing to do with the amount of work or its quality.
00:53 Kapace_ NotFound: i agree
00:53 Kapace_ theres a quote:
00:53 NotFound I'll obviously prefer a *right* one line fix than a big diff.
00:53 Kapace_ "measuring progress by lines of code is like measuring aircraft progress by weight" or somelike like that
00:53 Yuki`N whiteknight, that would be the case, however, many of the asm functions that need to be written have already-written analogues.
00:53 NotFound Kapace_: yeah, I know it
00:54 Yuki`N And most of the algorithms are the same, the only things that differ are data size or small factors.
00:54 whiteknight Yuki'N: Creating a "listing file" of assembly from C code is trivialy easy
00:54 whiteknight I think it's gcc -s
00:54 NotFound In fact, I've done in parrot several one line fixes of difficult bugs.
00:55 Yuki`N Yeah, these are written to use SIMD functions.
00:55 whiteknight combine that with gcc -s -o2, and suddenly you're producing optimized assembly files for any architecture in less than 10 seconds
00:55 whiteknight Yuki`N: right, -o2 will produce those in some cases
00:55 smash left #parrot
00:55 whiteknight he could hand-tune a few places to make it look better, but the hard work can all be done by computer
00:56 Yuki`N x264 assembly has a lot of macros that make writing the equivalent asm version of a c function easier.
00:56 Yuki`N Yeah that is true.
00:56 Yuki`N The hardest part imo is determining what the c function is trying to do.
00:56 Yuki`N And Dark_Shikari gives a lot of assistance.
00:57 Yuki`N The problem that I see is that that task type has a learning curve, but after you go over that it is no longer "difficult."
00:57 NotFound And BTW I don't think the number of lines of complaints against the rules of the contest qualify for anything ;)
00:57 Yuki`N Again, it's a problem with the rules and the definitions put forth by google.
00:57 Yuki`N NotFound, :P
00:58 whiteknight Yuki`N: it's not really a problem with google. If jumpyshoes can do a million and one tasks, they shouldn't be listed as difficult
00:58 rfw i've given up on trying to catch up to him
00:58 kid51 joined #parrot
00:58 rfw i think i'll do fuck all for the rest of gci
00:59 whiteknight if they are easy and fast to do, you shouldn't be listed as difficult
00:59 whiteknight that's a problem of the mentoring organization, not google
00:59 dukeleto rfw: i think you shouldn't be discouraged by jumpyshoes
00:59 dukeleto rfw: google will sort things out, i know the people who run GCI and they are very fair
00:59 rfw i hope so
00:59 NotFound Also, you are learning a hard lesson: programming merits aren't always appreciated with justice.
01:00 dukeleto rfw: if they see that someone gamed the system, they will even things out somehow
01:00 kid51 rfw: Parrot will repay you in beer ... when you're old enough!
01:00 rfw kid51: :D
01:00 rfw dukeleto: i'd still be happy to do anything for parrot if you want
01:00 NotFound But seriously: we appreciate your work.
01:00 dukeleto rfw: awesome! i know we don't have lots of tasks right now, but I am thinking some up
01:00 rfw dukeleto: ah cool
01:00 dukeleto rfw: i would like some documentation to get written for PL/Parrot
01:00 dukeleto rfw: check out http://pl.parrot.org
01:00 Yuki`N :P
01:00 Yuki`N I will also help out.
01:00 kid51 rfw: We'll dispatch our New Zealand emissary to thank you!
01:01 rfw kid51: haha
01:01 rfw dukeleto: i have some stuff to follow up for gluon first, though
01:01 dukeleto rfw: we only have a tiny bit of docs. I want a nice manual of examples for how to write various kinds of functions in PIR and Perl6
01:01 Yuki`N I like this project more than any of the others so far, I can say that. :P
01:01 NotFound parrots fly everywhere!
01:01 dukeleto rfw: no worries
01:01 dukeleto Yuki`N: you want to work on PL/Parrot docs?
01:02 Yuki`N Documentation isn't my strong point as I'm not a good writer but if I see some I can help out with then I will.
01:02 dukeleto Yuki`N: it would mostly be reading the tests of PL/Parrot to see how it works, then writing docs and examples
01:02 Yuki`N I remember rfw's first parrot tasks were some extensions to the linear algebra module
01:02 rfw tests, actually
01:02 rfw those were fun :D
01:02 dukeleto Yuki`N: you can basically use the tests in the docs, just add a bit of explanation to them
01:02 Yuki`N Hm, there were some tasks to do domething like add different types of matrices though.
01:03 Yuki`N Ah, that's nice.
01:03 dukeleto rfw: have you ever played with GSL ?
01:03 rfw ah, nope
01:03 Yuki`N Seeing as I wrote a good amount of the tests. :P
01:03 dukeleto i want to have GSL bindings to parrot
01:03 rfw hm i'd have to brush up a lot on my c
01:03 whiteknight Yuki`N, rfw: are either of you getting old enough for university?
01:03 rfw whiteknight: probaboy going next year
01:03 rfw test results pending
01:04 whiteknight because I think you guys would make great GSoC students if you want to do that
01:04 whiteknight and we have many bigger, funner projects for that
01:04 NotFound whiteknight: Are you plannig to create a Pafo University? ;)
01:04 rfw whiteknight: i think you mean old enough for gsoc
01:04 rfw since gsoc has the 18 age restriction
01:05 rfw you can be <18 and in university :p
01:05 * rfw is probably going to still be 16 for the first month in university
01:05 kid51 Yes, but you probably have to be >= 18 to sign a CLA to get a Parrot commit bit
01:05 kid51 :-(
01:05 rfw :(
01:06 dukeleto Yuki`N: i don't have much time to make tasks in a near future, but i can tell you to work on something and them make a task later :)
01:06 dukeleto Yuki`N: do you want to write docs for PL/Parrot ?
01:06 Yuki`N Ah/
01:06 whiteknight kid51: that's not a problem anymore. A GSoC student can do all their work in a fork
01:06 Yuki`N whiteknight, I'm in junior year.
01:06 NotFound kid51: I think a signed autorization from parents or legal tutors qualifies.
01:06 dukeleto Yuki`N: there is a pl/parrot repo on github
01:06 dukeleto Yuki`N: http://github.com/leto/plparrot
01:06 rfw Yuki`N: high school junior? i thought you were older than me
01:07 Yuki`N I also have a part-time job as a developer at a local university.
01:07 Yuki`N Nope.
01:07 Yuki`N HS junior.
01:07 kid51 NotFound: Glad to know that!
01:07 rfw when's your birthday again
01:07 Yuki`N august
01:07 Yuki`N I'm 16.
01:07 rfw what year
01:08 dukeleto Yuki`N: the task is to improve the documentation of PL/Parrot in any way you see fit. Spend a day or two on it, and report back. I will make a task soon
01:08 kid51 Coke: ping
01:08 dukeleto Yuki`N: i would like better html docs on the site
01:08 * NotFound feeling really old
01:08 dukeleto Yuki`N: the PL/Parrot website lives in the html/ dir of the repo
01:08 Yuki`N I see.
01:08 Yuki`N rfw, 1994
01:08 Yuki`N I'm 16.
01:08 Yuki`N You can do math. :(
01:08 rfw Yuki`N: oh my brain switched itself off
01:08 rfw after reading all the google groups posts
01:09 rfw i'm probably still shorter though
01:09 dukeleto Yuki`N: i would like to see different docs for different types of users, i.e docs for DBA's on how to install/maintain it, and docs for end users who are writing stored procedures
01:09 dukeleto rfw: what kind of parrot tasks are you interested in?
01:09 rfw dukeleto: i don't know :D
01:09 kid51 What file is it that has all the documentation about different 'make' targets?
01:09 Yuki`N fffff
01:09 dukeleto rfw: i need help setting up a continuous integration system for Parrot on our compile farm
01:10 Yuki`N firefox crashed while I was writing a mailing list response.
01:10 rfw Yuki`N: LOL
01:10 dukeleto rfw: i have some pieces of the puzzle but they need to be glued together properly
01:10 Yuki`N I would prefer code over documentation but I suppose I can take either.
01:10 rfw dukeleto: ah okay
01:10 Yuki`N `-`
01:11 rfw dukeleto: i've never had the hardware to set up a CI farm, heh
01:11 Yuki`N Damnit.
01:11 Yuki`N xchat cleared my clipboard.
01:11 dukeleto rfw: we need to test parrot master and selected branches on various versions of perl and run HLL tests, such as the Rakudo spectest and the plumage test suite
01:11 dukeleto rfw: we have access to the GCC compile farm
01:11 dukeleto rfw: we can run parrot tests on a playstation 3!
01:11 rfw haha
01:11 rfw when's the first game written in pir coming out ;)
01:12 dukeleto rfw: we need a nice web interface to it, that is what i need help with. A mostly static interface, but we could use javascript to make it more interactive and pretty
01:12 rfw ah
01:12 rfw i don't have to write it in perl do i
01:12 dukeleto rfw: i think whiteknight is planning on rewriting SimCity in PIR ...
01:12 rfw awesome
01:12 Kapace_ what is continuous integration?
01:12 dukeleto Kapace_: continually running tests on every commit
01:12 rfw Kapace_: it's when you continually rebuild new revisions of software
01:12 NotFound dukeleto: Sin City?
01:13 Kapace_ oh
01:13 rfw Yuki`N: rep says he would have lunch with us
01:13 dukeleto rfw: basically, i will give you a datastructure, just as a directory of JSON files or YAML files, and you need to generate a static site that shows off the data, graphs the performance of stuff against each other, etc.
01:13 rfw dukeleto: ah
01:14 rfw dukeleto: funny, i did the other end (generating data etc.) for x264
01:14 dukeleto rfw: it could be written in any language, which would you prefer?
01:14 rfw python :D
01:14 dukeleto rfw: you could help with that end as well
01:14 dukeleto rfw: that is fine with me, i don't care about the language
01:14 Yuki`N rfw, fuck yes.
01:14 Yuki`N Haha, ps3 testing.
01:15 rfw dukeleto: what CI platform are you using
01:15 dukeleto last time i ran the parrot test suite on a PS3, it went into an infinite loop, but "make coretest" passed ;)
01:15 NotFound dukeleto: if you have the data in json files, make that files addresable from HTTP
01:15 Yuki`N I can help out with this CI testing if you're willing to let me help as well.
