Parrotsketch Every Tuesday @ 18:30 UTC | IRC Log at irclog.perlgeek.de/parrotsketch/today
Set by moderator on 29 January 2009.
00:52 Tene_ joined 08:24 integral joined 10:50 Tene joined 15:12 davidfetter joined 16:02 Tene_ joined 16:04 masak joined 17:04 rurban joined 17:45 PacoLinux joined 17:46 wknight8111 joined 18:08 barney joined 18:16 allison joined 18:17 pmichaud joined 18:18 NotFound joined 18:20 Util joined
pmichaud I may be 5-10 late to #ps 18:21
18:28 chromatic joined 18:29 kj joined
chromatic Hello everyone. 18:30
kj hello there 18:31
Whiteknight hello
rurban here
masak hi
cotto hi
NotFound ho
Util hi 18:32
chromatic Let's begin.
allison?
pmichaud here.
chromatic barney?
allison here 18:33
- Started a branch to streamline our calling conventions, addressing some speed issues reported in MMD and I/O (but likely affecting all cross-over between C and Parrot calling conventions).
- Fixed persistent install location problems in install script.
- Fixed package annotations in MANIFEST and MANIFEST.generated.
- Changed install to use proper versioned directories (e.g. /usr/lib/parrot/0.9.0-devel/library/...), and changed runtime library paths to search versioned directories. (Every compiled parrot knows its version, #define'd as PARROT_VERSION.)
- Added an 'install-dev' target, which installs the necessary tools to compile and build a language. (Not needed to simply run a language.)
- Replaced tools/build/dynoplibs.pl and tools/build/dynpmc.pl with makefiles, will commit after release. (We have a highly-developed infrastructure for template-generated makefiles.) 18:34
- Fixed various bugs and hardcoded paths in the Parrot::Pmc2c modules, to allow them to run from an installed copy.
- Patched Rakudo's Configure.pl and Makefile.in to work off an installed Parrot, and added makefiles for Rakudo's dynpmcs and dynops. Waiting for TT #343 to be applied after the release to fix a remaining failure there.
EOR
chromatic barney?
barney Busy with $DAYJOB. Won't make it to German Perl Workshop next week.
Moved Pipp to github.
.eor 18:35
chromatic I have been poking at some tickets.
I'm also looking at a log of the core test run and the SKIP and TODO tests, and I'll keep working on that today.
cotto?
cotto * "DOD" is ded
eor
chromatic japhb? 18:36
18:37 Infinoid joined
chromatic kj? 18:37
kj preparing release
will start soon after #ps
no other activities
i have 2 questions
and many thanks to people who have been working hard to fix tests this week 18:38
eor
chromatic Infinoid?
Infinoid ENOREPORT 18:39
chromatic masak?
masak * lots of fun Proto (perl 6 project installer) and Druid (board game) hacking
* a smattering of bug reports
* no actual Rakudo development, unfortunately
.eor
chromatic NotFound? 18:40
NotFound * some work in parrot_debugger and debugger ops
* work on embedding and related changes on headers, ready to apply after release, TT #343
* miscellaneous cleaning and fixes
1 question
EOR
chromatic particle?
rurban? 18:41
rurban - Added experimental pmc/bignum with two open bugs (sorry, no tickets yet).
- More work on float bit pbc transformers, 64 bit still broken somewhere.
- Installed ancient openbsd 3.8 and verified shared problem, but not the open -0.0 problem yet.
- Applied several patches.
- Added some more tickets.
EOR 18:42
chromatic Tene_?
Tene_ Busy with $dayjob and staying AFK for the rest of the day for RSI and other sanity reasons. Did nothing Parrot-related. EOR
chromatic Util?
Util Traveling; Only time for cage work this week, straightening out SVN repo properties.
Back to Rakudo sanity tests on Monday. 18:43
EOR
chromatic Whiteknight?
