Priorities for this week: gsoc_past_optimizations, generational_gc, packfile, Google Code-In | Don't forget to post closed tickets in your report. | Note: This channel is only for our Tuesday status meetings (at 20:30 UTC); you probably want #parrot instead. | irclog: irc.pugscode.org/
Set by moderator on 23 November 2010.
00:58 kid51_at_dinner joined 01:58 bluescreen left 02:15 bluescreen joined 02:22 bacek left 02:32 kid51_at_dinner is now known as kid51 02:57 bluescreen left
kid51 kid51's report 03:45
WORKED ON:
* trac.parrot.org/parrot/ticket/532 "headerizer refactor". tools/dev/headerizer.pl has been phalanxed. Down to approximately 20 untested statements/branches/conditions. See tt532_headerizer_refactor branch.
* trac.parrot.org/parrot/ticket/1862 "gcc -fvisibility=hidden". CLOSED.
* trac.parrot.org/parrot/ticket/1861 "ipv6 config probe too noisy". CLOSED. 03:46
WILL DO
* Will continue to work on TT #532 and then TT #855 "profiling options"
REQUESTS
* Still need feedback on trac.parrot.org/parrot/ticket/1840 "config step to detect RTEMS" (kid51 nudges dukeleto) 03:47
* trac.parrot.org/parrot/ticket/1844: git instructions for README: One last bit of feedback needed.
EOR
03:52 whiteknight left 04:28 kid51 left 06:49 contingencyplan left 08:13 bacek joined 09:08 bacek left 09:52 bacek joined 13:45 contingencyplan joined 15:33 mikehh joined 17:18 NotFound joined 17:23 Emanuel joined 17:30 atrodo joined 18:11 whiteknight joined
whiteknight WHAT I DID: 18:11
* Working with GCI students creating--and closing--lots of tasks.
* Digging through the IOTasklist on the Trac wiki, turning some easy things into GCI tasks and trying to motivate other devs into looking at it too
* embed_api branch is building, thanks in no small part to bluescreen++. I still haven't sorted out problems with error message reporting. Have some blogging to do about this.
WHAT I WILL DO:
* Attend PDS!
* Try to resolve error message issues in the embed_api branch ASAP. Will alert the list and all about it when I am done. Will prepare to talk about it at PDS
* Keep up with GCI.
WHAT I AM BLOCKING ON:
* Time
19:06 Emanuel left
NotFound What I did: 19:43
-parrot
* Fixed a bug in library loading that was causing random segfaults
* Moved find_codepoint from experimental.ops to string.ops
* Fix and un todo/skip a few tests.
* Developing and testing the MappedByteArray PMC in a branch.
-winxed
* Added options -I and -L to the installable compiler
* Allowed multiple declaraions in int, string and float statements.
* Several minor refactors and improvements.
What I will do:
* Try to merge to master the mappedbytearray branch.
EOR
cotto_work *did: 19:45
- approved/reviewed some gci tasks (fbrito++)
- wrote up initial core inclusion guidelines
- feedback and additions are welcome
- read some papers
- PDS
- more gci review/approval
- grok and critique bacek's proposed gc algorithm
- ramp up on Lorito, figure out knowns, unknowns and questions
- will have a face-to-face Lorito brain dump on the 9th with allison, chromatic and possibly dukeleto
- if you have any burning questions, ask me or put them on LoritoDesignQuestions
*blockers:
- none
*eor
some truncations: I also started to review bacek's gc code. The TODO begins with "read some more papers" 19:52
19:59 bacek left
Util No tuits last week. Work on pbc_to_exe this week. 20:02
curl -s 'trac.parrot.org/parrot/timeline?day...ticket=on' | perl -wlne '/<em[^>]+\\(([^"]+)\\)">/ or next; $h{$1}++; END {printf "%7d\\t%s\\n", $h{$_}, $_ for sort keys %h}'
2 closed: done
9 closed: fixed
4 closed: invalid
17 closed: wontfix
1 closed: worksforme
13 new
2 reopened
EOR
20:11 dukeleto joined 20:17 plobsing joined
plobsing What I Did: 20:19
- cut PIRC out into its own project
What I Plan:
- increase the awesome (see plobsingtasklist on wiki for probable targets)
EOR
20:26 kid51 joined 20:29 plobsing left
dukeleto What I did: 20:29
* Got a TPF grant to work on embed system tests and docs:
news.perlfoundation.org/2010/11/201...sults.html
* Organized Parrot's involvement in Google Code-In. We have students submitting really cool stuff,
many pull requests have been merged.