01:15 dukeleto rfw: bash and for loops ;)
01:15 rfw dukeleto: oh dear
01:15 dukeleto NotFound: yes, good idea
01:15 rfw dukeleto: that's just exact same thing dark_shikari was doing with x264
01:15 rfw s/just/the/
01:15 rfw how did that get in there
01:15 dukeleto rfw: we can use buildbot or something else, but i want the simplest possible infrastructure
01:16 dukeleto rfw: i am thinking of just using git hooks
01:16 rfw ah
01:16 rfw so, build -> run tests
01:16 dukeleto rfw: we will have a cron job that keeps all git repos of parrot and all HLLs up-to-date
01:16 NotFound dukeleto: then we can download aand process them with parrot tools, in addition to web usages.
01:16 dukeleto rfw: yep
01:16 rfw awesome
01:17 * NotFound likes eating our own dogfood
01:17 dukeleto rfw: then we have git hooks that submit jobs or something that say "generate data for sha1 deadcoffee of parrot.git"
01:17 rfw ah
01:18 dukeleto rfw: and that can then be turned into "test all HLLs with the newest parrot.git sha1"
01:18 dalek parrot: f9fdda1 | Whiteknight++ | config/gen/makefiles/root.in:
01:18 dalek parrot: make headerizer fails if we haven't made src/extend_vtable.c first. So, build it first
01:18 dalek parrot: review: https://github.com/parrot/parrot/commit/f9fdda13ab
01:18 dalek parrot: c5d71e3 | Whiteknight++ | / (4 files):
01:18 dalek parrot: Fix the g++ build. Add a new function to src/embed.c to handle legacy-style setting the config hash by byte array instead of PMC. Add some new functions to the output of parrot_config_c.pl to help aide legacy support without needing those generated files to include more header files
01:18 dalek parrot: review: https://github.com/parrot/parrot/commit/c5d71e3590
01:18 dukeleto rfw: and then we have a web interface to all that data, as well as some email notifications if Bad Things Happen
01:18 Yuki`N http://pastebin.com/5SVLHndr rfw read please
01:18 rfw aha, got it
01:18 dukeleto who needs gci tasks right now?
01:18 dukeleto we need to increase test coverage of the MappedByteArray PMC, it is like 7% right now
01:18 dukeleto that is at least 2 tasks to get it up to 100%
01:19 rfw Yuki`N: lol that quote
01:19 dukeleto that is another task y'all can work on
01:19 Yuki`N rfw, yes.
01:19 * dukeleto throws tasks into the GCI snake pit
01:19 Yuki`N haha
01:19 rfw Yuki`N: looks good to me
01:19 Yuki`N rfw, if you don't mind I'm interested in that CI testing as well.
01:19 rfw Yuki`N: sure
01:19 Yuki`N Perhaps we can work on it together.
01:19 Yuki`N I need to perfect my python anyway.
01:19 rfw i'm not going to do it as part of gci though
01:19 rfw since i've sort of slipped into apathy mode about gci
01:20 Yuki`N Think like Jumpyshoes
01:20 Yuki`N I'll do some work.
01:20 dukeleto rfw: just do some fun tasks that you enjoy
01:20 Yuki`N Then claim credit en masse later.
01:20 dukeleto Yuki`N: that is a good way to look at it
01:20 rfw Yuki`N: lolol
01:20 rfw dukeleto: there aren't any more :(
01:20 Yuki`N The point imo is to have fun and also get recognition.
01:21 dukeleto rfw: i am telling you tasks right now :) I just don't have them in the system right now
01:21 Yuki`N However jumpyshoes is bastardizing it, lol.
01:21 * kid51 listens to the "Bach and Bluegrass" segment of WKCR's Bach festival (wkcr.org)
01:21 rfw dukeleto: heh, if you insist
01:21 rfw dukeleto: also you should give Yuki`N +o
01:21 dukeleto rfw: there is PL/Parrot, MappedByteArray, continuous integration. which do you want to hack on?
01:21 Yuki`N Posted to gci-discuss.
01:21 rfw CI please
01:22 Yuki`N aloha, coverage?
01:22 aloha Yuki`N: coverage is http://cv.perl6.cz or http://tapir2.ro.vutbr.cz/cover/cover-results/
01:22 whiteknight cotto: ping
01:22 NotFound whiteknight: the problem with that config approach is that the functions are not decorated, so tou can't link a C++ embed with a C built parrot or viceversa
01:22 dukeleto rfw: ok. assume that you have a directory which has every parrot release pre-compiled in a directory, as well as at least 1 copy of the parrot git repo
01:22 Yuki`N MappedByteArray is riddled with C preprocessor statements.
01:22 Yuki`N Testing all branches is going to be a pain.
01:22 whiteknight NotFound: what functions?
01:22 Yuki`N And perhaps impossible, in one run.
01:23 rfw dukeleto: why not integrate it directly with git
01:23 dukeleto Yuki`N: yeah, it will be fun :) we don't need all branches, just some important ones
01:23 NotFound whiteknight: Parrot_get_config_hash_bytes et al
01:23 dukeleto rfw: sure
01:23 whiteknight NotFound: those are just temporary
01:23 dukeleto rfw: that is fine, but i like having a directory for releases, because i want to do benchmarking across them
01:23 rfw dukeleto: ah
01:23 dukeleto rfw: but just the git repo is fine too
01:23 cotto_work whiteknight: pong
01:23 NotFound whiteknight: there are temporary hacks in parrot that last for years
01:24 rfw dukeleto: can Yuki`N work on it with me
01:24 dukeleto rfw: do you want to research what the best solution for us is? I am not quite sure
01:24 whiteknight cotto_work: TT #1910. I want to review that old embedding API
01:24 rfw dukeleto: i'm not quite sure either
01:24 whiteknight NotFound: I'm very motivated to see this issue be resolved
01:24 dukeleto rfw: yes, you and Yuki`N can work on it. I will create a task for each of you that is a different part of this
01:24 rfw i'm not very clear on CI
01:24 Yuki`N dukeleto, I just realized that I'm going to need a file to map. What is the working directory for tests as they're normally run
01:24 rfw i understand the rationale behind it but
01:25 dukeleto rfw: ok, take a look at jitterbug on github
01:25 rfw as for proposing actual solutions i'm at a loss
01:25 rfw A just-in-time image creator for pretty text headers — Read more?
01:25 NotFound whiteknight: Glad to know, but IMO in the meantime having a declaration in a header will not harm.
01:25 rfw oh, small smoker for Perl projects hosted on github
01:25 dukeleto Yuki`N: tests are run from the root of the checkout. you can put test data files in the same directory as the test
01:25 dukeleto rfw: we need to test that parrot works with various versions of perl, so jitterbug does a lot that we want
01:25 whiteknight NotFound: if you want a header, that's fine. This works without one and I think that's fine for now
01:25 Yuki`N Ah.
01:26 Yuki`N So I'd have to go t/pmc/mappedbytearray.t
01:26 rfw dukeleto: oh, so over various versions of perl too?
01:26 dukeleto anch)
01:26 dukeleto rfw: we want to make sure Parrot works with perl 5.8.x, 5.10.x, 5.12.x and probably blead (perl 5's master branch)
01:26 dukeleto rfw: yes
01:26 rfw ah, got it
01:26 dukeleto rfw: we can start out with one version of perl, but the solution needs to scale to multiple versions
01:27 rfw sure
01:27 dukeleto rfw: on the gcc farm, we have access to every version of gcc every known to man, pre-compiled
01:27 rfw awesome
01:27 dukeleto rfw: so we can go hog-wild with those as well, but probably just testing a few gcc versions is sufficient
01:27 rfw i'll probably need to brainstorm an architecture with Yuki`N though
01:27 dukeleto rfw: i need you and Yuki`N to come up with the architecture
01:27 Yuki`N lol
01:28 Yuki`N BRAIN BLAST.
01:28 rfw lol that was fun timing
01:28 Yuki`N I'm looking at MappedByteArray in the meantime.
01:28 NotFound For extra points, you can try to build with any supported perl version built with any gcc version supported by perl ;)
01:28 Yuki`N Testing is actually not going to be as difficult as I thought.
01:28 dukeleto rfw: yes, do that. brainstorm with Yuki`N and get back to me. Do some research and see what will give use something useful soon and be scalable in the future to do what we need
01:28 dukeleto Yuki`N: awesome!
01:28 rfw Yuki`N: we will call it squawk, y/y
01:28 Yuki`N yes.
01:29 Yuki`N parrot squawk
01:29 dukeleto rfw: there is already a langague called "squaak", the example parrot language
01:29 NotFound Yuki`N: I wrote MappedByteArray, feel free to ask any doubt.
01:29 Yuki`N I will.
01:29 Yuki`N I think I don't have much time today however.
01:29 rfw dukeleto: is there anything called plumage
01:29 Yuki`N Yes.
01:29 dukeleto rfw: yes
01:29 rfw oh wow
01:29 Yuki`N It's some package management or optional feature thingy.
01:30 dukeleto rfw: it is kind of like the CPAN toolchain but for Parrot
01:30 Yuki`N From what I can gather.
01:30 rfw running out of parrot/bid-related words again
01:30 Yuki`N Polly's a parrot-related name.
01:30 NotFound Yuki`N: if I'm not alive in the channel, email to the parrot list
01:30 dukeleto rfw: what about teradactyl ?
01:30 Yuki`N But it's kinda :|
01:30 rfw dukeleto: that's not exactly a parrot, lol
01:30 Yuki`N NotFound, will do.
01:30 dukeleto rfw: is is the grandpa of a parrot ;)
01:30 dukeleto s/is is/it is/
01:30 kid51 aloha, plumage?
01:30 aloha kid51: plumage is a package manager for Parrot
01:31 dukeleto rfw, Yuki`N, Kapace_ : you are all doing a fantastic job in GCI. Keep it up!
01:31 rfw why is every parrot release name by whiteknight in lolspeak
01:31 * dukeleto goes back into the shadows for a few hours
01:31 NotFound aloha: plumage is also https://github.com/parrot/plumage
01:31 aloha NotFound: Okay.
01:32 NotFound aloha: plumage?
01:32 aloha NotFound: plumage is a package manager for Parrot or https://github.com/parrot/plumage
01:32 Yuki`N Because it's whiteknight.
01:32 rfw oh, of course
01:32 Yuki`N aloha's pretty cool, who wrote it?
01:33 rfw is there anything called kea
01:33 rfw i know there's a kakapo
01:33 rfw Kea: Factor on the Parrot Virtual Machine
01:33 rfw why do you do this to me ;__;
01:34 Yuki`N How about Macaw.
