Perl 6 - the future is here, just unevenly distributed

IRC log for #parrotsketch, 2008-08-26

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

All times shown according to UTC.

Time Nick Message
10:36 gaz joined #parrotsketch
11:37 kid51 joined #parrotsketch
11:37 kid51 Pre-posting my report:
11:37 kid51 * Patched several config/auto/*.pm and associated t/steps/auto*.t files
11:37 kid51 to clean up any 'test_*' files created during Configure.pl or
11:37 kid51 pre-configuration tests.
11:37 kid51 * Applied patch providing functionality to specify >=1 config steps in
11:37 kid51 tools/dev/opsrenumber.pl.  Gave RT back to Coke for further work or closing.
11:37 kid51 * Applied patch by Vyacheslav Matjukhin re codas.
11:37 kid51 (I believe that was his first patch that we applied.)
11:37 kid51 * Posted patch implementing particle's request for file-based configuration.
11:37 kid51 * Feedback welcome (but save additional feature requests for a future RT).
11:37 kid51 Will apply at end of week unless I hear otherwise.
11:38 kid51 Pre-posting my report:
11:39 kid51 joined #parrotsketch
11:40 kid51 Pre-posting my report:
14:56 moritz joined #parrotsketch
15:22 Wknight8111 joined #parrotsketch
16:04 pmichaud joined #parrotsketch
16:20 davidfetter joined #parrotsketch
16:58 cjfields joined #parrotsketch
16:59 kj joined #parrotsketch
18:04 NotFound joined #parrotsketch
18:21 jhorwitz joined #parrotsketch
18:22 cotto_work joined #parrotsketch
18:27 allison joined #parrotsketch
18:27 chromatic joined #parrotsketch
18:29 barney joined #parrotsketch
18:31 chromatic Hi all.
18:31 moritz hi
18:31 Wknight8111 'ello poppet
18:31 particle hi there
18:31 NotFound h
18:31 kj hi
18:31 barney hi
18:31 cotto_work hi
18:31 mst joined #parrotsketch
18:31 jhorwitz hola
18:31 allison hi
18:32 moritz mst++ thanks
18:32 mst right
18:32 mst that's a random collection of +os
18:32 mst also, topic is now settable without +o
18:32 mst and now, I figure you guys can work it out for yourselves
18:32 mst left #parrotsketch
18:33 chromatic Shall we begin?
18:33 Topic for #parrotsketchis now weekly meeting at 19:30 UTC, Tuesday
18:33 chromatic Alphabetical order: allison, barney, chromatic, and everyone knows how to collate from there.
18:34 Topic for #parrotsketchis now weekly meeting at 18:30 UTC, Tuesday
18:34 allison - I'm back home in Cape Town, with lots of time to work on Parrot the next few months.
18:34 allison - Thanks to a refinement in my email process, I'm also finally able to read or skim all messages on the parrot mailing list and commit list.
18:34 allison - Modified Pmc2c libraries to parse and compile MULTI declarations in .pmc files, and to emit PMC initialization code that installs the MULTIs in a global table.
18:34 allison - Worked with Nuno and Andrew breaking off tasks for them to do in the MMD branch. Debugged and added features needed for their changes.
18:34 allison - Implemented signature-object multiple dispatch, making it possible to clear away large chunks of old MMD code.
18:34 allison - Last step (what I'm working on today) is implementing signature-object based invocation from C.
18:34 allison - Created a branch for testing and developing the Cygwin patches.
18:35 allison EOR
18:35 barney nothing to report :=(
18:35 barney .eor
18:36 chromatic I'm back in town and have time again.
18:36 chromatic I'm happy to fix reproducable segfaults people send or assign me.
18:36 chromatic I plan to work on vtable initialization this week, namely removing the need to put PARROT_API on all vtable functions.
18:37 chromatic Other than that, I'd like to see our bug count trend down for the first time in a few weeks.
18:37 chromatic cotto_work?
18:37 cotto_work * alter ego is rewriting phparray tests in PIR so they're faster and so I can release some code
18:37 cotto_work * no code is being copied from the encumbered implementation, of course
18:37 cotto_work * given the huge increase in speed from pure PIR tests, rewriting core PMC tests is looking like a nice target
18:37 cotto_work * queue one question
18:38 cotto_work kthxbye
18:38 cotto_work s/target/next target/
18:38 chromatic japhb_?
18:39 chromatic jhorwitz?
18:39 jhorwitz looked at preloading modules in mod_perl6.  not as easy as i thought, so i'm still working on it.
18:39 jhorwitz experimented with the november wiki to see if it would run under ModPerl6::Registry.  did see speedups of 10-20x in my initial tests.
