"Tuesday at 20:30 UTC"
Set by moderator on 20 August 2010.
01:24 whiteknight left 01:44 eternaleye left 01:45 eternaleye joined, eternaleye left 01:46 eternaleye joined 01:47 eternaleye left, eternaleye joined 02:18 tcurtis joined 02:25 contingencyplan joined 06:17 integral left, integral joined 07:02 tcurtis left 07:52 integral left 07:54 integral joined 11:07 kid51 joined
kid51 kid51's report 11:08
* Tried to close several old tickets.
* trac.parrot.org/parrot/ticket/1255: PARROT_EXPORT visibility in gcc 3.3: will close tomorrow unless someone objects.
* trac.parrot.org/parrot/ticket/1725: headerizer.pl documentation: will close tomorrow unless someone objects.
* trac.parrot.org/parrot/ticket/1751: links to DEPRECATED.pod: ticket opened by Paul_the_Greek; made small changes, had discussion
* trac.parrot.org/parrot/ticket/677: what should go in tools/dev/, tools/build/, etc?: developing proposal; please see ticket.
* trac.parrot.org/parrot/ticket/1750: packfile testing problems on Darwin/PPC: still befuddled; sometimes tests pass, sometimes they don't
* Posted comments in several tickets owned by mikehh and others stating that I thought such tickets were now closable. 11:09
EOR
12:11 bluescreen joined, kid51 left, integral left, PerlJam left, pmichaud left 12:12 contingencyplan left, Coke left, Util left, cotto left, Tene left 12:14 PerlJam joined 12:23 integral joined 12:48 Tene joined, cotto joined, Util joined, Coke joined, contingencyplan joined, pmichaud joined 15:17 mikehh joined 15:29 tcurtis joined 15:52 cotto_work joined 16:03 SingAlong joined 17:31 mikehh left 17:42 mikehh joined
mikehh What I did since my last report: 18:30
* released parrot 2.7.0
- my daughter-in-law had a stroke (fortunately mild) on Sunday night (15th)
- this resulted my missing most of Monday and Tuesday before the release
- I got everything ready after #ps last week, somehow managed to overwrite
- the NEWS I had been keeping updated and had to re-create it from scratch,
- but after making the changes had a test failure in t/pmc/threads.t
- (not in trunk, just the release directory) I eventually committed what I
- had got to see if anyone else had the problem. I then did a clean checkout
- to a new directory and that passed all tests. Still have not worked out
- what the problem was. Tagged the release and with help from cotto++ got it
- up on the site. Put a release notice on Parrot.org then, being totally
- exhausted had to get some sleep. When I woke up, I posted other notices,
- not sure if I got everything, or if I missed anything. The rest of the
- week was a complete write-off, getting family business sorted.
- Only managed to get back to building and testing parrot today.
- (excuse the long post - just had to rant/comment a bit)
* building and testing parrot on amd64/i386, with gcc/g++
* some fixes
What I intend to do in the next week:
* what I intended to do last week
* work on html_cleanup branch
* testing and fixing
.eor
18:52 plobsing joined 18:55 TimToady joined 18:57 nwellnhof joined
nwellnhof What I did 18:57
* got my commit bit
* started to hang out on #parrot 18:58
* finished my GC timing fixes
* fixed some GC leaks
* optimized hash access 18:59
* closed TT#1603, #1465, #945, #1121 and #1737 19:00
Plans/ideas
* distinction between charset and encoding for strings doesn't make sense 19:02
- they could be merged which would simplify the code
* hash cloning could be optimized by storing bucket indices instead of bucket pointers in hashes 19:03
EOR
19:16 whiteknight joined
whiteknight WHAT I DID: 19:20
* Still working on PLA.
* Improved support for BLAS (including support for the reference version written in Fortran)
* Adding support for LAPACK, looking for a sane way to add an interface to it without going overboard in bloat
* Ran into a problem in distutils (#1755) that makes life a little harder
* Following along with all the activity in GC and Hash code. +1 on everybody's effort!