01:34 cotto_work Yuki`N: aloha is bacek's project, though I think it's pretty close to stock
01:35 Yuki`N Ah.
01:35 dalek parrot: fc913cc | dukeleto++ | tools/dev/findhacks.pl:
01:35 dalek parrot: [tools] Script which generates a trac wiki page of hacky files, Yuki`N++
01:35 dalek parrot: review: https://github.com/parrot/parrot/commit/fc913cc752
01:35 rfw Yuki`N:bpa / Macaw Parrot browser plugin
01:35 rfw :(
01:35 Yuki`N damnit
01:35 rfw what about cockatoo
01:35 dalek TT #1910 created by whiteknight++: Review and Dissolve old embedding API
01:35 dalek TT #1910: http://trac.parrot.org/parrot/ticket/1910
01:36 kid51 OSU OSL will be doing some maintenance work around 2100 PST today.  Anything Trac-related may be affected.
01:36 Yuki`N Toucan?
01:36 rfw ooh, that's free
01:36 rfw toucan it is
01:37 Kapace_ thanks  dukeleto  :)
01:38 kid51 Do we have anyone who can do a presentation about Parrot in Italian?
01:40 davidfetter how much prep might they get?
01:40 nwellnhof left #parrot
01:41 rfw actually Yuki`N
01:41 rfw what about darter
01:41 rfw The darters or snakebirds are mainly tropical waterbirds
01:41 rfw hurr >snakebird
01:42 * davidfetter wonders what hurr is overwriting snakebird with, and realizes he needs to be outdoors more
01:43 rfw davidfetter: hah
01:43 cotto_work rfw: there's no reason you have to use a real parrot.
01:44 rfw cotto_work: as opposed to a fake parrot?
01:44 davidfetter are we running out of real parrots to name the releases for? :)
01:44 cotto_work imaginary parrot
01:44 Kapace_ virtual parrot
01:45 rfw the square root of a negative parrot?
01:47 Nol888 joined #parrot
01:47 Nol888 So VMWare Workstation just took a dump on my computer.
01:47 rfw Nol888: yeah virtualbox does that to me too
01:47 Nol888 I was playing l4d2 and suddenly all my usb devices disconnect.
01:48 Nol888 My NIC's lights aren't on anymore.
01:48 rfw oh
01:48 Nol888 I have several segfault error messages.
01:48 cotto_work rfw: I'm thinking more like "European Attack Macaw"
01:48 rfw mine just dumps memory into the host
01:48 Nol888 And I can't do a thing because my keyboard/mouse are usb.
01:48 rfw cotto_work: "Super Awesome Killer Parrot From Outer Space"?
01:48 Nol888 It's horrible.
01:49 Yuki`N left #parrot
01:49 Nol888 is now known as Yuki`N
01:50 NotFound Zombie Ninja Mutant Parrots
01:51 Kapace_ Yuki`N: you were playing games on a vm?
01:51 Yuki`N No, I was running a vm at the time.
01:52 Kapace_ oh
01:52 davidfetter o/` teenage mutant ninja parrots o/`
01:52 Kapace_ I think the names are rapidly moving away from parrots and more into epic movies
01:53 NotFound The amazing parrotman
01:53 kid51 is now known as kid51_at_dinner
01:53 NotFound Epic? Parrot the barbarian?
01:54 rfw could someone in the parrot organization on github create a "darter" repo and give me and Yuki`N access to it
01:55 whiteknight Every time I do a release, I ask my wife for an idea about what to name it. Every time, without fail, she suggests "Dana"
01:55 whiteknight ...so I pick a random name instead
01:55 rfw lolol
01:55 whiteknight For 2.11 I was seriously thinking about "blue crown"
01:56 rfw how did that become at the roflbbq
01:56 whiteknight I don't know. I was thinking about names, then I accidentally went to 4chan for a while, and when I came back everything was rofl something
01:56 rfw why would you go to 4chan while naming a parrot release
01:57 rfw at least you didn't call it PUDDI PUDDI
01:57 whiteknight I'm not the best at time management
01:59 Kapace_ so I'm working on the embed/api.c test https://gist.github.com/758046
02:00 Kapace_ trying to test Parrot_api_set_stdhandles... I can make the interpreter
02:01 Kapace_ and the call works, but I don't know how to get the interpreter to run some pir...
02:01 whiteknight Kapace_: That's a good question
02:03 whiteknight Kapace_: Comment that function out. We don't use it yet and it's not really testable
02:03 whiteknight at least not yet, not without more work
02:04 Kapace_ hmm, ok
02:04 whiteknight We probably need a function to get the standard IO handles. Maybe that will become a new task, so don't throw away your test code yet
02:04 Kapace_ right, just commented it out for no
02:04 whiteknight I may want to modify that set_handles function too, since I don't know if I like it as it is
02:05 Kapace_ *now
02:05 whiteknight Kapace_: okay, that's good.
02:05 whiteknight This is good, testing helps to identify weak areas in the code
02:06 whiteknight Kapace_: I'm heading to bed now. I'll take a look at whatever you have in the morning
02:06 davidfetter left #parrot
02:06 Kapace_ whiteknight: ok, good night
02:06 whiteknight goodnight
02:06 whiteknight left #parrot
02:21 x19290 joined #parrot
02:21 x19290 Hello.
02:22 sorear Hi.
02:22 x19290 I tried to create my trac.parrot.org account and clicked the "Register" link.
02:23 x19290 And got https://trac.parrot.org/parrot/parrot/register Not Found error.
02:23 x19290 What's wrong?
02:23 rfw x19290: i think you mean https://trac.parrot.org/parrot/register
02:23 rfw ;)
02:24 x19290 Oh! it works.
02:26 rfw Yuki`N: did you run away ヽ(´ー`)ノ
02:28 lidden left #parrot
02:30 silug joined #parrot
02:30 x19290 I found "New Ticket." Thanks.
02:31 Yuki`N rfw, yes
02:31 Yuki`N I did.
02:31 rfw Yuki`N: :<
02:31 bacek_at_work aloha, humans
02:31 bacek_at_work aloha, aloha?
02:31 aloha bacek_at_work: I have no idea.
02:32 bacek_at_work aloha, aloha is simple Bot::BasicBot::Pluggable hosted on https://github.com/bacek/aloha
02:32 aloha bacek_at_work: Okay.
02:32 cotto_work aloha, bacek_at_work
02:32 bacek_at_work aloha, cotto_work :)
02:33 rfw Yuki`N: darter would have to be centralized, wouldn't it
02:33 Yuki`N Well, I suppose.
02:34 Yuki`N The setup only exists on one machine.
02:34 rfw well
02:34 rfw what
02:34 rfw it runs on the gcc compile farm, so you would have multiple nodes
02:38 Yuki`N Oh true.
02:38 Yuki`N Some client-server interaction going on.
02:38 rfw but then wyou would have a single point of failure
02:39 Yuki`N That is true.
02:39 rfw alternatively you could send it to any given node and let it propogate
02:39 Yuki`N That is true.
02:39 Yuki`N P2P for the win I suppose.
02:39 rfw but that could get messy
02:39 Yuki`N Decentralized is always cooler though.
02:39 rfw since that decentralized node would act is a temporary hub
02:40 rfw s/is/as/
02:40 rfw we should probably make a #darter :|
02:40 Yuki`N pfft this is parrot-related.
02:41 rfw heh
02:41 rfw hm having a distributed architecture is messy though
02:41 Yuki`N Were we thinking about integrating with git?
02:41 rfw i think so
02:41 rfw but that's not too complicated
02:41 rfw just have every node periodically pull
02:43 rfw but yeah Yuki`N a centralized server would be easier to aggregate data with
02:43 Yuki`N It would be.
02:43 Yuki`N Each server has a different set of tools set up on it, I take it?
02:44 rfw i'm not sure
02:44 rfw i'd like to try that playstation one out :D
02:44 dalek TT #1911 created by x19290++: cannot read binary file stream
02:44 dalek TT #1911: http://trac.parrot.org/parrot/ticket/1911
02:45 rfw oh Yuki`N
02:45 rfw GCC release sources and misc tarballs are under /opt/cfarm/ftp/ (writable for everyone if you wish to add something)
02:45 rfw GCC release binaries for the default languages plus Ada are available under /opt/cfarm/release/X.Y.Z/bin (to be put in your PATH 4.3.3 is available on most CFARM machines)
02:45 rfw ada, what
02:45 Yuki`N loool.
02:46 sorear rfw: you've never heard of Ada?
02:46 rfw sorear: i have but why
02:46 rfw why do they have ada :|
02:47 * cotto_work takes off
02:48 sorear rfw: why not?  it's a major language and gcc supports it
02:48 rfw is it? i've never seen it used much
02:48 sorear it's not used much *today*
02:49 sorear Ada was the Pentagon's sweetheart for some years a while back and a lot of existing code uses it
02:50 rfw ah
02:51 sorear http://en.wikipedia.org/wiki/Ada​_(programming_language)#History
02:54 rfw oh
02:57 x19290 left #parrot
03:05 rfw Yuki`N: would it be appropriate for nodes to store the tests or for the hub to dispatch them
03:12 contingencyplan left #parrot
03:12 Yuki`N What do you mean?
03:13 rfw as in
03:13 rfw should the tests be marshalled and transferred to the leaf nodes
03:13 rfw or should each leaf node have that test stored
03:15 Yuki`N rfw, well the tests might be stored on a central server and maybe the leaf nodes can retrieve.
03:15 Yuki`N If tests get outdated, etc.
03:15 rfw hm
03:15 rfw or should they be marshalled every time?
03:15 Coke 22:13 < [Coke]> email does work.
03:15 Coke 22:15 < [Coke]> You can open a new parrot ticket by emailing tickets at  parrot.org
03:16 Coke kid51_at_dinner: pong.
03:17 dalek TT #1912 created by coke++: Test 2
03:17 dalek TT #1912: http://trac.parrot.org/parrot/ticket/1912
03:19 kid51_at_dinner is now known as kid51
03:20 kid51 coke: I subsequently posted to http://trac.parrot.org/parrot/ticket/1908
03:20 fbrito1 is now known as fbrito
03:21 Coke I just reposted. it's in the docs for perlcritic.t
03:24 kid51 Aha!  But I wonder if there is a way to get that into 'make help'?  'cause that's where I would have thought to look for it
03:24 Coke I don't see the need, but knock yourself out.