Whiteknight CORE
- Did some random GC cleanups, looking to find a better home for platform-specific GC code that's living in src/gc/system.c
- Ripped out src/bignum.*
- Converted VTABLE_morph to take a class PMC instead of an INTVAL type id.
- Been working to replace Parrot_PCCINVOKE with Parrot_pcc_invoke_method_from_c_args. Tough going so far.
- Still working on that in the rename_pccinvoke branch,
BOOK
- added information to docs/book about various PMCs, including their vtables, methods, and uses
- Wrote first section about IO
- put out a call for documentation-reviewing help on the parrotblog. Got lots of good responses
MISC
- Getting working on my Octave port again. Found another developer interested in the project. Details to come.
chromatic pmichaud? 18:44
pmichaud - primarily worked on Rakudo build issues, and answering miscellaneous questions
- updated Rakudo so that it can work with Test::Harness and 'prove'
- working on a rewrite of t/harness for this
- added a '--gen-parrot' option to Rakudo's Configure.pl to automatically download and build an appropriate copy of parrot 18:45
- refactored Rakudo's Makefile a fair bit, more refactors coming
EOR
chromatic Did I miss anyone? 18:46
Let's do a quick milestone review. 18:47
PDD 14 numbers -- no progress from me. particle?
Tene_ No progress from me on anything I'm listed on.
kj chromatic: shall I mark that as missed?
chromatic Let's start shuffling them tomorrow, if we don't hit them today. 18:48
rurban I've added get_bignum and set_bignum
chromatic As far as I know, the HLL class registry works; TT #71 is closed.
PDD 26 AST, pmichaud and Whiteknight?
Whiteknight is my name on that> 18:49
?
pmichaud no progress, but expect to have it done this week.
chromatic Sorry, that was Tene_.
PDD 19 PIR, Whiteknight and Tene_?
allison looks like the book covers that
Whiteknight Yeah, I think that's on track 18:50
chromatic On track or landed?
kj the "user docs" part is the book, yes, but pdd19?
Whiteknight the milestone was to improve PDD19-related documenation
kj (afaic, pdd19 is complete enough to mark as 'landed')
Whiteknight let's call it landed, but the book is still a work in progress 18:51
allison kj: I'd agree
chromatic dynops and dynpmc creation from installed Parrot, allison and pmichaud?
pmichaud allison's report indicates it's largely finished, I guess?
allison chromatic: nearly there, running into a few include problems in an installed parrrot 18:52
pmichaud I haven't had a chance to test it yet.
chromatic I'll mark that as on track then.
kj allison: when do you expect to have that done?
allison pmichaud: but, getting safely all the way through generating the C files
chromatic todo/skip review is in progress.
PDD planning and milestone edits, allison and pmichaud?
allison kj: I don't want to make more commits today, with the release coming up, but will commit right after release (so, today or tomorrow)
kj ok. just before we all agree that all these things are 'on track'; if they're not 'landed' within 2 hours, they'll be 'missed' ;-) 18:53
allison planning and milestone edits: landed
kj: yes, they'll be "missed/rescheduled"
chromatic PDD 23 exceptions, Tene_ ? 18:54
kj (but I agree we can mark them 'on track' for the next release)
Tene_ chromatic: I havdn't done any work.
chromatic PDD 20 lexicals, pmichaud?
pmichaud didn't make it for this release, but I should be able to have it by next #ps 18:55
(barring any further unforeseen non-Parrot events)
q2q
chromatic Windows porting, particle and allison? 18:56
allison there's been some good work there this month 18:57
I'd call it landed, though we'll still do general platform testing next month
chromatic on track? landed? A vast portal of despair?
Resolving mmd tickets, allison?
allison missed
(I picked the "critical" task first)
kj just curious: what exactly was windows porting/subprocesses? 18:58
I've been a bit inactive lately, so I didn't really got that
allison kj: essentially, give a good review of Parrot on windows, make sure the tickets were getting attention
kj ok.thx. 18:59
with the several tasks moved to next month's release, and the tasks that were already there, what's the level of confidence in making them all?
chromatic Very little of next month's work is code.
kj chromatic: true.. 19:00
rurban confident
pmichaud that prompts my question, though: is the calling conventions re-work a milestone for March release?