WHAT I WILL DO:
* Continue work on PLA to get LAPACK bindings up and working
ROADBLOCKS:
* None
19:24 particle1 is now known as particle
cotto_work #done: 19:28
- good progress on the github plugin
- can detect links (git and svn) and redirect them to github correctly
- stub tooltip support (will expand shortly)
- co-conspirator: dafrito++ for bugfixes and a testing tool 19:29
- trac-- for having a strange and unclearly documented plugin interface
- I should have something to play with by next #ps
- code is at github.com/cotto/github-trac
#hope to:
- finish the github plugin, set up a test site to show how it works
#eor
Coke Did: hacked on partcl-nqp, some TT (parrot) and RT (rakudo) chasing. Will miss #ps this week due to work. 19:30
will: hack on partcl-nqp. 19:31
.eor
19:48 dukeleto joined
plobsing ~ What I Did: 19:49
* not much (no internet access for a while) 19:50
* a couple rakudo startup improvements
What I Plan:
* more rakudo startup speedups
* eliminate costly collision checking in hash thaw
* manipulate packfiles more efficiently in imageio
* land gsoc_nci branch
EOR
dukeleto What I did: 19:59
* Worked on parrot github mirror. I now have a MooseX::Runnable app that will sync
all svn branches to github, created new git tags from svn tags and remove dead
branches from github when they are deleted from svn. Still needs some docs,
and could be abstracted for other projects, but this is what it looks like now:
github.com/leto/Util/blob/master/pe.../Parrot.pm
* Did GSoC stuff
* Tried to close some tickets 20:00
* Worked on PL/Parrot. Running into this coredump gist.github.com/546982
on this PL/Parrot branch: github.com/leto/plparrot/tree/global_grammar
* Gave a talk about PL/Parrot at PDXPUG, slides :
github.com/leto/presentations/raw/m...G/pres.pdf
What I will do:
* Setup the parrot github mirror so that it is run from feather or some other server.
Currently I am running it manually to make sure it does not set things on fire.
* Stuff
Blocking on:
* IRL
.EOR
tcurtis What I did: 20:04
* Not much Parrot-wise.
* Tried unsuccessfully to switch IO.readall to use a StringBuilder.
* Blog post shortly.
* Relaxed a bit after GSoC.
What I will do: 20:05
* Upload GSoC tarball to googlecode.
* Hopefully work on tree-optimization some.
EOR
20:05 bluescreen left 20:06 whiteknight left 20:07 atrodo joined 20:14 smash joined 20:17 bluescreen joined 20:30 masak joined
cotto_work helol 20:30
mikehh hi there
tcurtis Hi.
dukeleto hola hola hola (oatmeal and granola)
nwellnhof hi everyone
cotto_work dukeleto: can you lead this thing? 20:31
dukeleto cotto_work: yeah 20:32
Welcome to the weekly Parrot Dev meeting. Please buckle up.
mikehh click click 20:33
dukeleto so how did we do on our goals for the week? I am pretty sure we closed more than 13 tickets
cotto_work That was a good goal. We should try it again.
mikehh yes - gotta a couple to close still
nwellnhof But this time *everyone* closes 13 tickets ;) 20:34
20:34 NotFound joined
particle you first. 20:35
nwellnhof just kidding
NotFound Sorry, will report later.
dukeleto anybody have any comments about what we accomplished in the last week? critiques or condemnations?
particle dafrito's cla arrived.
cotto_work good news there 20:36
particle have we "accomplished" the vote for the parrot board?
dukeleto voting is open until the end of the month
mikehh I think smash was working on that
particle oh, right, i forgot it's another week still. 20:37
vote early, vote often!
cotto_work Should there be something on parrot.org about that?
particle seems like a newsworthy item, indeed. 20:38
would you mind contacting smash about that? 20:39
cotto_work np
dukeleto will tell @parrotvm/!parrot about the board election as well 20:41
let's talk about which goals we want for this week 20:42
mikehh maybe to the list as well
dukeleto anything critical that we should all know about?
nwellnhof i have some plans for the string subsystem 20:43
i think the current distinction between charsets and encodings doesn't make sense 20:44
mikehh the g++ build is broke at the moment, dunno how critical that is
nwellnhof charsets and encodings should be merged IMO 20:45
dukeleto nwellnhof: that sounds like you should make an RFC ticket and email parrot-dev about it
mikehh: that is critical as well
mikehh darbelo has been working on that
nwellnhof dukeleto: ok
mikehh charsets and encodings that is 20:46
NotFound mikehh: I'll take a look at the c++ as soon as possible
mikehh NotFound: tried fixing it but got more broken
dukeleto looks like we closed 14 tickets this past week, according to trac.parrot.org/parrot/query?status...changetime 20:47
nwellnhof didn't plobsing fix the g++ build 65 minutes ago?