03:24 Coke (I'd be happy to only ever run that test during fulltest, it's a pig)
03:25 kid51 Well, at least I can incorporate that syntax into 'make moretests' (or whatever we'll end up calling it)
03:25 kid51 Alright, let me start a branch for that and then head out
03:28 Coke Yuki`N: as long as you have privs, you should re-run faces.pl and upload it.
03:28 kid51 Hmm ... looks like bot nopaste has gone missing
03:32 kid51 http://pastebin.com/RX35MU9h
03:34 dalek TT #1912 closed by coke++: Test 2
03:34 dalek TT #1912: http://trac.parrot.org/parrot/ticket/1912
03:34 dalek TT #1909 closed by coke++: g++ build broken
03:34 dalek TT #1909: http://trac.parrot.org/parrot/ticket/1909
03:41 Matt_ left #parrot
03:43 kid51 Yuki`N: ping
03:43 Yuki`N Coke, faces?
03:44 Yuki`N kid51, pong
03:44 kid51 Yuki`N: See http://trac.parrot.org/parrot/ticket/552
03:44 Yuki`N yes, that's the one I wrote a script for.
03:44 kid51 Can you post in that ticket the extent to which tools/dev/findhacks.pl addresses the issues raised there?
03:44 Yuki`N Ok.
03:46 kid51 dukeleto: By accident, I assigned you this ticket:  http://trac.parrot.org/parrot/ticket/552#comment:5
03:47 kid51 But if you want to assign it back, feel free.
03:49 kennym left #parrot
03:50 Coke Yuki`N: tools/*/*faces*.pl
03:50 Coke or ParrotFaces on the wiki.
03:50 * kid51 must sleep
03:51 Yuki`N Coke, ok.
03:51 fbrito kid51: good night
03:53 dalek parrot: 067bf3a | Coke++ | / (14 files):
03:53 dalek parrot: remove init::headers config step
03:53 dalek parrot:
03:53 dalek parrot: hardcode include directory here instead of config step.
03:53 dalek parrot: Removing NONGEN headers entirely still passes checkdepend.t
03:53 dalek parrot: and the build works at a high -j.
03:53 dalek parrot: review: https://github.com/parrot/parrot/commit/067bf3a22f
03:53 dalek parrot: a32e328 | Coke++ | t/src/checkdepend.t:
03:53 dalek parrot: Enable testing of src/pmc dependencies.
03:53 dalek parrot: review: https://github.com/parrot/parrot/commit/a32e328c4a
03:53 dalek parrot: d3ca078 | Coke++ | config/ (2 files):
03:53 dalek parrot: Avoid duplicate deps on PMCs; add minimum deps
03:53 dalek parrot: review: https://github.com/parrot/parrot/commit/d3ca078538
03:53 dalek parrot: b2b52b6 | Coke++ | config/auto/pmc.pm:
03:53 dalek parrot: Minor fixups on calculated dependency paths.
03:53 dalek parrot: - don't bother scanning <> includes at all.
03:53 dalek parrot: review: https://github.com/parrot/parrot/commit/b2b52b669b
03:53 dalek parrot: 6d6911d | Coke++ | config/auto/pmc.pm:
03:53 dalek parrot: every PMC has this as a dep.
03:53 dalek parrot: (safe now that we de-dup)
03:53 dalek parrot: review: https://github.com/parrot/parrot/commit/6d6911d0cc
03:53 dalek parrot: d9b858c | Coke++ | config/ (2 files):
03:53 dalek parrot: move PMC_H_FILES into the pmc dep generator.
03:53 dalek parrot: lets us de-dup more elements
03:53 dalek parrot: add some more dynamic reqs.
03:53 dalek parrot:
03:53 dalek parrot: * down to 4 failures in checkdepend.pl
03:53 dalek parrot: review: https://github.com/parrot/parrot/commit/d9b858c48c
03:53 dalek parrot: a71ab59 | Coke++ | config/auto/pmc.pm:
03:53 dalek parrot: Track cases where our include includes.
03:53 dalek parrot: (only 1, so cheat.)
03:53 dalek parrot: review: https://github.com/parrot/parrot/commit/a71ab59eb0
03:53 dalek parrot: dd03d3c | Coke++ | config/auto/pmc.pm:
03:53 dalek parrot: Fix very old typo
03:53 dalek parrot: Since we were never checking dependencies, we never noticed that this code
03:53 dalek parrot: and the pmc2c code were doing 2 different things.
03:53 dalek parrot:
03:53 dalek parrot: - checkdepend.t passes again (now with no .c skips)
03:53 dalek parrot: review: https://github.com/parrot/parrot/commit/dd03d3caec
03:53 dalek parrot: c8161fc | Coke++ | config/auto/pmc.pm:
03:53 dalek parrot: whitespace
03:53 dalek parrot: review: https://github.com/parrot/parrot/commit/c8161fcddb
03:53 dalek parrot: f668f73 | Coke++ | t/codingstd/pccmethod_deps.t:
03:53 dalek parrot: Fix same bug that was cut/pasted from auto::pmc
03:53 dalek parrot:
03:53 dalek parrot: Add note about merging this dependency check into checkdepend.t
03:53 dalek parrot: review: https://github.com/parrot/parrot/commit/f668f73882
03:53 dalek parrot: 5a71f25 | Coke++ | config/ (2 files):
03:54 dalek parrot: Remove now-redundant .mak variable.
03:54 dalek parrot: review: https://github.com/parrot/parrot/commit/5a71f25d05
03:54 dalek parrot: e12aa0f | Coke++ | frontend/parrot_debugger/main.c:
03:54 fbrito wow
03:54 dalek parrot: remove trailing whitespace.
03:54 dalek parrot: review: https://github.com/parrot/parrot/commit/e12aa0f40e
03:54 dalek parrot: 0cc0d8e | Coke++ | src/embed.c:
03:54 dalek parrot: add missing ASSERT_ARGS
03:54 dalek parrot: review: https://github.com/parrot/parrot/commit/0cc0d8e831
03:54 dalek parrot: 98cc69b | Coke++ | config/gen/makefiles/root.in:
03:54 dalek parrot: more includes? more dependencies.
03:54 dalek parrot: review: https://github.com/parrot/parrot/commit/98cc69b18b
04:04 dalek parrot: 40569a6 | Coke++ | t/steps/auto/pmc-01.t:
04:04 dalek parrot: "fix" steps test - METHOD, not PCCMETHOD
04:04 dalek parrot: review: https://github.com/parrot/parrot/commit/40569a69d6
04:05 fbrito left #parrot
04:08 kid51 left #parrot
04:09 Yuki`N Coke, no change.
04:24 dalek TT #1911 closed by plobsing++: cannot read binary file stream
04:24 dalek TT #1911: http://trac.parrot.org/parrot/ticket/1911
04:24 frodwith left #parrot
04:30 dalek parrot: 84fb5d3 | mikehh++ | MANIFEST:
04:30 dalek parrot: re-generate MANIFEST
04:30 dalek parrot: review: https://github.com/parrot/parrot/commit/84fb5d375e
04:30 dalek parrot: 58d6f87 | mikehh++ | tools/dev/findhacks.pl:
04:30 dalek parrot: fix codetest failures
04:30 dalek parrot:
04:30 dalek parrot: remove cuddled elses
04:30 dalek parrot: change platform-specific perl shebang line
04:30 dalek parrot: add perl coda
04:30 dalek parrot: review: https://github.com/parrot/parrot/commit/58d6f873a8
04:31 Coke Yuki`N: figured for sure some GCI students had been added to CREDITS, ah well.
04:39 Yuki`N Coke, I added myself, but for some reason I didn't see my gravatar.
04:39 Yuki`N You try to run it.
04:40 Coke you're on line 104.
04:41 Coke Running in master?
04:50 dalek TT #1768 closed by coke++: Profiling runcore breaks Rakudo.
04:50 dalek TT #1768: http://trac.parrot.org/parrot/ticket/1768
04:54 plobsing msg dukeleto, sorry I missed you. I'd like to assuage your concerns regarding Complex.pmc. Please comment on the TT or hit the list.
04:54 aloha OK. I'll deliver the message.
05:13 Yuki`N Coke, pretty sure.
05:23 bacek aloha, humans
05:23 bacek seen cotto
05:23 aloha cotto was last seen in #parrot 21 hours 37 mins ago saying "are they all ok to merge without reviewing?".
05:25 Yuki`N Coke, never mind, I didn't paste properly apparently.
05:25 Yuki`N Updated.
05:27 cotto hi
05:27 cotto bacek, did you need something?
05:28 bacek cotto, http://trac.parrot.org/parrot/wiki/Parro​tDeprecations?action=diff&amp;version=28 is it good enough?
05:29 cotto looking
05:31 cotto I'd prefer to see a full notice.  I can write the English part if you prefer.
05:31 cotto thanks
05:32 cotto Coke, thanks for catching #1768.  It probably was a dupe of the "-Rprofiling breaks nqp-rx" ticket.
05:33 dukeleto ~~
05:33 * dukeleto is adding gci tasks
05:34 bacek cotto, basically it was part of immutable strings. Which was landed in 2.3 afair
05:34 dalek tracwiki: v6 | Yuki`N++ | ParrotFaces
05:34 dalek tracwiki: rerun tools/dev/faces.pl
05:34 dalek tracwiki: http://trac.parrot.org/parrot/wiki/P​arrotFaces?version=6&amp;action=diff
05:34 dalek tracwiki: v28 | bacek++ | ParrotDeprecations
05:34 dalek tracwiki: http://trac.parrot.org/parrot/wiki/Parro​tDeprecations?version=28&amp;action=diff
05:35 dukeleto Yuki`N: i just added some mappedbytearray tasks
05:35 rurban_ joined #parrot
05:36 * dukeleto muses that parrot test coverage, in general, has skyrocketed since GCI started
05:37 rurban left #parrot
05:38 rurban_ is now known as rurban
05:38 cotto dukeleto++
05:51 Yuki`N dukeleto, ah.
05:51 Yuki`N I'm going to sleep now though
05:52 Coke left #parrot
06:26 Kapace_ Yuki`N: mind if I take one? :P
06:30 Kapace_ If I need the MappedByteArray test to read a testfile, where should I put the file?
06:32 dukeleto Kapace_: same directory as the test
06:33 Kapace_ ok http://www.google-melange.com/gci/task/show/goog​le/gci2010/parrot_perl_foundations/t129360088964
06:35 Yuki`N Kapace, Ok you take one, I'll complete your work.
06:36 Kapace_ Yuki`N: awesome :)
06:38 Yuki`N left #parrot
06:41 dukeleto melange is dragging along right now
06:42 Kapace_ ok, I have a pull request already open
06:42 dukeleto Kapace_: accepted
06:42 Kapace_ should I submit the same link?
