Perl 6 - the future is here, just unevenly distributed

IRC log for #parrotsketch, 2012-09-04

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

All times shown according to UTC.

Time Nick Message
00:46 benabik joined #parrotsketch
07:52 lucian joined #parrotsketch
13:32 bluescreen joined #parrotsketch
13:55 benabik joined #parrotsketch
14:04 benabik_ joined #parrotsketch
15:07 benabik joined #parrotsketch
15:08 benabik_ joined #parrotsketch
15:36 benabik joined #parrotsketch
15:43 wagle joined #parrotsketch
15:44 cotto joined #parrotsketch
15:45 contingencyplan joined #parrotsketch
15:45 rurban joined #parrotsketch
15:47 sorear joined #parrotsketch
15:47 slavorg joined #parrotsketch
15:47 Tene joined #parrotsketch
16:25 bluescreen joined #parrotsketch
16:34 bluescreen joined #parrotsketch
16:42 bluescreen joined #parrotsketch
17:20 benabik joined #parrotsketch
17:57 benabik joined #parrotsketch
18:32 rurban now or in one hour?
18:39 lucian joined #parrotsketch
19:29 lucian joined #parrotsketch
19:36 tadzik joined #parrotsketch
19:37 rurban I'll wait for whiteknight
19:56 rurban no ps today?
19:57 cotto no whiteknight, I guess
20:00 rurban I've got 30 lines to post
20:01 cotto go for it
20:01 rurban DONE:
20:01 rurban * merged several minor branches and fixed tickets.
20:01 rurban * bumped PBC_COMPAT to 12.1 for the new pmc/os, and need 13.0 for
20:01 rurban the new PBC format in native_pbc.
20:01 rurban * had a lot of parrot performance discussions with the rakudo folks.
20:01 rurban parrot performance is the only problem, not bugs, not features.
20:01 rurban TODO:
20:01 rurban * replace nqp-rx with nqp-p6 (started as git submodule)
20:01 rurban * started working on missing or wrong number converters in native_pbc,
20:01 rurban now that the tests are stable.
20:01 rurban * started looking into replacing parrot with a llvm/vmkit based
20:01 rurban vm with natural types and calling convention for nqp. esp. no
20:01 rurban continuations and a cdecl/fastcall ABI. libraries from java or .NET.
20:01 rurban * sorear hates the lorito design. why should this be faster at all?
20:01 rurban easier jittable yes, but the calling convention still sucks and jitting
20:01 rurban this mess would make it only slower, not faster.
20:01 rurban * lack of management: who made all these crazy parrot decisions?
20:02 rurban My summary is: "premature pessimation is evil"
20:02 rurban e.g. pbc broken by design on purpose.
20:02 rurban nqp switched to a different ffi, because libffi did not work on windows
20:02 rurban for them. I do not understand why this could have happened. libffi works
20:02 rurban better than dyncall on windows for me.
20:02 rurban * performance:
20:02 rurban * esp. our calling convention looks incredibly stupid and slow, methods go via
20:02 rurban Parrot_ext_call really? Parrot_ext_call as the only generic calling function
20:02 rurban is a joke also. Parse signature strings at run-time. Every simple ffi does better.
20:02 rurban * create a new object for each function/method call? not long ago even two.
20:02 rurban BLOCKS:
20:02 rurban * threads look a bit racy here and there, but adougherty fixed the
20:02 rurban biggest blocker, the sleep deadlock. looks mergable.
20:02 rurban * native_pbc is ready and needs votes to get merged. replaced
20:02 rurban the parrot version triple for opslib with the PBC version.
20:02 rurban * without parrot performance soon rakudo will switch to something else.
20:02 rurban me also.
20:03 benabik We should consider moving #ps.  Judging by attendance, this isn't a great time for people anymore.  I know I'm in class.
20:09 Coke we also just don't have many developers anymore. rurban and whiteknight, now that GSOC is over.
20:10 Coke I'm happy to kibbitz, but I'm not contributing code at this point.
20:10 benabik I'm sticking around, although my output will be lessened.
20:12 Coke regarding your question about who made these decision decisions -- it's been 11 years, and a lot of groups have come and gone.
20:16 rurban it was more rhetorical
20:17 rurban but someone should overlook the efforts - whiteknight seems to be the pm - and talk to the rakudo folks.
20:17 rurban their lorito and calling convention opinions
20:18 rurban but our threads look really good and fast.
20:18 pmichaud joined #parrotsketch
20:18 benabik Doesn't rakudo ignore our calling conventions for the most part?
20:18 rurban I just have to bing the default number to the run-time num of cpus.
20:19 rurban sure, because ours is too stupid
20:20 rurban I just have to bind the default number of threads to the run-time num of cpus
20:20 rurban The GC also looks good.
20:21 Coke our GC has issues, to be sure.
20:22 rurban Only with --leak-test on destruction.
20:22 rurban nci and threads seem to work fine.
20:23 Coke no, with the amount of memory allocated depending on what parrot thinks is available, the inability to say "only use this much memory and no more.."
20:23 sorear o/ pmichaud
20:24 Coke (the actual collection itself I know of no issues with)
20:47 pmichaud I think Coke++ is referring to https://github.com/parrot/parrot/issues/795, perhaps among other things
21:11 rurban Oh, no libsigsegv, and no malloc-fail check
21:31 rurban Fixed now all pbc number converters
21:32 tadzik left #parrotsketch
21:33 pmichaud left #parrotsketch
21:43 benabik joined #parrotsketch
21:48 benabik joined #parrotsketch
21:48 lucian joined #parrotsketch
23:06 whiteknight joined #parrotsketch

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