Parrotsketch Every Tuesday @ 18:30 UTC | IRC Log at irclog.perlgeek.de/parrotsketch/today
Set by moderator on 18 February 2009.
00:28 japhb joined 03:31 Tene_ joined 14:30 diakopter joined 15:31 Tene joined 15:57 rdice joined 16:21 diakopter joined 16:35 Whiteknight joined 17:04 davidfetter joined 17:05 pmichaud joined 17:07 Infinoid joined, Coke joined, NotFound joined 17:26 Util joined 17:46 PacoLinux joined 17:54 moritz joined 18:12 PacoLinux left, PacoLinux joined 18:13 barney joined 18:24 allison joined
Coke Early greetings. 18:27
Tene Hi! 18:28
moritz hello 18:30
Whiteknight hello
moritz it's time now
barney Hallo
NotFound hi
Coke Alright. lacking a chromatic, I'll lead.
cotto hi
Coke My report: 18:31
- another issue of TWIP, this one rather lame and placeholdery.
- cleaned up some of the experimental ops. (TT #415 for the rest)
- moved languages/APL to code.google.com/p/paraplegic/
- make test harness use new runcore options and ripped out the old ones. some of the tests rely on the old short versions, much of the fallout fixed by rg++. [1]
- minor progress trying to get 'make examples_tests' 100% clean by 1.0 - may involve deleting extremly old, broken examples.
[1] Highlights that once something is deprecated and a replacement exists, we should start warning on the old behavior so we can quickly remove any old usage. We tried to do that for the runcore options, but failed; and the warnings from the old "-j" options were swallowed by the harness. So we thought we had fixed this months ago but hadn't.
.
Whiteknight?
Whiteknight * Been spending most of my time working on Matrixy, the new M-on-Parrot compiler
* Testing on Ubuntu8.10 x86_64, WinVista x86_64, and WinVista x86. Opening tickets as I find errors.
* Calling conventions work on hold till after the release for stability.
EOR
Coke Thanks. 18:32
barney?
barney Pipp:
Worked on building, testing and installing outside the parrot tree.
http;//pipp.org now redirects to wiki.github.com/bschmalhofer/pipp
cotto++ fixed critical bug.
.eor
Coke Danke.
Since you mentioned cotto... cotto?
cotto * converted more PMCs to ATTRs
- Continuation (+ RetContinuation, ExceptionHandler)
- ParrotLibrary
- ParrotInterpreter (+ ParrotThread, ParrotRunningThread)
- Rational, GDBMHash
- FixedPMCArray (+ ResizablePMCArray )
* no HLL breakage thanks to the wonders of VTABLE functions
* the remaining PMCs (Hash and ManagedStruct families) should be easy, having no impact on HLLs
EOR
q1q 18:33
interesting order, btw
allison here
Coke Thanks. Alphabetical is officially out of date.
moritz so web 1.0 ;-)
Coke moritz, you're up.
moritz * some Rakudo hacking
* trying to answer Rakudo questions on #perl6 - they seem to increase 18:34
18:34 chromatic joined
moritz * spent rest of my free time writing articles about Perl 6 18:34
* observed that Rakudo works fine on optimized parrot (on amd64)
.eor
Coke (optimized parrot) nice.
NotFound, you're up.
moritz aye ;-)
NotFound * Some more work in extend/embed
18:35 kj joined
NotFound * Added the program pbc_checker 18:35
* Working on annotations, showing them in traces and backtraces
* Miscellaneous cleans and fixes
EOR
Coke danke. 18:36
kj comes in late, without a report due to $work
Coke so reported.
chromatic, how goes?
chromatic Working on the bug triaging guidelines; will have a draft in a couple of hours. 18:37
q1q about where to put it.
Util here
Coke util, shoot.
Util Many typos and codingstd fixes.
More to come. 18:38
`make fulltest` vs `make fulltest_all` in TT#432 resolved.
Fixing copyright formatting, toward re-enabling the .t that is skipping it.
Emailing remaining copyright hold-outs.
Alert: main LICENSE file says "Perl Foundation"!
Found evidence of a a hole in the release procedures, causing parrot.spec and other files edited during the release to to have DOS line-endings, only in the tarball. Ticket in the works.
Plan to continue along these lines until after the 1.0 release, then to help Infinoid with docs. 18:39
EOR
Coke q1q in response to util.
allison, i think you're all that's left. anyone else have a report but hasn't spoken up yet? 18:40
pmichaud (here, hadn't spoken up yet)
Tene
Coke ok. tene? 18:41
(whoops, I missed you from the early responders.)