* Just attended a great talk about Rubinius and learned a lot about how Parrot needs to use LLVM
* Hacked on the tt1798_ipv6 branch, which adds IPv6 support to Parrot.
What I will do:
* Write "how to merge pull requests with signed-off-by lines" and "how to submit a patch"
sections for our git_workflow.pod
Blocking on:
* Too much cool stuff to hack on.
.EOR
I might miss some of #ps today.
whiteknight Hello 20:31
NotFound Hola
Util Hello 20:32
whiteknight tell me we have more than 3 people today?
dukeleto hola 20:33
NotFound After thansgiving effects?
dukeleto cotto: you here? 20:34
cotto_work yes
whiteknight who's driving this train?
dukeleto whiteknight: you are.
whiteknight okay, let's get started
how did we do last week?
20:34 tcurtis joined
whiteknight I didn't see any movement on gsoc_past_optimizations 20:34
dukeleto what happened with gsoc_past_optimizations ? Did it get merged?
whiteknight no, definitely not merged 20:35
tcurtis failed to get around to working on it over Thanksgiving.
Sorry about that.
whiteknight I was looking at generational_gc branch notes too, but failed miserably
I think this was a bad week with the holiday
We have been kicking ass with Google Code-In though 20:36
cotto_work a note about the gen_gc branch: the comments are for a proposed algorithm that hasn't been implemented yet
kid51 Hmm, my impression was that we had a lot more activity than in previous weeks -- but not on the first 3 priorities listed in topic.
whiteknight 23 tasks completed already, most very high quality code/doc submissions
kid51: astute observation
cotto_work My understanding is that the algorithm needs review and eyeballs and that bacek doesn't see much future for the current code in the branch.
kid51 plobsing (I think) moved PIRC out of main repository -- which eliminated 20+ Trac tickets
whiteknight cotto_work: yes, that's my understanding too 20:37
kid51: yes, but he opened them all as tickets on github
kid51 We may actually get < 600 tickets for the first time since I've been involved with Parrot
whiteknight I'm hoping we can get down to 550 soon, with a concerted effort 20:38
cotto_work sounds like a weekly goal
dukeleto meh 20:40
20:40 bacek joined
dukeleto I am not motivated by "close X tickets", but maybe others are 20:40
I prefer a more focused goal, but that is just me. 20:41
kid51 I would welcome a comprehensive summary of the status of our Garbage Collection efforts in advance of this Sunday's PDS. 20:42
whiteknight kid51: yes, we are working on that
kid51 (I'd also welcome "Garbage Collection for Dummies" -- but that's a different request ;-) ) 20:43
whiteknight Okay, I think that's enough for current status. What goals do we want to focus on for the coming week?
kid51 Preparing for PDS and having a good one.
cotto_work PDS prep for one
20:43 PacoLinux joined
whiteknight definitely PDS prep 20:44
kid51 Also: Do we need to do anything more to keep up with google code-in than we did this past week?
whiteknight and GCI tasks. That program is not showing any signs of slowing down 20:45
kid51 has not gotten any assignments, per se
whiteknight The students are eating tasks as quickly as I can make them
kid51: The way mentors are assigned is goofy, you basically have to suggest the task yourself if you want to become a mentor
coming up with task ideas is the hardest part so far
We've been through ~23 tasks in a week 20:46
kid51 Did any students bite on the Spanish translation tasks yet?
dukeleto kid51: yes. We need more GCI tasks.
We need more code tasks.
whiteknight There are some translation tasks open, I think
dukeleto These students are more talented that I thought they would be. Which is a good thing. 20:47
whiteknight Okay, PDS and GCI. Any other goals we want to attempt for this week?