Tene_ And do we have a track record of better productivity at docs than at code?
chromatic Not so far...
pmichaud and is it a 'critical' milestone?
i.e., do we go ahead and release Parrot 1.0 with the current calling conventions in place?
Tene_ "not code" doesn't necessarily mean easier.
allison pmichaud: it's not in a milestone for March
NotFound IMO a solution for the invoke problem is critical 19:01
kj NotFound: what's the invoke problem?
pmichaud allison: as in, it's not a 1.0 feature?
NotFound kj: be able to call: object()
allison pmichaud: I wouldn't consider any current calling conventions a show-stopper for 1.0, more "nice to have"
pmichaud allison: okay.
NotFound Instead of having to do: object(object) 19:02
kj allison: the calling-conventions refactor doesn't mean a changing interface right?
allison pmichaud: and about optimization
kj NotFound: ok self. I see.
Whiteknight a lot of cc-related fixes will land next month, but they will land piecewise
pmichaud allison: it won't affect language-specific PMCs?
allison Whiteknight: and may be delayed until after 1.0, depending on their stability
Whiteknight True. some are going to be more invasive then others 19:03
pmichaud NotFound: we've lived for years without being able to invoke objects. I don't know why it would now be "critical".
Whiteknight pmichaud: they *should* be mostly transparent to PMCs
pmichaud I'm not saying it's unimportant -- Rakudo and PGE and PCT could really use the ability to invoke objects, but it hasn't been a showstopper. 19:04
allison basically, at this point, a pile of unstable changes to the calling conventions is much worse than a handful of know but non-disruptive problems
NotFound pmichaud: I think is very important for language and libraries interoperability
pmichaud NotFound: interoperability is 1.5
allison pmichaud: aye 19:05
NotFound pmichaud: rigth, but will be far better if the basis were stablished before. Maybe not critical, aye.
pmichaud we do all need to remember that 1.5 is only four months after 1.0 19:06
chromatic We can't afford to do a lot of non-critical tasks if they delay critical tasks.
pmichaud that's basically the amount of time that existed between PDS and 1.0
do we have any critical milestones that haven't landed? 19:07
looks like just the dynoplibs issue
rurban pdd14 numbers maybe
pmichaud pdd14 numbers isn't listed as critical. 19:08
allison rurban: not critical
pmichaud (nor do I think it's critical)
chromatic Let's go through the list of tasks for next month very quickly. Some are missing people.
pmichaud "critical" in this context means "we don't ship 1.0 without it"
chromatic bug/issue tracking and triage guidelines
I can work on that.
pmichaud list me down for the tutorials
NotFound clean up core headers - TT #343 is a step 19:09
pmichaud some of that is done already, but I really need to get it all in place for my upcoming talks
chromatic PCT user docs, pmichaud and Tene_.
Move languages out of repo/tarball.
Lots of coke, fperrad there. 19:10
pmichaud I'm planning to move pynie to github
Tene_ No idea whether I'll have cycles or not.
chromatic Web site updates?
allison pmichaud: does python use git?
pmichaud shall we leave lolcode in the repo or move it out? I think it's a useful demo language, but I can go either way.
chromatic +1 to keep
Tene_ +1 keep 19:11
pmichaud allison: there's some discussion about that within the python community, iiuc
kj especially if we can get the any-to-lolcode converter :-)
pmichaud kj: that's coming also :-)
kj pmichaud: I was hoping for that :-) That's really a USP!
chromatic Anyone for web site updates?
allison I want to move all languages out of the repository, even if only to parrot-languages on svn.parrot.org 19:12
Util keep++ ; I am still feeding the video of PM's talk to PerlMongers, stirring up interest in Parrot.