18:39 jhorwitz discovered that %*ENV doesn't properly update when setting a value due to issues with the Env PMC.  pmichaud thinks this is due to some confusion btw set & assign somewhere.  opened RT #58362 to track this.
18:40 jhorwitz EOR
18:40 particle kid51 pasted his report beforehand, so see the log.
18:40 chromatic kj?
18:41 kj * worked on pirc, I'm quite happy with the result so far.
18:41 kj * checked in code to allow re-entrancy testing.
18:41 kj * I'm really stuck w.r.t. bytecode generation; need to build some motivation to work on that.
18:41 kj .eor
18:41 chromatic NotFound?
18:41 NotFound More work in PDB, approaching to an usable state.
18:41 NotFound Closing some old tickets
18:41 NotFound EOR
18:41 chromatic particle?
18:42 particle ~ ongoing but minor parrot foundation/parrot.org setup work
18:42 particle ~ participated in a few design/implementation discussions on various parrot/rakudo topics this week
18:42 particle .end
18:42 chromatic Tene?
18:42 chromatic tewk?
18:43 chromatic Wknight8111?
18:43 Wknight8111 * Created a new clean GC branch for work
18:43 Wknight8111 * Separated MS collector into a separate file for consistency, waiting for thumbs-up to merge this change into trunk.
18:43 Wknight8111 * Added a lot of function-level documentation for RT#48260
18:43 Wknight8111 * Did some bumbling work on pdd27mmd/src/pmc/MultiSub.pmc.
18:43 rdice joined #parrotsketch
18:43 Wknight8111 EOR
18:44 chromatic Question time.  cotto_work?
18:44 moritz chromatic: you forgot pmichaud and me
18:44 pmichaud (I'm present.)
18:44 chromatic Oh yeah, ops.
18:44 Wknight8111 he's made his decision
18:44 chromatic moritz?
18:44 Wknight8111 :)
18:44 moritz * Usual testing + applying patches work
18:44 cotto_w0rk joined #parrotsketch
18:44 moritz * Wrote a blog series about Perl 6, got lots of positive feedback <http://perlgeek.de/blog-en/perl-5-to-6/>
18:44 moritz * Hacked on November (the Perl 6 wiki engine), hitting rakudo bugs every now and then. Grammers rock, though ;-)
18:44 moritz * Queue two questions
18:44 moritz EOR.
18:44 chromatic pmichaud?
18:44 pmichaud == Rakudo stuff
18:44 pmichaud : various method and code refactors
18:44 pmichaud : Added @*INC and %*INC
18:44 pmichaud : precompiled modules now work
18:44 pmichaud : improve "make spectest_regression" speed by 67%
18:44 pmichaud == Other
18:45 pmichaud : answering questions in RT, IRC, and other
18:45 pmichaud EOR
18:45 chromatic Did I miss anyone else?
18:45 cotto_w0rk sorry, lost connection
18:45 cotto_w0rk is a question from me ok now?
18:45 chromatic Yes, go ahead.
18:46 cotto_w0rk custom structs don't seem to work as ATTRs in pd17 PMCs
18:46 cotto_w0rk is this something I'm doing incorrectly or do they need to be fixed?
18:47 allison as in 'ATTR struct foo *bar' ?
18:47 chromatic The Pmc2c modules should parse those fine, at least if you've #typedef'd them.
18:47 allison or some equivalent declaration that includes the word 'struct'?
18:48 allison it'll parse anything of the form 'ATTR <type> *<name>' or 'ATTR <type> <name>'
18:48 cotto_w3rk joined #parrotsketch
18:48 allison if you have to use the word 'struct' then you haven't properly named the struct
18:49 cotto_work ok
18:49 cotto_work thanks
18:50 chromatic moritz, you had two questions.
18:50 moritz 1) It seems usuful to me to give Reini Urban a commit bit. Has this been discussed already? Any comments?
18:50 chromatic We're waiting for his signed CLA.
18:50 moritz ok
18:50 moritz that's on his way
18:50 moritz 2) Rakudo needs an rindex function. Do we want that in parrot? Or rather in Rakudo directly?
18:50 moritz (might be common with other languages)
18:50 pmichaud we can certainly implement rindex without having to have a rindex opcode
18:51 Wknight8111 what is rindex?
18:51 pmichaud rindex == find last occurrence of a substring
18:51 Wknight8111 gotcha
18:52 allison doesn't sound like a terrible addition to parrot
18:52 barney also for strrpos  in PHP
18:52 allison though, can we make it some kind of option to index instead?
18:52 chromatic Wouldn't that make index a variadic opcode?
18:53 allison say, pass an integer to index as a position to start at, and if the integer is negative, work backwards?
18:53 pmichaud personally I think as a method on String might be better
18:53 allison well, all opcodes allow multiple signatures
18:53 allison yes, a method on String sounds quite sensible
18:54 moritz so index stays an opcode, rindex as a method?