06:43 Kapace_ or open new pull request?
06:45 dukeleto Kapace_: what is your question?
06:45 Kapace_ I have a pull request already open: https://github.com/parrot/parrot/pull/48
06:46 Kapace_ Should I create a new one for this task, or just reuse this one?
06:46 cotto Kapace_, I was just looking at that one trying to figure out if it got merged
06:52 dukeleto Kapace_: not sure, whichever makes the most sense. One pull request per task makes things simple
06:53 dukeleto cotto: we need to work on our pull requests, having too many pending will increase the chance of merge conflicts
06:53 dukeleto cotto: do you want to split them up and get them merged?
06:53 Kapace_ ok, hmm, maybe for this task I'll just keep it in this pull request, but next time I'll try to separate them
06:54 cotto I like one request per task.
06:56 cotto I'm not a fan of github's pull request notificatoins
06:56 dukeleto cotto: what do you mean?
06:56 dukeleto cotto: how it sends it out to everyone in the org?
06:57 cotto The notifications don't reflect whether the requests have been closed
06:57 dukeleto cotto: i ignore all notifications and just look at the pull requests for the parrot org
06:57 cotto that's probably a good way to go about it
07:04 dukeleto cotto: should the Object PMC be using Parrot_ext_call internally?
07:04 dukeleto cotto: that seems odd to me
07:07 cotto dukeleto, good question
07:07 cotto It strikes me as odd but plausible.
07:08 dukeleto cotto: eventhandler, object and class are the only PMCs that use Parrot_ext_call
07:08 cotto Object and Class are special, so they don't surprise me too much.
07:15 rfw bah, there's some work on the darter client/server architecture done
07:15 rfw seems really quite crude though
07:15 Kapace where?
07:15 rfw #!/usr/bin/env python26
07:15 rfw er
07:15 rfw completely wrong paste
07:15 rfw https://github.com/rofflwaffls/darter
07:16 * dukeleto looks
07:17 dukeleto rfw: i like your docs so far, good description of the architecture
07:17 rfw i need to add more pydoc to the actual classes and functions, though
07:18 rfw i was prototyping this really quickly
07:19 dukeleto cotto: why do we have a OrderedHash PMC? Almost nothing seems to use it
07:20 cotto dukeleto, istr that it's needed for *ManagedStruct
07:22 dukeleto cotto: unmanagedstruct is the only thing that uses it
07:22 dukeleto cotto: just seems a bit lonely
07:24 cotto One use is a lot more than zero.
07:27 dukeleto cotto: indeed
07:27 dukeleto cotto: how do you feel about the Complex PMC ?
07:27 dukeleto cotto: do you think it belongs in core?
07:27 cotto I don't see a compelling reason to keep it in core.
07:27 * dukeleto is thinking about what else needs to get deprecated before 3.0
07:28 dukeleto cotto: but where should it live?
07:28 dukeleto cotto: i don't think parrot-linear-algebra is the right place
07:28 cotto shot down before I could finish typing my reply
07:28 dukeleto cotto: parrot-data-structures could work, but that would need to get more love
07:29 dukeleto cotto: i don't want to see the complex pmc leave core and then go unmaintained and broken
07:29 dukeleto cotto: if there are bugs in it now, i will fix them
07:29 cotto If we get automated testing via plumage working, it won't be a problem to get external projects regular automated lov^H^H^Htesting.
07:29 dukeleto cotto: but i understand that it is at the boundary of what should be in core and what should be optional
07:29 NotFound A first step may be moving it to dynpmc
07:30 dukeleto NotFound: i was thinking that too
07:30 cotto dukeleto, do you happen to know if/how Rakudo uses it?
07:30 dalek parrot: 45282e1 | NotFound++ | frontend/parrot_debugger/main.c:
07:30 dalek parrot: delete a TODO comment outdated since long time
07:30 dalek parrot: review: https://github.com/parrot/parrot/commit/45282e15de
07:32 dukeleto cotto: rakudo rolls their own
07:34 cotto ok
07:38 dukeleto cotto: Pointer PMC has 23% code coverage. Do people actually use it?
07:38 dukeleto cotto: oh, i see.
07:38 dukeleto cotto: is is a base class which things override
07:40 * dukeleto is on a gci task rampage
07:42 Kapace_ yay, parrot tasks! :)
07:43 dukeleto Kapace_: take a look, i just added a half dozen and will be adding another half dozen in a few minutes
07:45 Kapace_ I'm looking, can't wait until I can claim the next one
07:50 dukeleto Kapace_: if you wait a few more minutes, you will have a bigger selection :)
07:53 Kapace_ right, but first I need my current task closed: http://www.google-melange.com/gci/​task/show/google/gci2010/parrot_pe​rl_foundations/t129360088964#c2002
07:54 cotto I'll review that unless dukeleto's already on it.
07:55 cotto Kapace_, there's a lot of extra stuff in that pull request.  In the future, please keep separate work separate.
07:55 Kapace_ cotto: right, I will try to do that
07:56 cotto thanks
07:57 dukeleto Kapace_: yeah, pull requests should not mix tasks
07:58 dukeleto Kapace_: if you keep the master branch in your fork the same as the master in parrot.git, and make branches from it for each task, that is one way you can do that
07:58 Kapace_ I wasn't sure how to make a new pull request with just the last commit.
07:58 Kapace_ ah, perfect Ill do that
07:59 dalek parrot: 2b4c364 | cotto++ | tools/dev/findhacks.pl:
07:59 dalek parrot: color "hack" red in findhacks output
07:59 dalek parrot: review: https://github.com/parrot/parrot/commit/2b4c364219
08:01 dukeleto NotFound: what was your question about "git targets and old git versions" in #ps ? i don't quite get it
08:04 NotFound dukeleto: It was in machine at payjob and I forgot to take note, I'll file a ticket when I can.
08:05 dukeleto NotFound: you had issues with using git 1.5.x or something like that?
08:06 NotFound An outdated version of git failed to run a command while installing a git project, that's all I remember.
08:07 NotFound I think it was an ubuntu 8.04
08:07 dukeleto Kapace_: ok, lots more tasks for you to choose from
08:08 Kapace_ dukeleto: thanks :)
08:08 Kapace_ Capture pmc sounds interesting
08:09 fperrad joined #parrot
08:13 bacek left #parrot
08:14 theory left #parrot
08:17 cotto dukeleto, what's the git equivalent of svnclobber?
08:29 rfw cotto: rm -rf *; git checkout -f HEAD?
08:29 rfw make sure not to accidentally delete .git, heh
08:30 sorear rfw: I think you mean git reset --hard HEAD; git clean -dfx
08:31 rfw sorear: oh, is that the proper way
08:31 dalek parrot: 7dc6d62 | kapace++ | t/pmc/ (2 files):
08:31 dalek parrot: Added file read test to MappedByteArray, increases coverage to 48%
08:31 dalek parrot:
08:31 dalek parrot: Signed-off-by: Christoph Otto <christoph_git@mksig.org>
08:31 dalek parrot: review: https://github.com/parrot/parrot/commit/7dc6d62934
08:31 rfw yeah i knew i was probably telling a lie :|
08:32 cotto Kapace_, I closed your pull request.  If you want anything else to get into master, make a separate request for it.
08:32 Kapace_ :) ok, well thats it for today, good night everyone, and thanks for your help!
08:32 cotto thanks for the improved test coverage
08:33 Kapace_ ok, np, night, kthxbai
08:33 cotto 'night
08:43 dukeleto rfw: what is svnclobber supposed to do again?
08:43 rfw dukeleto: apparently delete all files that aren't committed to svn
08:45 dukeleto rfw: git clean -fdx
08:45 dalek parrot: a72b588 | (Tony Young)++ | t/pmc/continuation.t:
08:45 dalek parrot: Increased coverage of Continuation PMC to 86.60%.
08:45 dalek parrot:
08:45 dalek parrot: Signed-off-by: Christoph Otto <christoph_git@mksig.org>
08:45 dalek parrot: review: https://github.com/parrot/parrot/commit/a72b588db7
08:45 rfw yeah i should really do that from now on
08:45 dukeleto rfw: make realclean
08:45 rfw and i should really change my name on github to rfw
08:46 dukeleto rfw: i like rofflwaffls :)
08:46 rfw dukeleto: yes but right now it's my full name :|
08:47 cotto rfw, how would you like to be known in CREDITS?  rfw or rofflwaffls?
08:47 rfw cotto: just rfw
08:48 cotto wfm
08:48 * dukeleto is making a GCI task for the Task PMC. muhahahahaha!
08:48 * cotto is making a task for going to sleep
08:49 tadzik . o O ( insrease the GCI coverage of the Task PMC)
08:49 dalek parrot: f6f2af2 | cotto++ | CREDITS:
08:49 dalek parrot: add rfw++ to CREDITS
08:49 dalek parrot: review: https://github.com/parrot/parrot/commit/f6f2af2bcb
08:49 tadzik that'd be something: GCI::Cover. To check if really the entire project is covered with GCI tasks
08:49 rfw \o/
08:50 cotto rfw, are you in New Zealand?
08:50 rfw cotto: yes
08:51 rfw cotto: why do you ask?
08:51 cotto Cool.  That part of the world is underrepresented among Parrot developers.
08:51 rfw ah
08:51 rfw do you have anyone else from there :D
08:51 rfw s/there/here/
08:51 cotto bacek is in sydney
08:51 cotto or somewhere nearby
08:51 rfw ah
08:52 rfw do i get the title "first new zealand parrot developer"
08:52 cotto Probably not, but I don't know of any others.
08:52 rfw aw
08:53 cotto In a ten year old project it's hard to be the first anything.