20:48 nwellnhof joined
cotto_work GC algorithm review for those who feel like getting their feet wet. 20:48
whiteknight yes, I need to re-attempt that again
NotFound I'd like testing of MappedByteArray
whiteknight okay, that's a good one too. NotFound: is it mergable with testing?
NotFound I think I can merge right now, but have tested only in linux and windows. 20:49
whiteknight okay. More testing on that then
dukeleto is going to lunch 20:50
NotFound: do we have code coverage stats for MappedByteArray ?
NotFound Maybe merge is the quicker way. Any brokeness will be easily fixable with some more #if
dukeleto +1 to merge
whiteknight +1
NotFound dukeleto: I don't have coverage stats for anything.
mikehh +1
NotFound make cover looks broken
whiteknight blah. We might want to get that fixed 20:51
cotto_work Just make sure to mark it experimental until we feel like it's ready to have a supported interface.
NotFound cotto_work: yeah, that was my intention
cotto_work full steam ahead, then
mikehh make cover fixes as a goal? 20:52
whiteknight mikehh: if we have somebody who can make it happen
I don't know anything about it myself
mikehh I'll certainly have a look at it
NotFound I haven't a clue about it.
mikehh: then +1
whiteknight PDS, GCI, Make cover, and testing some branches? 20:53
that's really all we should have, we don't want to get overloaded
moderator PDS Dec 05 2300 UTC !!! | Priorities for this week: PDS, Google Code-In, MappedByteArray test&merge, fix `make cover`, GC algorithm review | Don't forget to post closed tickets in your report. | Note: This channel is only for our Tuesday status meetings (at 20:30 UTC); you probably want #parrot instead. | irclog: irc.pugscode.org/ 20:53
cotto_work What's the best way to prepare for PDS? 20:54
whiteknight 1) Send an email to parrot-dev about anything you want to see discussed
that's it
kid51 will test mappedbytearray on Darwin
whiteknight I think we want to talk about embed_api and the GC, so I am going to send summaries about those to the list soon
any other topics are welcomed
atrodo Lorito? 20:55
Util cotto_work: shadow boxing, Rocky Balboa-style
whiteknight Lorito will wait till next PDS
kid51 Fine
cotto_work Lorito won't be discussed, but I'll be blogging about it after meeting with allison and chromatic on the 9th
kid51 Good
atrodo Works for me
whiteknight okay, shall we move on to questions? 20:56
NotFound Talking about testing, did we have some standard way to create temporary files in PIR tests?
whiteknight I'm not aware of any 20:57
kid51 IIRC, we don't. I think there is a TT about that somewhere.
mikehh need the equivalent of File::Temp or something 20:58
kid51 trac.parrot.org/parrot/ticket/955
NotFound I'll take a look at it. 20:59
whiteknight Does anybody have any questions? I have several 21:01
...? 21:02
Okay, I guess I'll go first :)
user bluescreen has sent in a CLA. He's made a number of good commits to my fork and has been a big help on the embed_api branch. Vote to make him a committer? 21:03
NotFound +1 21:04
mikehh I think so, he's been around for a while
+1 21:05
Util +1
whiteknight I'm willing to mentor
cotto_work have we received his cla?
whiteknight yes 21:06
earlier this week
cotto_work +1
whiteknight okay, We'll set that up soon. Next question: 21:07
GZipHandle PMC. (TT #1580). It's experimental but many people use it and rely on it
I suggest it no longer be experimental
NotFound Just a matter of deleting it from experimental and close the ticket, ot there is some more work needed? 21:08
whiteknight no, just remove it from DEPRECATED.pod
NotFound Good
+1
Util Does it use native code, or call an external C lib?
whiteknight PMC wrapper around an external C lib 21:09
libgzip2, I think
Util +1
whiteknight GZipHandle is a dynpmc, optionally compiled only if the necessary prereqs are found during configure 21:10
Util wonders about a future API to say "this dynpmc is not available" without haveing to know the config dependencies.
