Perl 6 - the future is here, just unevenly distributed

IRC log for #parrotsketch, 2012-05-08

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

All times shown according to UTC.

Time Nick Message
07:50 contingencyplan joined #parrotsketch
08:30 lucian joined #parrotsketch
09:29 schm00ster joined #parrotsketch
12:53 alvis joined #parrotsketch
14:01 jashwanth joined #parrotsketch
15:20 crab2313 joined #parrotsketch
15:45 jashwanth joined #parrotsketch
16:16 contingencyplan joined #parrotsketch
18:08 lucian joined #parrotsketch
19:01 tadzik joined #parrotsketch
19:03 Coke Will not be able to make #ps today, probably: covered coke/rm_pasm on parrot-dev mailing list. feedback, more PASM removal commits welcome.
19:03 Coke EOR.
19:03 whiteknight joined #parrotsketch
19:04 benabik joined #parrotsketch
19:07 NotFound joined #parrotsketch
19:09 alvis What I did:
19:09 alvis * NOTE: Primarily, the same report as last week, except
19:09 alvis - Almost complete with 'ayardley/READMEs' - a local branch to update Parrot's 'README' files to 'README.pod' files and to add 'README.pod' files to the top-level directories.
19:10 alvis EOR
19:11 benabik DID:
19:11 benabik * More work on pct_winxed
19:11 benabik * submitted many issues to Winxed (sorry NotFound)
19:11 benabik * Put together a gist on the possible PACT assembly language: https://gist.github.com/2599989
19:11 benabik WILL DO:
19:11 benabik * Probably not much, maybe a bit more in pct_winxed
19:11 benabik BLOCKERS:
19:11 benabik * Want to try to pile through as much of my thesis as possible in the next couple weeks before GSoC starts
19:11 benabik EOR
19:12 dukeleto joined #parrotsketch
19:12 moritz DID: some rm_pasm work. Lots of Rakudo hacking. EOR
19:12 * dukeleto bought a house and is getting married on Saturday. Just about nothing parrot-related was accomplished.
19:13 moritz dukeleto++ # double congrtulations
19:15 dukeleto moritz: thanks! It has been a lessen in organized chaos :)
19:16 whiteknight WHAT I DID:
19:16 whiteknight * Digging through the issue tracker. We have almost 500 open tickets, we need to get that number much lower. I'm fixing/closing them as I go. We have over 100 branches, many of those are old and may be deletable.
19:16 whiteknight * Started a new branch to remove VTABLE_can (but not the "can" op). Going surprisingly smoothly. May be ready to merge shortly after 4.4.0, with testing.
19:16 whiteknight * Started a branch to remove singleton PMCs. This branch is not going smoothly and, because of the pointer-based interface for singletons, is segfaulty.
19:16 whiteknight * Fixed pbc_merge to properly deduplicate strings, merge annotation segments, and properly merge debug segments.
19:16 whiteknight * Fixed pbc_disassemble to do something reasonable when passed no arguments.
19:16 whiteknight * Fixed LexPad.get_iter to do the correct thing instead of returning the iterator for the corresponding LexInfo.
19:16 whiteknight * Fixed parrot_debugger to not segfault or do other stupid things when we pass in something that isn't a .pbc. Now, it expects all input files to be .pbc files.
19:16 whiteknight * Several cleanups to Key PMC and src/key.c, add some introspection methods to Key PMC for working with register-reference keys.
19:16 whiteknight * Started a branch to enable GC finalization globally. Testing seems to be going well so far, but I don't think anybody has built/tested Rakudo with it yet.
19:16 whiteknight * Helped rurban++ fix several memory corruption errors he was finding using the clang AddressSanatizer tool. I'm not able to run AddressSanatizer myself because of various config-related issues I'm having
19:16 whiteknight * Rosella.Net is now tested and considered stable.
19:16 whiteknight * A large assortment of other small cleanups and fixes as I dig through the ticket queue.
19:16 whiteknight * Following along with Coke++'s efforts to remove PASM. I have a patch to rip PASM out of the IMCC interface routines and rip out the PASM compreg PMC. This opens the door to detect and remove much pasm-related IMCC code in the future.
19:16 whiteknight WHAT I WILL DO:
19:16 whiteknight * 4.4 release is next week. I'm going to fix/cleanup/merge anything that I can do before then. Probably have a freeze on master starting on Sunday.
19:16 whiteknight EOR
19:19 alvis dukeleto++ # My congratulations as well!
19:21 NotFound What I did: fixed some of the winxed issues reported by benabik and added some new features. EOR
19:22 cotto did:
19:22 cotto * resumed hacking on M0 (made a big todo list, implemented register name aliases, chatting with jimmyz++ and nbrown++ about other things)
19:22 cotto * started working with benabik++ on his pact assembly language, reviewed pct_winxed awesomeness
19:22 cotto * flipped the crazy bit, gave my 2 weeks notice at $dayjob, will start working at a tiny startup on the 22nd.
19:23 cotto will do:
19:23 cotto * continue working through the m0 todo list (https://gist.github.com/2581153 - comments/forks welcome!)
19:23 cotto * gsoc stuff with benabik
19:23 cotto * -Ofun
19:23 cotto EOR
19:24 dukeleto cotto++ # congrats on the crazy bit!
19:30 cotto It's that time again.
19:30 alvis hello
19:30 NotFound Hola
19:31 benabik o/
19:32 * dukeleto waves temporarily
19:32 cotto How's this week been?
19:33 whiteknight good
19:33 whiteknight very good
19:33 cotto indeed
19:34 cotto It's great to see development pick up.
19:35 whiteknight We've got way too many open tickets in the queue. We need to start focusing on that again
19:35 whiteknight I think we can get down to 450 open tickets by the end of the month without too much stress
19:35 cotto jimmy++ was also looking through old branches.  If he asks about anything, we should make sure to give him a hand.
19:35 whiteknight yes
19:35 whiteknight We also have over 100 branches, so if we can prune the tree a little that would be great
19:36 cotto A issue bash sounds great.
19:36 whiteknight We've got a week until 4.4, and if we can trim some fat in preparation that would be great
19:36 NotFound I think we are having again "wish list" issues
19:36 cotto Should we focus on issues, branches or both?
19:36 whiteknight issues
19:36 cotto NotFound: can you elaborate?
19:37 benabik I poured through issues a while back... gave up when #185 led to #346 led to "how do I test this"?
19:37 whiteknight As I found in the last few days, we have several issues which are easily closable. Either they're already fixed but nobody updated the ticket, or the ticket is WONTFIX, or the ticket is a vague wishlist item, etc
19:38 cotto Consistently closing even a single issue per person-week is meaningful progress.
19:38 whiteknight yes it is
19:38 whiteknight A lot of old tickets we need to be able to say " I mean no disrespect, but we're not going to keep this open"
19:38 NotFound cotto: for example, "consider some extern lib for hashing"... I thought the idea was to put that kind of things in wiki pages, not in issues.
19:38 whiteknight yes.
19:39 dukeleto +1 to cleansing issues
19:39 cotto NotFound: that can go away or go to the wiki.  It doesn't really need an issue and should be very easy to implement for a mildly motivated hacker.
19:40 dukeleto respectfully close old issues that are ambiguous/wishy
19:40 cotto +1
19:40 dukeleto as long as we ask people to create a new, specific issue with updated data, I think it is fine
19:40 benabik +1 to saving useful information in wiki
19:40 dukeleto Keeping open issues which are impossible to close is a fools errand. Nobody wins.
19:41 dukeleto benabik: yes. And closed issues are, of course, still there for public archaeology/etc
19:41 benabik dukeleto: The information isn't lost, but its discoverability is.
19:41 dukeleto Having a graph of the current number of issues + branches will help us visualize progress.
19:41 dukeleto benabik: maybe. Closed issues are just as easy to search as open issues, imho
19:43 benabik dukeleto: As an open issue it's "hey, I could work on this."  Nobody cruses closed issues for things to do.  Just throwing my support behind putting those thing in the wiki.
19:44 alvis While we're on issues: I promised I'd bring up kid51's email about "7 .pmc files with undocumented functions ..." to parrot-dev; so, here it is. He would like to know how we should proceed, as well.
19:45 cotto alvis: thanks
19:46 cotto I'll take object.pmc
19:47 cotto also needed: namespace, eval, complex, callcontext, bigint, capture
19:47 alvis So, folks should pick one and work on it, then? If so, I'll take bigint.
19:48 whiteknight open a ticket, assign it to me. Will be done tonight
19:48 cotto That's what I failed to explicitly suggest, yes
19:48 alvis cotto: ok, I'll follow up on parrot-dev with the approach.
19:48 alvis whiteknight: ok. how about eval?
19:48 whiteknight quick warning: most of the comments in eval pmc will be along the lines of BLAH BLAH BLAH THIS PMC IS TEH STOOPID
19:48 cotto alvis: ok.  I'll open an issue
19:49 whiteknight alvis: that's fine
19:49 cotto whiteknight: those are the best kind
19:49 alvis cotto: ok. I'll leave it to you to open the issue.
19:52 cotto https://github.com/parrot/parrot/issues/768
19:53 lucian joined #parrotsketch
19:53 cotto and firefox crashed just after submitting the issue
19:55 cotto any other questions?
19:56 alvis that's it from me.
19:57 cotto anyone mind if I make that test fail?  Right now it just TODOs underdocumented PMCs
19:58 whiteknight go for it
19:59 cotto and done
20:05 alvis we wrapped?
20:05 cotto I think so.
20:05 alvis left #parrotsketch
20:06 moritz left #parrotsketch
20:06 crab2313 left #parrotsketch
20:23 NotFound left #parrotsketch
20:26 tadzik left #parrotsketch
21:16 benabik left #parrotsketch
22:14 whiteknight joined #parrotsketch
22:22 lucian joined #parrotsketch
23:22 contingencyplan joined #parrotsketch

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