Perl 6 - the future is here, just unevenly distributed

IRC log for #parrotsketch, 2012-06-19

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

All times shown according to UTC.

Time Nick Message
03:32 benabik joined #parrotsketch
08:57 lucian joined #parrotsketch
09:41 kjs joined #parrotsketch
09:57 kjs Reporting now, as I won't be available today and rest of week:
09:57 kjs REPORT:
09:57 kjs Much progress on M1, with help of jimmy++.
09:57 kjs HIGHLIGHTS:
09:57 kjs + recursive functions work; (fact(), fib())
09:57 kjs + structs and nested structs work (x.y.z = 42).
09:57 kjs + multi-dimensional arrays work (x[3][4][5] = 42).
09:57 kjs + combination of the above works too (x[1].y[3].z = 42);
09:57 kjs OTHER PROGRESS:
09:57 kjs + working on examples/life.m1 (doesn't work ATM).
09:57 kjs + working on examples/crc.m1 (different output from C version but prob. a 32 v. 64 bit issue).
09:57 kjs + various bug fixes; many issues are ironed out.
09:57 kjs --> making assumptions explicit with assert() works great.
09:57 kjs + fix up the type checker (WIP)
09:57 kjs --> most problems are solved by fixing the type checker; it needs to walk the whole AST to visit all nodes and set the right pointers.
09:57 kjs + added support for hex., oct. and binary contants (0b01010)
09:57 kjs + work on docs/pddxx_m1.pod to explain the various "subsystems", even though they are not modular.
09:57 kjs TODO & OPEN ISSUES:
09:57 kjs + the type checker needs much more work to emit warnings and errors on improper use.
09:57 kjs + the register allocator/optimizer needs more testing, and need a proper register spiller. whenever running out of registers, it just resets at the moment.
09:57 kjs + parameters and arguments seems to be wonky. This needs more testing and ironing out.
09:57 kjs + probably want to have callee set up its own frame; this code would
09:57 kjs only be generated per function, rather than per call; this reduces the bytecode file size.
09:57 kjs + add support for various types such as unsigned types, int32, etc.; this depends on M0 support
09:57 kjs HELP WANTED:
09:57 kjs + It would be great to increase the bus number for M1 as my availability is likely to go down over the next months. Please try it out, report bugs, ask questions, code-reviews; code walk-throughs can be arranged.
09:57 kjs .eor
13:35 wagle_ joined #parrotsketch
13:38 alvis_ joined #parrotsketch
13:50 jlaire joined #parrotsketch
13:53 PerlJam joined #parrotsketch
14:05 PerlJam joined #parrotsketch
15:00 contingencyplan joined #parrotsketch
16:14 jashwanth joined #parrotsketch
19:15 whiteknight joined #parrotsketch
19:17 alvis What I did:
19:17 alvis * almost complete on the fixup to 'tools/release/parrot_github_release.pl' (It does work as is; it's just a'bit tricky)
19:17 alvis * cut the 4.5.0 release (but still working on a few bits.)
19:17 alvis EOR
19:18 whiteknight WHAT I DID:
19:18 whiteknight * GSOC Wrangling.
19:18 whiteknight * Working on the io_cleanup1 branch. Most tests are passing. function-level docs are in place. A few small fixes are left, nothing major.
19:18 whiteknight WHAT I WILL DO:
19:18 whiteknight * Would like to merge the threads branch soon. I was waiting for nine, but he has been absent. I will make preparations
19:18 whiteknight * Continue on the io_cleanup1 branch. Would like to have it passing all tests and be merge-ready this week (even if we wait a while before doing it)
19:18 whiteknight * Continue on GSOC stuff
19:18 whiteknight * Away the end of the week and the weekend.
19:18 whiteknight EOR
19:18 darbelo joined #parrotsketch
19:27 brrt joined #parrotsketch
19:30 cotto howdy
19:30 alvis hio
19:30 whiteknight cotto: howdy
19:30 rurban hi
19:31 cotto how's this week been?
19:31 rurban ok me next
19:31 rurban 2012-07-19 rurban
19:31 rurban DONE:
19:32 rurban * @YAPC talked with Patrick and Jim for the upcoming supported release I'll be doing. Looks fine, but...
19:32 rurban * beware of threads and lorito and io_cleanup
19:32 rurban * aside: started on the perl5 type system :) e.g. my const int @array[5] = (0..4);
19:32 rurban TODO:
19:32 rurban * watch threads, io_cleanup and lorito stableness. better merge earlier to
19:32 rurban see no surprises and torn apart features