NotFound Just try to load it. 21:11
Making a way to check if something can be loaded without loading it is more code to maintain and debug for no good cause. 21:12
mikehh q1q
Util NotFound: And have it return/throw something that means "disallowed/not available" instead of "sorry I can't do that Dave"?
whiteknight okay. mikehh: you can go
Util NotFound: good point
whiteknight (we're losing steam, so I'll save the rest of my questions)
mikehh essentially opengl - you have a ticket #1870 I think, but I get a lot of config failures because we do not have a parrot type longlong 21:13
NotFound I have that failures too, but it works anyway. 21:14
mikehh well it works for me in winxed
whiteknight we voted last time to maybe move the OpenGL bindings out into a separate repo. This might be a moot point 21:15
mikehh do we need an equivalent in int64 or something in parrot anyway? I think so 21:16
especially on 32 bit platforms 21:17
whiteknight that gets tricky because Parrot is C89 21:18
there isn't a long long type, and not all 32 bit compilers have a 64-bit long
msvc is a perfect example. long is 32 bits
mikehh yeah, but I have always felt we should move on to at least C99 21:19
NotFound In platforms that doesn't have it, surely there will not be libraries that use it wich need to interface.
mikehh and do workarounds for those platforms that don't have it
whiteknight mikehh: that could be a topic for serious discussion. I know msvc doesn't really support C99, so we need to consider portability 21:20
but it is a discussion that could definitely be raised. this or the following PDS might be a good forum to broach the topic
kid51 IIRC, it was only that we wanted to support Win32 that led us to remain at C89. 21:21
mikehh we have to do all sorts on fiddly things on Win32 anyway 21:22
NotFound Win32 no, Visual C++
There are sane compilers for win32 21:23
cotto_work I'd rather not lose msvc support if we can help it. 21:25
Util cotto_work: I agree (even though I heart MinGW)
whiteknight win32 does have partial compliance with C99. it does support long long for instance
NotFound Anyway, I think we can support long long for platforms that have it with conditional compilation without much trouble, at least for interfacing NCI. 21:26
whiteknight we will have to weigh it out
mikehh I seem to remember having 64 bit int types available in msvc
haven't used it for a while though
NotFound cotto_work: I can help. People that use it can help. 21:27
cotto_work Can someone volunteer to find out if it's possible to support msvc and use C99?
NotFound But without that last part, the first is pointless.,
We can't support compilers nobody uses or nobody report problems. 21:28
dukeleto is back 21:29
NotFound We can declare that we support it, but will be a lie.
whiteknight okay, I have to leave now. I think we're about at the end of the meeting anyway
dukeleto can take over if there is anything else to talk about
21:29 whiteknight left
dukeleto what does our PDS goal of this week actually mean? 21:30
kid51 It means: prepare for it and attend it
NotFound Doing it without anyone being hurt? ;)
kid51 NotFound: apt! 21:31
If you want to shorten the Topic by taking that out, that would be okay.
mikehh dukeleto: prep for the PDS, make sure we consider the necessary discussion topics
dukeleto mikehh: what does that mean? 21:32
mikehh dukeleto: set an agenda?
dukeleto kid51: what does "prepare" for the PDS mean?
kid51 I've requested that whiteknight prepare a summary of where we stand on API work/team. 21:33
I'm hoping that there be something similar re GC.
I will post to parrot-dev a proposal for scheduling our PDSes for next 12 months.
mikehh dukeleto: set up a list of topics that need discussion in the PDS 21:34
kid51 In other words, post stuff in advance of the meeting so that the meeting moves effectively.
As opposed to, "Alright, we're here; what happens now?"
kid51 has to refocus on $job 21:35
21:35 kid51 left
mikehh dukeleto: and also "be prepared" for the meetin' 21:36
dukeleto sounds good to me. Does anybody else have questions or comments? 21:37
mikehh I know I had something else I wanted to bring up, but it has slipped my mind at the moment :-{ 21:39
21:41 plobsing joined
dukeleto it's a wrap. 21:44
21:44 dukeleto left, atrodo left 21:45 plobsing left, nwellnhof left
NotFound There is no info in ParrotDeprecations for the functions recently renamed. 21:48
Wrong chanel.
21:48 NotFound left 22:08 PacoLinux left, bluescreen joined 23:22 tcurtis left 23:46 tcurtis joined 23:49 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!