08:53 rfw heh
08:53 rfw you know kakapo is a native new zealand parrot though
08:53 cotto yup
08:53 cotto charming bird, if a bit dim
08:53 rfw not nearly as bad as the kea
08:57 dukeleto rfw: i have a fledling project named Kea already :)
08:57 dukeleto rfw: Factor on Parrot, mostly an idea at this point : http://github.com/leto/kea
08:57 rfw what's a factor
08:58 dukeleto rfw: it is a stack-based programming language
08:58 rfw aren't most of them
08:59 dukeleto rfw: factorcode.org
08:59 dukeleto rfw: many are stack based. This is stack-based in the sense that all functions have "stack effects" instead of arguments to the function
09:00 rfw oh, i see what you mean
09:00 dukeleto rfw: stack effects describe the necessary structure of the stack before and after the function has executed, such as "there are at least 2 elements before and 1 after" and such
09:00 rfw haha, reverse polish notation
09:00 dukeleto rfw: but they can be dynamic
09:00 dukeleto rfw: yes, it reads from right to left
09:01 rfw ah this looks pretty cool
09:01 dukeleto rfw: but algorithms are extremely short in Factor. It is very expressive, not much code is needed
09:01 rfw i've actually thought about something like this
09:01 * cotto sleeps
09:01 rfw night cotto
09:01 cotto 'night
09:01 rfw dukeleto: i didn't realize such a thing actually existed :D
09:01 dukeleto rfw: it is also like smalltalk in some sense that it lets you save images of environments and start them up again
09:02 rfw oh, not too familiarw ith smalltalk
09:02 dukeleto rfw: i know the guy that started Factor and a few other Factor hackers, and they are all good people that do awesome stuff
09:03 rfw ah
09:03 dukeleto rfw: nor am i, but i am learning a bunch because i am reading the Smalltalk-80 book
09:03 rfw hehe, enjoy
09:04 dukeleto rfw: Factor has a small (13K line C++ vm) and *everything* else is written in Factor
09:04 rfw dukeleto: what does the vm even have?
09:05 dukeleto rfw: only the most low level things that are architecture-specific
09:05 rfw ah, cool
09:05 dukeleto rfw: just the smallest bootstrapping that is needed
09:05 dukeleto rfw: it used to has asm in the vm but that was taken out for portability reasons
09:05 rfw heh
09:06 dukeleto rfw: do you get to see Kea birds at all?
09:06 rfw yeah they're annoying as hell
09:06 rfw they tip over rubbish bins and stuff
09:07 dukeleto rfw: flying raccoons, it sounds like
09:07 rfw haha yeah
09:07 dukeleto rfw: which task are you working on now?
09:07 rfw dukeleto: just that ci thing
09:08 dukeleto rfw: i added a slew of test coverage tasks, take a look when you need a break :)
09:08 rfw sure :D
09:08 dukeleto rfw: darter sounds cool
09:09 rfw i hope it will be cool
09:09 dukeleto rfw: what datastructures does it use?
09:09 dukeleto rfw: what goes in and what goes out?
09:09 rfw um basically
09:09 rfw the hub sends marshalled functions to the leaf nodes, which run them
09:09 rfw and they send some kind of result back (not sure what yet)
09:11 rfw the good thing about sending marshalled fucntions down the wire is that you don't need to individually update tests on each machine
09:11 rfw the bad thing is that really weird crap happens if you're not careful
09:11 dukeleto rfw: what do you mean by a "marshalled function" ?
09:11 rfw well, it sends the bytecode of a test function to the leaf machines
09:12 rfw i guess its a less crude way of sending a source file to another machine and evaluating it
09:13 rfw it's*
09:13 dukeleto rfw: all machines NFS mount the same filesystem, so none of that is needed
09:13 rfw oh bleh
09:13 dukeleto rfw: assume every machine can see any file it needs
09:13 dukeleto rfw: and at first, the master node will also be running the tests, for simplicity
09:14 rfw ah
09:14 rfw well, the hub would need to run a leaf node on it as well and have it connect to itself
09:14 dukeleto rfw: yep
09:15 dukeleto rfw: i am thinking in terms of a cron job at first, just something that updates a git repo, copies it to a bunch of temp-repo dirs, and then kicks off a bunch of build+tests
09:15 dukeleto rfw: so i think you can just write the leaf code at first
09:15 dukeleto rfw: don't worry about the hub
09:15 dukeleto rfw: that can be added once we have leaf code working
09:16 rfw dukeleto: i was actually thinking of using a git hook that makes an rpc call to the hub
09:16 dukeleto rfw: why RPC ?
09:16 dukeleto rfw: seems overly complicated
09:16 rfw well, not strictly rpc
09:16 rfw just some kind of procedure call
09:16 dalek parrot: 1f25a8e | fperrad++ | t/dynpmc/foo.t:
09:16 dalek parrot: [t] fix test 4, concat_s_s is gone
09:16 dalek parrot: review: https://github.com/parrot/parrot/commit/1f25a8e6b9
09:16 dukeleto rfw: sure, i was misunderstanding
09:17 rfw i wasn't quite clear :D
09:18 dukeleto rfw: i am thinking that a git hook adds an entries to some kind of datastore, then a cron job comes along (or a hub daemon) and checks for new elements in the datastore and then sends them to leaf nodes
09:18 dukeleto rfw: jitterbug solves most of that
09:18 dukeleto rfw: but i have never actually used jitterbug
09:18 rfw i'll need to look into jitterbug first
09:18 dukeleto rfw: so you can decide if using jitterbug is worth it
09:19 dukeleto rfw: jitterbug solves the "test this against N perl's" problem, which is quite nice. You don't want to deal with that.
09:19 rfw heh
09:19 dukeleto rfw: but we can just think about 1 perl for now
09:19 dukeleto rfw: we can use perlbrew to add other perls later
09:20 dukeleto rfw: jitterbug uses perlbrew to choose between different perls
09:20 rfw i better add that to the list of things i need to look up later
09:20 dukeleto rfw: indeed!
09:20 dukeleto rfw: also, I don't want to use Hudson, so don't waste your time on that :)
09:20 contingencyplan joined #parrot
09:22 rfw dukeleto: i should probably talk some of this over with nol888
09:23 kennym joined #parrot
09:24 dukeleto rfw: i think you need to step back and think about data structures
09:24 rfw dukeleto: i do have some conception of how they would work
09:24 dukeleto rfw: what goes in? what goes out? How do you represent that data?
09:25 rfw dukeleto: i was just htinking to serialize the results in json, like i did with x264's regression testing framework
09:25 rfw which then can be parsed by whatever
09:25 dukeleto rfw: your input is git repos and git sha1, with various configurations and your output is some kind of output of "make" and "make test" and possibly "make cover" and who knows what else
09:26 dukeleto rfw: yes, of course. JSON or YAML, whichever makes more sense
09:26 dukeleto rfw: they are interchangeable
09:26 rfw yes, i know :D
09:26 rfw well, only json->yaml anyway
09:26 dukeleto rfw: if we want to have interactive javascript graphs, then JSON is more useful
09:27 dukeleto rfw: there is something called JSYNC that will do json <--> yaml, except for some very rare edge cases that shouldn't be used, as far as I understand
09:27 rfw hm i'm more of a json fan seeing it's more supported + it's the native javscript data format
09:28 dukeleto rfw: so i would go with JSON, because we want the optional of interactive visualization of the data
09:28 dukeleto rfw: you should take a look at tool_bench
09:29 dukeleto rfw: http://github.com/notbenh/tool_bench
09:29 rfw ahhhh so many links to click
09:29 dukeleto rfw: it is a small utility to run benchmarks. it can give you JSON output of running a test N times and stuff like that
09:29 rfw dukeleto: https://github.com/rofflwaffls/digress <-- this is the regression testing framework i wrote for x264
09:30 dukeleto rfw: i help hack on it too :)
09:30 rfw it's strictly cli though
09:30 dukeleto rfw: you could use tool_bench at the end of the line, to actually generate the JSON data of timing a specific test file
09:31 dukeleto rfw: but we need to focus. Are we going to focus on building, testing or benchmarking first ?
09:31 rfw building + testing first
09:31 dukeleto rfw: we will always need to build
09:31 dukeleto rfw: so that is a given. Yes, i agree. building+testing
09:31 rfw since i more or less have that written :D
09:32 dukeleto rfw: can we just use digress and add stuff to it and tweak it?
09:32 rfw should be able to
09:32 dukeleto rfw: what does digress actually do? i don't see any docs ...
09:32 rfw um
09:32 rfw http://git.videolan.org/?p=x264.git;a=​blob;f=tools/test_x264.py;h=57022ae74e​329f0353bf4f10d1ade1d2191e7f3d;hb=HEAD <-- this is the script for x264
09:33 rfw it just generates a set of test statsitics for every revision
09:33 rfw it doesn't parallelize or anything though, so i'd probably need to work with that
09:34 rfw s/with/on/
09:34 dukeleto rfw: we can work with that for now
09:35 rfw alright but first i'm going to sleep
09:35 dukeleto rfw: we don't need to do stuff in parallel yet. We don't do anything now :)
09:35 rfw night dukeleto
09:35 dukeleto rfw: have fun!
09:35 dukeleto rfw: dream of parrots...
09:35 rfw i just hope they're not keas :|
09:41 rfw left #parrot
10:00 dukeleto we now have about 25ish new parrot test coverage GCI tasks
10:00 dukeleto go buck wild, kids.
10:12 dalek parrot: 6009565 | mikehh++ | t/src/embed/pmc.t:
10:12 dalek parrot: add const to two tests to get tests to build with g++
10:12 dalek parrot: review: https://github.com/parrot/parrot/commit/6009565726
10:27 dalek parrot: fd68bb9 | mikehh++ | MANIFEST:
10:27 dalek parrot: re-generate MANIFEST
10:27 dalek parrot: review: https://github.com/parrot/parrot/commit/fd68bb9ff3
10:32 bacek joined #parrot
10:36 dalek parrot: 3e9a2f0 | mikehh++ | t/pmc/mappedbytearray.t:
10:36 dalek parrot: fix codetest failure - trailing whitespace
10:36 dalek parrot: review: https://github.com/parrot/parrot/commit/3e9a2f0259
10:50 Psyche^ joined #parrot
10:50 Patterner left #parrot
10:50 Psyche^ is now known as Patterner
10:56 Patterner left #parrot
10:56 Psyche^ joined #parrot
10:56 Psyche^ is now known as Patterner
11:14 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#1898) fulltest) at 2_11_0-335-g3e9a2f0 - Ubuntu 10.10 i386 (g++-4.5)
11:15 Patterner left #parrot
11:18 mikehh what's the difference when merging into a branch with git merge master and git merge origin/master?
11:21 mikehh I've seen references to both, but can't find docs covering the difference between them
11:38 bluescreen joined #parrot
11:46 Tene mikehh: 'master' is the commit your local master is pointing to, and origin/master is the commit you're aware of the remote repo's master pointing to.  In most normal cases, these will be the same.
12:31 mikehh Tene: I did a git pull --rebase --verbose, the git merge master and it claimed - already up to date, the I did a git merge origin/master and it got a lot of stuff, what did I do wrong?