dukeleto i think trac.parrot.org/parrot/ticket/1746 needs more eyes as well. Segfaults from Rakudo means a deep bug. 20:48
plobsing nwellnhof: I thought I did
mikehh might have done - haven't tested yet
dukeleto Shall we make our goal for the week to close 20 tickets? Is that reasonable? 20:49
cotto_work +1
dukeleto we have 676 open tickets
tcurtis +1
nwellnhof but not all of the open tickets are bugs 20:50
NotFound r48638 builds for me
dukeleto nwellnhof: yes, you are correct. And not all bugs have tickets :)
closing 20 tickets would mean closing about 3% of our tickets in one week, which is a nice goal 20:51
do we want to make a stipulation that a few important TT's should be included in that 20, or make it any 20?
nwellnhof we have 305 open bug tickets
149 todo 20:52
cotto_work dukeleto: narrowing down the scope might make it easier to pick a ticket and run with it
dukeleto cotto_work: sure. how do you want to narrow the scope?
NotFound Some todo should be rfc
cotto_work no useful idea
NotFound But usually no one comments on rft, so it doesn't matter %-) 20:53
s/rft/rfc
dukeleto I think we should concentrate on closing open bug tickets, but I think lowering the total number of open TT's is a good thing. 676 open tickets is ... a lot.
NotFound: i hear ya. Maybe we need a better-defined RFC process. 20:54
NotFound dukeleto: we can create an RFC to discuss that X-)
dukeleto nwellnhof: so how about "Close 20 TT's this week, concentrating on open bug tickets" ? Sound reasonable?
ERECURSIVERFCPROCESS 20:55
nwellnhof +1
dukeleto ok, we have a goal for the week.
cotto_work FWIW, I'd be glad to help anyone who wants to help hack on the github trac plugin set up a test environment.
dukeleto cotto_work: i can help on the front where the trac plugin needs to talk to the github mirror
So, i think we are now at the quiz-show portion of the meeting. Questions? 20:56
mikehh anything happened with smolder? 20:57
dukeleto i haven't heard anything. last i heard, particle was talkign to OSU folks
tcurtis++ # blog posts much appreciated 20:58
The smolder situation is ... suboptimal. 20:59
nwellnhof what's the deal with smolder? 21:00
dukeleto our current smolder server is down. this means that "make smoke" has nowhere to post test reports 21:01
this makes people sad
nwellnhof i see
dukeleto we may need to look into other options for receiving test reports
21:01 whiteknight joined
dukeleto wouldn't it be nice if we can something piggyback onto CPANtesters or something? 21:02
s/something/somehow/
having to run our own server to accept test results requires maintenance
mikehh smoke/smolder allows us to look at problems on other platforms than your own 21:05
dukeleto I think we should bring this issue to parrot-dev and continue with the meeting, though.
Any other questions?
mikehh taptinder is of use there but it is x86 - 32 and 64 bit only 21:06
dukeleto taptinder is saving our feathers right now. Without taptinder, we would be running blindly. 21:07
mikehh sure, but we need test results from other platforms, including sparc, arm and power
dukeleto I think we should talk about git and github a bit. As a community, which features of github do we want to use? 21:08
mikehh: agreed. Everyone wants more tests on more platforms.
nwellnhof does anyone regurlarly test on windows? 21:09
i think there are a lot of windows related tickets
mikehh taptinder does
tt.taptinder.org/ 21:10
nwellnhof trac.parrot.org/parrot/search?ticke...;q=windows
mikehh it runs builds/tests whenever there is a new revision 21:11
dukeleto anything else that needs to be talked about in #ps, or shall we go back to #parrot ? 21:12
mikehh we were going to look at git and github?
dukeleto I think #ps is the right place to talk about what we want to accomplish this week regarding converting to git, but this is not the place for extensive discussions. Those are for #parrot and parrot-dev 21:13
cotto_work: can you give a quick status report on your github trac plugin and what we can expect in the next week? 21:14
cotto_work: does it talk to the parrot github mirror ?
I think we should close this #ps and take other questions to #parrot and parrot-dev. Good luck to everyone in our quest to close 20 tickets! 21:19
dukeleto goes back to #parrot
21:20 nwellnhof left 21:27 masak left 21:33 atrodo left 21:35 bluescreen left 21:57 dduncan joined 22:04 dduncan left 22:17 mikehh left 22:20 plobsing left 22:23 kid51 joined 23:21 smash left