18:54 pmichaud index is used very frequently -- PGE makes heavy use of it
18:54 allison or 'index' as a method with multiple options
18:54 pmichaud rindex is used very rarely
18:55 allison the starting position could be generally useful (find the first occurence of X char after Y position)
18:55 * moritz doesn't care either way
18:56 pmichaud we already have index taking a starting position
18:56 chromatic Implementing the rindex method in terms of a partially applied index opcode seems sane.
18:56 allison for now, index stays as an opcode
18:56 pmichaud yes, index needs to be an opcode (see comment about heavy usage above)
18:57 allison pmichaud: (starting position) ah, excellent, so we can simply extend it to allow negative starting positions
18:57 pmichaud allison: I think it might already allow negative starting positions
18:57 NotFound IMO rindex is a differente beasts, becauseof variable length encoding issues.
18:58 allison NotFound: so index shouldn't handle variable length encodings?
18:58 allison sounds odd
18:58 NotFound allison: it must, but is easier than doing it backwards.
18:59 allison after we finish implementing the strings pdd, it'll just convert to NFGs, and go backwards
18:59 allison we shouldn't have two implementations of index, one that's multiple encoding safe and one that isn't
18:59 pmichaud (fwiw, after checking source, the index op always returns -1 on negative indexes)
18:59 pmichaud er, negative starting locations
19:00 allison pmichaud: so, it's spec'd but broken.
19:00 moritz src/ops/string.ops doesn't talk about negative starting positions
19:00 allison okay, step 1) fix index so it handles negative indexes as an rindex
19:00 pmichaud yes.  I'd be opposed to having negative start position as rindex
19:00 pmichaud I think it should just start index from end of string, the way some other opcodes handle negative start locations
19:00 allison step 2) possibly add an 'index' method to String that accepts more options
19:01 allison pmichaud: if it starts the index from the end of the string, does it loop around to the start? or does it just never find anything
19:02 allison searching in reverse seems more useful than either option
19:02 allison ah, unless you're starting at, say -10, and working forward
19:02 pmichaud yes.
19:02 pmichaud that's what I man.
19:02 pmichaud *mean.
19:02 allison fair enough
19:03 pmichaud rindex is a different operation from index in that sense
19:03 allison then index as opcode, and rindex as method
19:03 moritz ok
19:03 NotFound We can also add index as method, for completeness.
19:03 allison but let's call it 'reverse_index'
19:03 moritz eoq for me
19:03 chromatic Did you have another question, moritz?
19:03 moritz chromatic: no
19:04 chromatic Are there any other questions?
19:04 allison Are the parrot mailing list and commit list delayed by 12+ hours for anyone else? Methinks something is screwy with my perl.org email alias.
19:04 Wknight8111 it's the samefor me
19:04 chromatic RT seems slow to me, but many of my mails to the list come back quickish.
19:05 allison certainly makes email conversations difficult
19:05 pmichaud it's variable for me.  sometimes mails come back very quick.  sometimes it's many hours.
19:05 Wknight8111 (many hours)--
19:05 NotFound messages posted in RT delays a few hours to reach the list for me.
19:05 allison okay then, probably has to do with spam backlogs on the perl.org servers
19:05 moritz (same as pmichaud for me)
19:06 allison yeah, I'll get them delayed by 18 hours one day, then all flooding in at once the next day
19:06 Wknight8111 on the bright side, very rarely see any spam
19:06 allison no worries, we're about to switch to a new mailing list, just wanted to know if it was a general thing
19:07 allison EOQ
19:08 chromatic If there's nothing else, let's call it a week.
19:08 allison Thanks c!
19:09 pmichaud left #parrotsketch
19:10 chromatic left #parrotsketch
19:10 Wknight8111 left #parrotsketch
19:10 NotFound left #parrotsketch
19:11 moritz is anybody staying in this channel until next week?
19:16 particle moritz: as long as my connection is up, yes
19:17 moritz more random ops ;)
19:20 cotto_work left #parrotsketch
19:20 kj left #parrotsketch
19:29 allison left #parrotsketch
19:29 jhorwitz left #parrotsketch
19:36 rurban joined #parrotsketch
20:55 cjfields left #parrotsketch
21:09 rurban left #parrotsketch

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