12:33 kid51 joined #parrot
12:34 Psyche^ joined #parrot
12:34 Psyche^ is now known as Patterner
12:40 mtk joined #parrot
12:58 Coke joined #parrot
13:12 Patterner left #parrot
13:18 Psyche^ joined #parrot
13:23 Psyche^ is now known as Patterner
13:26 dalek Some 'mikehh' person just gave me a 38887 byte commit packet.  They need to be more careful in the future.
13:32 Psyche^ joined #parrot
13:32 Patterner left #parrot
13:34 Psyche^ is now known as Patterner
13:37 rurban_ joined #parrot
13:38 rurban left #parrot
13:39 rurban_ is now known as rurban
13:41 Patterner left #parrot
13:41 Psyche^ joined #parrot
13:42 Psyche^ is now known as Patterner
13:42 kid51 left #parrot
13:45 whiteknight joined #parrot
13:47 contingencyplan left #parrot
13:51 whiteknight good morning, #parrot
13:52 Patterner left #parrot
14:02 mikehh left #parrot
14:05 mtk left #parrot
14:06 mtk joined #parrot
14:31 dalek parrot: cf870e9 | Whiteknight++ | / (3 files):
14:31 dalek parrot: Merge remote branch 'origin/gci_deprecate_io_position'
14:31 dalek parrot: review: https://github.com/parrot/parrot/commit/cf870e99ed
14:49 Psyche^ joined #parrot
14:53 Psyche^ is now known as Patterner
15:01 Patterner left #parrot
15:04 SingAlong joined #parrot
15:04 SingAlong it would be nice to have a timestamp in the docs to show when the doc page changed
15:12 lidden joined #parrot
15:14 NotFound SingAlong: use git log
15:28 SingAlong left #parrot
15:32 Psyche^ joined #parrot
15:32 Psyche^ is now known as Patterner
15:32 Matt_ joined #parrot
15:36 hercynium joined #parrot
15:41 SingAlong joined #parrot
15:41 SingAlong left #parrot
16:22 Coke I could see it being useful on, say, docs.parrot.org to know when it was last updated.
16:23 Coke We could easily have "make html" do that if we're in a git repo, though.
16:23 Coke (rather than try to track it manually)
16:29 mtk left #parrot
16:35 * Coke wishes there was an option for "make" that showed build times.
16:46 Patterner left #parrot
16:46 mtk joined #parrot
16:47 Psyche^ joined #parrot
16:47 Psyche^ is now known as Patterner
16:57 pmichaud joined #parrot
16:58 pmichaud left #parrot
16:58 pmichaud joined #parrot
16:58 Coke pmichaud: hio
16:59 pmichaud hello, Coke
17:12 hercynium left #parrot
17:13 bluescreen left #parrot
17:20 hercynium joined #parrot
17:23 he__ joined #parrot
17:27 he_ left #parrot
17:28 cotto_work ~~
17:54 Patterner left #parrot
17:57 dalek parrot: 41afaba | Coke++ | examples/benchmarks/fib.pir:
17:57 dalek parrot: Update fib.pir to match .pl output.
17:57 dalek parrot: (+ minor updates to style and clarity)
17:57 dalek parrot: review: https://github.com/parrot/parrot/commit/41afabaa55
17:59 Coke there's a nice little benchmark to fix. perl version runs in .7s realtime, pir runs in 3.5 or so.
18:00 Psyche^ joined #parrot
18:00 Psyche^ is now known as Patterner
18:03 dalek parrot: 6496141 | Coke++ | editor/mk_skeleton.pir:
18:03 dalek parrot: Don't generate the svn Id tag in the skeleton
18:03 dalek parrot: review: https://github.com/parrot/parrot/commit/649614108f
18:05 Psyche^ joined #parrot
18:05 Patterner left #parrot
18:06 Psyche^ is now known as Patterner
18:06 cotto_work whiteknight: ping
18:06 whiteknight pong
18:06 cotto_work I made a new branch with some of kapace's work for your review
18:06 dalek parrot/kapace/api-serialize: 04ee8f9 | cotto++ | / (3 files):
18:06 dalek parrot/kapace/api-serialize: add Parrot_api_pmc_serialize from kapace++
18:06 dalek parrot/kapace/api-serialize: review: https://github.com/parrot/parrot/commit/04ee8f9bae
18:06 cotto_work there it is
18:07 Kapace_ morning
18:07 cotto_work hi Kapace_
18:07 Kapace_ hey cotto_work
18:08 cotto_work I love that trick of adding .diff to github urls.  dukeleto++ for making that known.
18:19 dalek parrot: 93f3219 | cotto++ | src/embed.c:
18:19 dalek parrot: typo fix
18:19 dalek parrot: review: https://github.com/parrot/parrot/commit/93f3219636
18:23 dukeleto ~~
18:24 Patterner left #parrot
18:27 dalek parrot: e9680f1 | nwellnhof++ | src/ (10 files):
18:27 dalek parrot: Fix usage of Parrot_str_to_cstring
18:27 dalek parrot: review: https://github.com/parrot/parrot/commit/e9680f1529
18:29 whiteknight cotto_work: I'll look at it in a bit
18:30 dmalcolm joined #parrot
18:44 cotto_work quiet day today
18:46 bluescreen joined #parrot
18:53 dukeleto cotto_work: we are still in the holiday lull
18:53 dukeleto cotto_work: what are you hacking on?
18:53 dukeleto cotto_work: i created many tasks last night, hopefully that keeps the gci students busy for a few days
18:54 cotto_work I've been busy just reviewing incoming commits.  Packfiles will probably be my next target.
18:57 dukeleto cotto_work: i saw some braindeadness in the PackfileSegment PMC last night
18:57 dukeleto cotto_work: the get_pointer VTABLE just does "return NULL"
18:57 dukeleto cotto_work: and there is lots of "XXX TODO" to implement various vtables that just throw exceptions right now
18:58 dukeleto cotto_work: seems suboptimal
18:58 cotto_work "suboptimal" is a polite word
18:58 cotto_work I'm really looking forward to having everything go through the Packfile PMCs.
19:00 dukeleto cotto_work: you know me, always subtle
19:02 NotFound blizkost is broken
19:02 dukeleto NotFound: OH THE HUGE MANATEE
19:03 dukeleto NotFound: in what way?
19:03 NotFound symbol lookup error: /home/julian/runparrot/lib/2.11.0​-devel/dynext/blizkost_group.so: undefined symbol: key_next
19:03 dukeleto NotFound: blizkost is another thing that needs to get regularly tested in our CI setup
19:04 Coke left #parrot
19:05 cotto_work dukeleto: can you start a wiki page for CI-worthy project?
19:11 whiteknight cotto_work: a big packfile redesign/reimplemenation is sorely needed
19:11 whiteknight I'm sure you know that
19:11 whiteknight I just want to make sure it gets plenty of brain-time after 3.0
19:12 cotto_work whiteknight: it will
19:13 theory joined #parrot
19:16 whiteknight cotto_work: with the new embedding API in, I think there are some obvious opportunities to simplify some logic, and plobsing's recent work is *exactly* what I I want to be doing
19:17 whiteknight Some of the pbc-related functions in src/embed.c can go away or be refactored for big wins too, I think
19:17 whiteknight I don't know if you had a chance to really look through that file yet
19:17 rfw joined #parrot
19:20 whiteknight TT #1913 looks like a perfect GCI task
19:22 dalek blizkost: a2833b9 | moritz++ | src/pmc/p5scalar.pmc:
19:22 dalek blizkost: update to newer parrot (incomplete)
19:22 dalek blizkost: review: https://github.com/jnthn/b​lizkost/commit/a2833b93f0
19:22 dalek tracwiki: v11 | NotFound++ | ModuleEcosystem
19:22 dalek tracwiki: New repository
19:22 dalek tracwiki: http://trac.parrot.org/parrot/wiki/Modu​leEcosystem?version=11&amp;action=diff
19:22 dalek tracwiki: v12 | NotFound++ | ModuleEcosystem
19:22 dalek tracwiki: mistake
19:22 dalek tracwiki: http://trac.parrot.org/parrot/wiki/Modu​leEcosystem?version=12&amp;action=diff
19:22 dalek TT #1913 created by ajs++: README issues
19:22 dalek TT #1913: http://trac.parrot.org/parrot/ticket/1913
19:24 cotto_work whiteknight: which file?
19:25 whiteknight cotto_work: src/embed.c
19:27 cotto_work "This file implements the Parrot embedding interface." <- How's that different from the embedding api?
19:27 whiteknight cotto_work: the difference is that statement is old and is a lie. The API is the embedding interface
19:27 whiteknight that file contains the old API, which we need to disolve
19:27 whiteknight dissolve
19:28 cotto_work ah
19:28 whiteknight I basically want to deprecate that entire file ASAP. Some of the functions can continue to live, but will be moved elsewhere and maybe renamed
19:28 whiteknight some of those functions need to dei
19:29 whiteknight die
19:29 dalek parrot: 65e770b | cotto++ | src/embed.c:
19:29 dalek parrot: less lies
19:29 dalek parrot: review: https://github.com/parrot/parrot/commit/65e770b0af
19:30 whiteknight cotto++
19:32 cotto_work at the very least I'll review the ticket you started and look through the functions you list and what you want to do with them.
19:32 Psyche^ joined #parrot
19:32 Psyche^ is now known as Patterner
19:35 whiteknight Many of the functions in that file are expressely for setting and manipulating flags, Some of them are pbc related and belong in src/packfile/*
19:36 whiteknight the flag ones, now that I think about it, basically duplicate effort in src/call/context_accessors.c, so those can probably go away too
19:36 whiteknight I need to update the ticket
19:39 cotto_work whiteknight: should packfiles have a Parrot_pf_* api or should everything just go through the PMCs?
19:40 cotto_work I see plobsing++ already took care of one of the items.
19:43 NotFound moritz: ping
19:44 moritz NotFound: pong
19:44 NotFound moritz: Why 'incomplete'?
19:44 Patterner left #parrot
19:44 NotFound It builds for me
19:45 cotto_work whiteknight: nm.  I was thinking of the external interface.  Parrot_pf_* will be internal.
19:45 moritz NotFound: because my build failed
19:45 NotFound Oh
19:45 moritz NotFound: I later found that it's because my perl was configured wrongly
19:46 NotFound Ah, good, then
19:49 NotFound My blizkost-Gtk2 experiments works, so it seems blizkost is fine.
19:54 NotFound Did we have some license for usage of the Parrot logo?
19:54 cotto_work you mean the one on parrot.org?  istr that being allison's work.