PerlJam random opinion from the lurker: abc and lolcode are both nice languages to keep with parrot.
pmichaud at any rate, I really like git's model, so it's a good choice for pynie. If/when someone wants to lead development of pynie, they can easily move it somewhere else.
allison abc and lolcode could be moved to "examples"
Infinoid chromatic: Given the proper access, I can work on website stuff.
chromatic user forum?
NotFound We need some languages in the repo if we want to be able to have tests of language loading and compiling 19:13
allison pmichaud: I was thinking I might invest in pynie after 1.0, but doesn't matter to me where it's hosted
pmichaud allison: I'll be curious if you find any takers at the pycon devel summit
allison pmichaud: me too
Infinoid: will make sure you have access 19:14
chromatic Integrated language testing?
Infinoid A task list more specific than "web site updates" would help too :)
kj Infinoid: I think most of it would be moving the old website? 19:15
Infinoid Moving wiki pages and stuff? Can do.
pmichaud I think there are tickets for web site update thingies.
kj so there's no bits in the old site anymore.
Infinoid great, I'll have a look
pmichaud closing out the perlfoundation wiki pages would be good.
I can do integrated language testing for NQP.
can't speak for other langs. 19:16
allison Infinoid: also, scanning the pages looking for references to old resources, old ways of submitting bugs, etc
Infinoid Makes sense. I probably already have enough access for that task, I'll raise an interrupt if I run into any trouble
kj next task is: make html 19:17
allison Infinoid: I don't see an "Infinoid" account on www.parrot.org
kj: I fixed up 'make html' last month 19:18
kj allison: so 'landed'?
rurban almost
kj 'on track' then
allison kj: what it needs, more than anything is a clean up of the docs content, but that's a separate task on the list
kj: that ticket may have also had the 'docs.parrot.org' site attached to it 19:19
kj allison: there's so much old documentation, hard to know what to keep and what not
it'd be a waste just to throw it out
pmichaud toss anyting that's wrong. :-)
allison kj: yes
pmichaud *anything
allison kj: but, we don't have to include all of it in 'make html' 19:20
kj allison: true.
allison and, yes, toss out the stuff that's wrong
NotFound docs.parrot.org does not resolve yet
allison NotFound: yes, hasn't been created yet
NotFound: a ticket has been filed with our admins 19:21
(requesting the domain)
Util I will have time after Monday to help with website updates. I would also need access. (did not parse c's statement as a call for volunteers, at first)
kj Whiteknight: ping
allison I don't see a Util account on www.parrot.org either, create one and I'll grant access 19:22
kj oh sorry, wrong channel
Util ok, thx
chromatic Are there any other questions?
Infinoid I've got a minor one.
kj I have 3
NotFound Me 1
chromatic kj?
kj 1: copyright 19:23
Util 1Q
kj all files except LICENSE will be Parrot Foundation ?
including all perl scripts, basically any file
allison also not Pod::Simple
or other external Perl modules in the repository)
otherwise, yes 19:24
because Rakudo already moved out
kj ok, so I'll leave most of lib/ alone then
allison most of lib is lib/Parrot, which is Parrot Foundation
kj: if you want to divide by directories, I can help on the copyrights 19:25
kj great; if you can make a list or whatever
allison kj: would like to have that done before the 0.9.1 release
kj of directories, that are to stay (c) TPF
allison kj: none stay TPF
kj ok, I can spare a bit, but I'd rather not postpone the release too long, otherwise it'll be very late here...
allison kj: if it's not TPF, then it stays as whatever it is 19:26
kj allison: but what about the exernal perl mdoules?