19:32 rurban * exotic platforms. I have a few, but Jim is the only one with a PPC
19:32 rurban * kjs: when do you thing you can merge M1?
19:32 rurban * how are threads doing this month? still the windows blocker around?
19:32 rurban EOR
19:32 rurban s/thing/think/
19:32 whiteknight cotto: Week has been slow in master. Looks like m0 is moving along like crazy
19:33 cotto yes and very yes
19:34 rurban kjs is not here. what is the time-frame for m0 and m1 to be merged?
19:35 NotFound joined #parrotsketch
19:35 NotFound Hello
19:36 rurban NotFound: your turn
19:36 cotto I think it's early to talk about merging.  We'll need to be a lot closer to implementing parrot guts in M1.
19:36 NotFound Nothing important to report.
19:36 whiteknight next merges on the table are threads and hopefully io_cleanup1
19:37 whiteknight the later still has some kinks to work out
19:37 whiteknight threads I would like to do this week or next week
19:37 rurban cotto: great to here. I was worried
19:37 rurban great to  hear. (oh my)
19:38 rurban do we announce threads as experimental?
19:38 cotto has it been merged?
19:39 NotFound Someone has tried declaration-after-statement in non-gcc environments?
19:39 rurban no, just preparing myself for a desaster
19:40 rurban can we exclude threads somehow from the "supported" policy?
19:40 whiteknight threads is pretty stable. We can announce it as experimental
19:40 whiteknight It might be marked that way in the branch already. I'll make sure of it
19:41 cotto thanks
19:42 rurban I am mainly worried about our supported API for threads. I'm pessimistic with the feature, but no prob for me
19:42 whiteknight rurban: the concept of "supported" is much weaker than it used to be
19:42 whiteknight we don't follow that old policy very closely anymore
19:43 rurban for the next 3 releases I know :)
19:43 cotto thanks for bringing up that concern.  We'll handle it with kid gloves.
19:43 whiteknight so long as Rakudo continues to build and run like normal, that's all the support I currently care about
19:43 rurban The rest is pretty stable anyway
19:43 rurban And if they decide to try out threads? :) just kidding
19:44 whiteknight Ideally the threads should work, and if we find bugs we will fix them
19:44 rurban Good
19:46 rurban So lorito in about 3 months or so?
19:46 NotFound We can even find and fix them at the same time.
19:46 cotto rurban, hard to say
19:46 cotto not sooner than that, at least
19:46 rurban Ok
19:47 rurban kid left us? I need him
19:48 whiteknight kid51 is still around
19:53 rurban Thanks, that's it I guess. No questions?
19:55 rurban The webpage has to be updated for the new release, btw
19:55 alvis yep, that's one of the bits left to do.
19:56 rurban PDD 13 has an odd ordering in http://docs.parrot.org/parrot/latest/html/pdds.html
19:58 alvis I'll see what I can do to fix it.
19:58 rurban So looks like I'm calling it a wrap now, not cotto.
19:59 cotto wfm
20:20 brrt joined #parrotsketch
20:25 brrt left #parrotsketch
20:47 brrt joined #parrotsketch
20:48 brrt ok, repeating my report here for the logs
20:48 brrt what I did:
20:48 brrt - reshuffle a bunch of code (move the settting up of IO handles to winxed, about which I'm really happy)
20:52 brrt - I have not written as much tests as I would have liked, mostly because I feel things are not done enough, and also because I work primarly within gdb (use make debug, works like a charm)
20:52 brrt what I'm busy with right now:
20:53 brrt - getting the request variables and headers imported into parrot, this is mostly a copying and moving stuff arround issue
20:53 brrt - getting output headers out of parrot into apache, which is tricky-ish (although #parrot just gave me good advice)
20:54 brrt What I'm going to do:
20:54 brrt - unify script-loading into its own namespace and classes
20:58 brrt - write the different loaders (CGI, PSGI, inline-based-on-hinxed)
20:59 brrt - and move different utiltiy functions together
20:59 brrt EOR
21:37 brrt left #parrotsketch
21:43 NotFound left #parrotsketch
22:22 whiteknight joined #parrotsketch
22:29 darbelo left #parrotsketch
22:53 alvis_ joined #parrotsketch
22:53 alvis_ left #parrotsketch

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