19:55 dalek tracwiki: v3 | cotto++ | PackfileTasklist
19:55 dalek tracwiki: http://trac.parrot.org/parrot/wiki/Pack​fileTasklist?version=3&amp;action=diff
20:09 mikehh joined #parrot
20:16 mtk left #parrot
20:19 perlite_ joined #parrot
20:20 mtk joined #parrot
20:23 perlite left #parrot
20:23 perlite_ is now known as perlite
20:24 Psyche^ joined #parrot
20:26 Psyche^ is now known as Patterner
20:33 mtk left #parrot
20:34 mtk joined #parrot
20:35 moritz rakudo spectest on latest parrot lookos horrible
20:35 Patterner left #parrot
20:36 cotto_work moritz: how so?
20:36 cotto_work slow or explodey?
20:36 moritz cotto_work: more than a screen page of failures
20:36 cotto_work Thanks for testing.  Have you had a chance to bisect?
20:38 cotto_work What's one test that fails?
20:41 cotto_work https://gist.github.com/759046 <- temporary build failure when running make install at high -j
20:46 mtk left #parrot
20:47 cotto_work moritz: I'm building Rakudo now.  Do you have a single test that I can use to repro before I dive into spectest_regression?
20:47 Coke joined #parrot
20:51 mtk joined #parrot
20:51 Kapace_ if capt = new ['Capture'], $N0 = capt["number test"] should call get_number_keyed_str, right?
20:52 Tene sounds reasonable, yes
20:52 Kapace_ but it seems to be calling get_number_keyed(pmc*) instead
20:56 rurban left #parrot
20:57 bacek Kapace_, it's kind of "known issue" with IMCC.
20:57 Kapace_ hmm, whats IMCC?
20:59 bacek parrot's pir compiler
20:59 Kapace_ oh, the parse?
20:59 Kapace_ are there any workarounds?
21:02 whiteknight Kapace_: There might be another op you can use
21:02 whiteknight Try something like $N0 = set capt, "number test"
21:02 Kapace_ whiteknight: Ok, i think I already tried that, but I'll keeping changing it etc
21:04 whiteknight Kapace_: No, Ijust looked at the list. None of the ops call VTABLE_get_number_keyed_str
21:05 Kapace_ so theres no way to test?
21:05 whiteknight that vtable in Capture may be dead code if there are no explicit uses of it and if we can't reach it through PIR
21:05 whiteknight Kapace_: There may be no way to use it at all. Open a ticket to suggest we delete it.
21:06 Kapace_ ok, ill try to remove it, and see what happens too
21:06 whiteknight okay, that's a good idea
21:06 whiteknight just don't commit that, since it may take a deprecation cycle to actually remove it
21:07 Kapace_ ok
21:09 cotto_work It's kinda stupid.  Everything dispatches through get_x_keyed
21:09 cotto_work or at least surprising
21:10 tadzik hello parrots
21:11 cotto_work hello parrots
21:11 cotto_work ;)
21:12 mtk left #parrot
21:14 mtk joined #parrot
21:14 Psyche^ joined #parrot
21:14 Psyche^ is now known as Patterner
21:19 dalek TT #1914 created by DavidCzech++: Dead code in src/pmc/Capture.pmc
21:19 dalek TT #1914: http://trac.parrot.org/parrot/ticket/1914
21:20 dalek parrot: 0264a39 | mikehh++ | t/benchmark/benchmarks.t:
21:20 dalek parrot: update to reflect changes to examples/benchmarks/fib.pir
21:20 dalek parrot: review: https://github.com/parrot/parrot/commit/0264a39a9b
21:20 moritz cotto_work: t/spec/S02-magicals/{args,pid}.t both fail on my machine
21:21 Psyche^ joined #parrot
21:21 Kapace_ whiteknight: theres 11% of coverage in those functions
21:22 Patterner left #parrot
21:22 Psyche^ is now known as Patterner
21:23 NotFound TT #1914 is not exact, the fact that the test passes if removed integer_keyed doesn't mean that vtable functions are not used.
21:25 NotFound And are used, in fact. Just put an abort(); in it to verify.
21:25 Kapace_ get_number_keyed_str
21:25 Kapace_ er, is being used?
21:26 NotFound Let me double-check...
21:28 NotFound I was wrong, I mistakenly placed it in the keyed_int variant
21:29 Patterner left #parrot
21:30 bluescreen left #parrot
21:31 whiteknight left #parrot
21:32 Coke but yah, just because tests pass doesn't mean they're not used. esp. since it probably has a PMC parent.
21:32 Coke er...
21:32 NotFound Note that not accesible from pir doesn't mean dead code. vtable functions can be used from extensions, from HLL that uses its own PMCs....
21:33 NotFound Default takes care, but the Default route is slower.
21:33 cotto_work moritz: spectest looks fine so far and both those test pass when I run them by themselves.
21:34 cotto_work latest Parrot and Rakudo
21:34 Coke mikehh: danke. I ran examples tests, but not the bench ones.
21:35 moritz cotto_work: hm. I'm going to wipe my install and try again
21:35 NotFound Someone has an idea about how to cover line 91 of this? http://tapir2.ro.vutbr.cz/cover/cover​-results/2010-12/2010-12-29-fc913cc/c​_cover/src-pmc-hashiterator-pmc.html
21:36 cotto_work moritz: I do see one failure for t/spec/S29-context/die.rakudo that's not TODO'd
21:40 moritz wow, that test file hasn't been touched in 2010 at all
21:42 rurban joined #parrot
21:42 cotto_work NotFound: ask bacek
21:42 cotto_work actually, luben wrote it
21:43 bacek aloha, who called my name?
21:43 aloha bacek: No clue. Sorry.
21:43 bacek yak... stupid girl.
21:44 cotto_work bacek: misfire
21:44 rurban left #parrot
21:44 cotto_work though if you know how to cover line 91 of http://tapir2.ro.vutbr.cz/cover/cover​-results/2010-12/2010-12-29-fc913cc/c​_cover/src-pmc-hashiterator-pmc.html, it'd be helpful
21:46 bacek cotto_work, no idea, sorry
21:46 Psyche^ joined #parrot
21:46 Psyche^ is now known as Patterner
21:46 NotFound I can "cover" it just by shortening a bit the way is written, but trigger the condition will be a lot better.
21:48 mtk left #parrot
21:52 Kapace_ If there are HLL users or something else that uses those functions, is there a way to test them?
21:54 Patterner left #parrot
21:54 NotFound Place an abort() in those functions and build rakudo with that parrot
21:54 Coke We can write tests for C code.
21:55 Coke ah, test that they are used rather than testing them for feature. yah, do what NotFound said.
21:56 NotFound If they are really unused and we are going to discuss its killing, there's no point in writing a difficult test
21:57 Kapace_ then make rakudo's tests, and if everything passes then, present that info to the ticket?
21:57 NotFound But I think we should have an op to call that vtables, like we have for the keyed_int variants.
21:58 NotFound And made imcc generate it, of course.
21:59 Kapace_ whiteknight said he couldn't find any op to access those vtables
21:59 Psyche^ joined #parrot
21:59 Psyche^ is now known as Patterner
22:03 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#1908) fulltest) at 2_11_0-343-g0264a39 - Kubuntu 10.10 amd64 (g++-4.5 with --optimize)
22:09 Patterner left #parrot
22:11 fperrad left #parrot
22:31 Psyche^ joined #parrot
22:31 Psyche^ is now known as Patterner
22:38 Yuki`N joined #parrot
22:40 Patterner left #parrot
22:43 cotto_work seen luben
22:43 aloha luben was last seen in #parrot 28 days 3 hours ago saying "yes. I was not expecting the merge to flood the channel".
22:44 mikehh rakudo (3750262) - builds on parrot (2_11_0-343-g0264a39) - make test, make spectest_smolder[(#1909), roast (f65933f)] PASS - Kubuntu 10.10 amd64 (g++-4.5 with --optimize)
22:45 Kapace_ I tried to build rakudo with clear parrot master, compile errors: http://dpaste.com/292091/
22:52 cotto_work Kapace_: what version of Rakudo?  It bitrots pretty quickly.
22:53 Kapace_ master
22:53 cotto_work ok.  I know that Parrot master and Rakudo master work fine for me as of now.
22:54 cotto_work was that a parallel build?
22:54 Kapace_ hmm.. I probably just need to try harder/ clean up
22:54 Kapace_ parallel build like -j3?
22:54 cotto_work yes
22:54 cotto_work also, did you install Parrot?  it could be seeing an older installed version
22:55 Kapace_ yeah I sudo make installed, but no this was normal build not parallel
22:55 cotto_work I'd try a reconfig and reinstall.  That shouldn't be broken.
22:56 Kapace_ ok
22:56 estrabd left #parrot
22:58 Psyche^ joined #parrot
23:04 Psyche^ left #parrot
23:04 Psyche^ joined #parrot
23:04 Psyche^ is now known as Patterner
23:08 Yuki`N dukeleto, ping
23:10 Patterner left #parrot
23:12 rfw am i allowed to make slavorg[n] trust people
23:14 rfw opbots trust Yuki`N
23:14 slavorgn Ok
23:14 slavorg Ok
23:14 rfw \o/
23:16 Kapace_ what is +o and why do we need it?
23:16 rfw Kapace_: channel operator
23:16 rfw i don't remember {how,why} i got mine
23:29 cotto_work With great power come great photo ops.
23:30 Psyche^ joined #parrot
23:30 Psyche^ is now known as Patterner
23:36 sorear lets you change the topic and kick people
23:37 Yuki`N But in reality nobody really uses that.
23:37 Yuki`N Maybe topic changing.
23:37 dalek parrot: 2368569 | nwellnhof++ | / (3 files):
23:37 dalek parrot: Fix deletion of zero integer hash keys
23:37 dalek parrot: review: https://github.com/parrot/parrot/commit/2368569db3
23:37 rfw also you get an @ next to your nick to confuse people unfamiliar with irc
23:37 * NotFound resisting temptation...
23:37 Yuki`N wwww
23:38 Yuki`N hurr that's where that went
23:38 Yuki`N ignore that
23:42 silug left #parrot
23:43 cotto_work nice fixes from nwellnhof today
23:48 silug joined #parrot
23:48 plobsing left #parrot
23:49 plobsing joined #parrot
23:54 Patterner left #parrot
23:55 Psyche^ joined #parrot
23:55 Psyche^ is now known as Patterner
23:56 nwellnhof joined #parrot

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

Parrot | source cross referenced