Tene Giving a presentation on PCT to the local LUG tomorrow. .EOR
(You didn't say a *good* report)
Coke well, let's follow that up with patrick, who no doubt will make up for you in volume. =-)
(with pmichaud, in case he's filtering on his nick. =-) 18:42
pmichaud doubtful, today. =-)
Primarily I've been working on rakudo tasks, documentation for developers and cleaning up rakudo tickets
I expect to spend the next day or so closing out my parrot release tasks
I'm also reviewing the impact of recent Perl 6 specification changes on PGE and PCT -- q1q regarding this 18:43
EOR
Coke alrightee.
(you still beat tene.)
Allison, you're up.
allison - Ticket maintenance, applied patches, fixed bugs, answered questions.
- Test failure review, fix/TODO/skip. 18:44
- Worked on Debian/Ubuntu packaging, prepping for 1.0.
- Documentation cleanup, revisions for clarity, revisions for out-of-date content, revisions for coding standards, gave 'make html' a detailed TOC, added version number to generated HTML.
EOR & q1q
Coke alright. my question for util was resolved in #parrot, so I'm done.
let's jump to chromatic's which is hopefully quick.
chromatic Where should I put the bug triaging guidelines document? 18:45
allison chromatic: docs/project seems sensible
Coke +1
chromatic docs/project/bug_triaging.pod ?
Coke perhaps the more generic "issue..." 18:46
but I don't care that much. 18:47
docs/project/bug_triaging.pod it is.
allison either is good
Coke alrightee. Cotto?
cotto There are two groups of PMCs left to convert to ATTRs (Hash and ManagedStruct). Assuming that all tests pass (and rakudo's make test is happy), what's the latest I should commit changes to these PMCs before the 1.0 release? 18:48
Coke before the weekend, I'd say.
chromatic It'd be nice to have at least all day Sunday and Monday to review smokes and fix things.
cotto sounds good 18:49
Whiteknight cotto, you need any help with that?
I'll take one if you take the other, etc
Coke You can divvy it up in #parrot. =-)
Util q1q 18:50
Coke Util: just go now. =-)
cotto eoq
Util Will languages/cola leave the nest before parrot's 1.0 release?
Melvin Smith's copyright statements are the biggest block to un-SKIPping t/codingstd/copyright.t test#2.
allison Util: yes
Coke Util: it will be deleted, if nothing else.
Util Thanks.
Coke pmichaud: shoot.
Tene q1q 18:51
pmichaud This past week has seen a lot of substantial changes to Match objects (what gets produced by PGE)
in the Perl 6 specification
both PGE and the compiler toolkit will need to be updated to match the spec
which means we have a deprecation issue 18:52
if I put notes for all of the deprecations into the current trunk (and in the 1.0 release) -- will anyone have an issue if they get removed shortly thereafter?
allison this has happened before, and we did parallel development for a while, and then deprecated the old one
Coke pmichaud: no, that's the plan on deprecations.
pmichaud this isn't something that's conducive to parallel development.
before the issue was syntax changes -- so I could implement the new syntax while supporting the old.
The new changes are fundamentally semantic. 18:53
Coke put in the deprecation notices now so they're in the release; then as soon as 1.0 is cut you can change it.
allison ok, then yes, standard deprecation policy, which means you can make the changes right after 1.0 if needed
Coke (sooner the better, so more testing before 1.4)
all set?
pmichaud okay. Where possible I will try to support the deprecated stuff until other languages have time to switch over, but some things might break shortly after 1.0.
thanks. 18:54
Coke as long as it's shortly after and not before. =-)
tene, you're next.
Tene How urgent is the request I saw on the ML to move languages out of trunk?
Do I need to relocate cardinal and lolcode?
allison Tene: this leads into my question 18:55
Tene I've only vaguely been kinda following it.
Coke then lets jump to allison's question. =-)
allison The languages/ directory is going away today (or within the next couple of days)
pmichaud I'm planning to move pynie to squawk
allison pmichaud: I'd rather move pynie to a dedicated repository
not within any umbrella repository 18:56
pmichaud okay, I can create a dedicated one then. Googlecode, I guess?
allison don't care if it's git or svn, though
googlecode sounds good (since Guido works there)
pmichaud one potential pynie developer I spoke with thought that svn might be better than git for acceptance
allison pmichaud: okay, useful info
My general Q: Other languages to move before I start relocating the remaining ones? 18:57
pmichaud where will abc and lolcode go?
allison I've currently got:
- abc, lolcode, and squaak -> examples/languages
- pheme, dotnet, c99, pod, json, PIR, ecmascript -> svn.parrot.org/languages
- cardinal -> ? dedicated repository?
Tene I'll put it on github
pmichaud for cardinal I highly recommend github 18:58
Tene: would you like me to do the transfer (preserving history), since I already did it once for rakudo?
Coke OOC, why is lolcode going to examples ?
allison Everything else that hasn't already been claimed or moved, is deleted.