(you just said...)
allison external perl modules aren't TPF, they're various external owners
kj ooh ok. So bascially, if it's (c) TPF, then it'lll be (c) PF
allison kj: right
Util Cancel Q; pmichaud answered in email (rakudo-commits not being fed)
kj ok, great. that's easy
Q2:
ports/cygwin 19:27
there's some patch files
rurban yep
kj do we need them?
rurban that's what I've used for the latest release
kj (just like to know why there's patch files named parrot-0.8.2-1....)
rurban I didn't make any 0.9.0
0.8.2-1 is the latest release 19:28
kj rurban: ok well I can harass you about this on #parrot, is that ok with you? (so we can move on here)
allison kj: those don't go into a release tarball, so really doesn't matter
kj that's automated? ( I mean, they'll be excluded automatically?) 19:29
allison kj: yes, the whole ports/ directory is excluded from the tarball 19:30
kj (in the mean time, Q3, last one: .include search order, in pdd19 there's a remark saying that the search order may be changed
it's in this ugly {{ NOTE ME }} notation. I'd like to get rid of that
(esp. because we mark pdd19 as 'landed' ;-) 19:31
allison kj: feel free to change to a regular sentence in the paragraph
kj okido. that's clear enough for me
thanks, eoq for now.
chromatic NotFound? 19:32
NotFound Install paths and compiling extern/embed: #include "parrot/*.h" will not work if we have include/parrot/release-number/*.h
allison NotFound: yes, need to strip the leading "parrot/" and just use "*.h" 19:33
NotFound: also need to move all .h files into include/parrot, instead of scattered all over the repository. 19:34
NotFound allison: that can be confusing, we have header names too generic
kj like string.h
allison NotFound: yes, but they're buried in a subdirectory, only accessed if someone explicitly adds -I 19:35
alternatively we can name everything parrot_*.h
but hardcoding paths is a disaster for an install
(this is the current problem that's biting me on building dynoplibs and dynpmcs from an installed Parrot) 19:36
rurban make includizer
spinclad .oO { <libdir>/parrot/<release-number>/include/parrot/ , and #include "parrot/*.h" ? } 19:37
allison spinclad: was <libdir>/parrot/<release-number>/*.h
spinclad: but I had to change it to <libdir>/parrot/<release-number>/parrot/*.h to work
NotFound I think a directory is better than a name prefix, because it can be symbollic linked into
allison you don't want to symbolically link include directories 19:38
the name prefix means it's always the same everywhere
NotFound allison: lots of packagers do it
allison you -I the directory, and you get the files, doesn't matter where they are 19:39
Infinoid NotFound: The two are not mutually exclusive.
NotFound Infinoid: if you have a directory, not 19:40
Infinoid <libdir>/parrot/<release-number>/include is enough of a directory to symlink to if you need to choose which version you want to build against 19:41
chromatic Are there any other questions for #ps, or should we move to #parrot?
rurban symlinks dont work on windows
allison chromatic: yes, this is a discussion for later
NotFound Ok 19:42
spinclad -> Infinoid.question { ... } 19:43
chromatic Any other questions? kj, did we get all of yours?
kj chromatic: yes all my qs were answered, thx
chromatic spinclad?
Infinoid Is everyone happy with dalek's behavior? The poor bot's now tracking 18 rss feeds, I've bumped up the polling interval accordingly.
I'm looking for more things to track, dropped commits, formatting errors... 19:44
NotFound Infinoid: looks like sometimes it miss parrot commits
Infinoid NotFound: Recently? I applied a fix 3 or 4 days ago which I hoped would fix that 19:45
(trac was taking too long to generate the changeset page to parse the filenames from)
NotFound Infinoid: not sure, maybe fixed, then
Infinoid If you see it happen again, please yell at me, loudly.
Okay. Anyway, if anyone has a problem with dalek, you guys know where to find me. 19:46
eoq
chromatic Other questions?
Okay, let's call it a week. 19:47
Finish those milestones, but don't get in kj's way.
19:47 chromatic left, Infinoid left, Util left, NotFound left 19:52 kj left 20:20 rurban_ joined 20:34 rurban left 21:56 Whiteknight joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!