Tene pmichaud: sure, that would be great
Coke (and not to svn.parrot.org/languages)
allison Coke: someone suggested it, don't remember who, would be just as happy to put it in the languages repo 18:59
pmichaud Tene: do we need a special cardinal account on github, or shall I just do it into mine (and someone can pull from there into a new master repo)?
chromatic lolcode has to go to examples/ so we can have a Perl 6 to lolcode converter.
Coke chromatic: ok, you win.
pmichaud actually, the lolcode conversion becomes part of PCT :-P
cotto +a lot
allison chromatic: :) we can do that in the languages repo too :)
Tene pmichaud: I don't know the benefits of having a dedicated account.
Coke that can be hashed out in #parrot. Any other language related questions? 19:00
pmichaud Tene: in rakudo's case, I did it primarily so that Rakudo wouldn't be attached to me personally.
Coke (and, any other reports we're missing?)
19:00 barney joined
pmichaud Also, I could potentially give out the password to rakudo's account on github to trusted folks without having to give out my own password. 19:00
(so there can be multiple meta-maintainers) 19:01
allison pmichaud: the same reason (not attached to person) seems to make sense for cardinal too
pmichaud we'll create a separate account for it.
Coke +1
Anyone have any other business? I have a general comment about documentation; we all need to be better about writing down decisions and pruning old stuff out of the repository. I suspect we'll see a lot of cleanup in the next week in that area as we get close to the 1.0 release. (I know that would have avoided some on-list confusion this week.) 19:02
Tene I know I've said it before, but I plan to work on docs the coming week. 19:03
Coke +1
NotFound Forgot I have a question
Coke nooooo
pmichaud (better than having a forgotten question)
Coke er, notfound, typo there. go ahead.
NotFound Short, I promise
allison Coke: +1 on documentation
NotFound pdd13 description of segmente header lacks some fields. Must we make it to refelect current state? 19:04
Coke yes, even if only to say "these fields exist but will be removed."
Infinoid peanut gallery: pdd13 is also out of date with regards to annotations
allison NotFound/Infinoid: could you submit patches or make fixes to pdd13 for segment header and annotations? 19:05
NotFound And they will be removed? There are no comments in the code indicating that.
And one of them is actually used in several segment types. 19:06
Infinoid I don't understand jonathan++'s recent annotations changes well enough to patch pdd13, otherwise I would love to.
Coke NotFound: my comment wasn't meant to imply that was the case, merely that it was a possibility.
allison NotFound: generally, the PDD reflects future plans, but something missing could either be intentionally excluded or simply missed
Coke in /either/ case, they should be in the docs.
cotto q1q (quick) 19:07
Infinoid in the case of annotations, pdd13 is outdated, not the current state of parrot
allison and, yes, need to be documented, we can make a decision later
Infinoid: then pdd13 definitely needs to be updated
Coke so, updates to pdd13 needed; a ticket to trac progress on that might be good.
NotFound allison: I think an accurate description of the current format is a must-have.
Coke agreed, even if it is subject to change.
allison NotFound: yes
Coke alrightee. cotto, you're back on the list. Shoot. 19:08
cotto Should docs/pmc.pod (which I'll be updating) be moved to docs/pmc/pmc.pod or otherwise renamed?
Coke docs/pmc/*.pod is for pod from pmcs.
allison cotto: leave it where it is
Coke generic docs about pmcs can probably stay in docs.
cotto ok. I'll update it in-place. 19:09
eoq
Coke woot. the queue is empty.
allison, I'll let you close it out with something heartfelt about next week's release, and we can all hold hands and hum.
NotFound Infinoid: What is the problem with annotations? I managed to work with them without much problem. 19:10
allison Oooooohm...
Infinoid NotFound: the PMC API documented in PDD13 doesn't reflect the current layout of structures and functions
NotFound Infinoid: ah, fine, I was looking just at the pbc format.
Infinoid Some extra stuff was added after pdd13 was written, and some things have been moved around 19:11
Yeah, and I'm just looking at the API :)
allison I'm planning to generate the tarball early this month, and pass it around for platform testing.
(that is, early on release day)
NotFound Infinoid: in the api part, I'd like to discuss a bit the lookup function.
But out of #ps, better 19:12
Infinoid NotFound: yeah, moving to #parrot
Coke allison: all set/ 19:13
?
allison yup
Thanks everybody!
Coke 7 days until 1.0; Tell your friends.
.
19:13 Coke left 19:14 moritz left, Util left, Infinoid left, PacoLinux left 19:15 chromatic left 19:17 allison left, diakopter left 19:24 contingencyplan joined 19:43 kj left 19:54 NotFound left 21:23 PacoLinux joined 21:43 davidfetter joined 21:58 Whiteknight joined 22:41 Tene_ joined 23:15 PacoLinux joined 23:18 PacoLinux left