Perl 6 - the future is here, just unevenly distributed

IRC log for #moarvm, 2015-11-05

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

All times shown according to UTC.

Time Nick Message
01:43 tokuhiro_ joined #moarvm
02:05 lizmat joined #moarvm
02:17 tokuhiro_ joined #moarvm
02:47 ilbot3 joined #moarvm
02:47 Topic for #moarvm is now https://github.com/moarvm/moarvm | IRC logs at  http://irclog.perlgeek.de/moarvm/today
05:19 vendethiel joined #moarvm
06:33 JimmyZ Unhandled exception: Bytecode stream corrupt (missing magic string) # ^^
06:34 JimmyZ no, it is 'set encoding requires an object with REPR MVMOSHandle'
06:38 JimmyZ flussence: looks like a  rakudo bug
06:39 domidumont joined #moarvm
06:44 domidumont joined #moarvm
07:04 domidumont joined #moarvm
07:19 domidumont joined #moarvm
07:19 vendethiel joined #moarvm
08:28 leont joined #moarvm
08:39 zakharyas joined #moarvm
11:02 * stmuk wonders why we have new deprecations in a beta peroid
11:24 jnthn nwc10: Thanks for the ASAN :)
11:48 dalek MoarVM: 1e4912d | jnthn++ | src/strings/normalize.c:
11:48 dalek MoarVM: Remvoe bogus tweaking of end of normalized chars.
11:48 dalek MoarVM:
11:48 dalek MoarVM: We're in the middle of normalizing at this point, and update this
11:48 dalek MoarVM: afterwards. And we can now be sure we're never re-normalizing chars
11:48 dalek MoarVM: since the recent patch fixing that.
11:48 dalek MoarVM: review: https://github.com/MoarVM/MoarVM/commit/1e4912d42e
13:53 JimmyZ joined #moarvm
14:44 leedo joined #moarvm
14:44 ashleydev joined #moarvm
14:54 _longines joined #moarvm
15:24 nwc10 jnthn: ASAN is tolerating you again.
15:25 nwc10 (although if you're not careful it will bring you another dead mouse to try to encourage you to learn to be a better hunter)
15:28 dalek MoarVM: f601e24 | jnthn++ | src/ (4 files):
15:28 dalek MoarVM: Be sure to chomp separators at eof.
15:28 dalek MoarVM: review: https://github.com/MoarVM/MoarVM/commit/f601e242ef
15:28 jnthn Phew :)
15:29 jnthn Latest one should be benign. :)
15:29 jnthn .oO( famous last words )
15:59 jnthn Bah, managed to get a runaway moar process that allocated 20 gigs before I could stop it
16:00 jnthn Oddly enough, 'cus of a bug in handling the BOM
16:00 [Coke] the BOOM?
16:01 dalek MoarVM: 9fd3005 | jnthn++ | src/strings/utf8.c:
16:01 dalek MoarVM: After eating BOM, mark we indeed used the bytes.
16:01 dalek MoarVM:
16:01 dalek MoarVM: Otherwise, we won't correct mark eof.
16:01 dalek MoarVM: review: https://github.com/MoarVM/MoarVM/commit/9fd3005a51
16:01 jnthn Somebody set up us the BOM...
16:01 nwc10 how *fast* is it at allocating 20 Gigs? Is our performance dangerously good, or safely pathetic?
16:01 jnthn I had to reboot.
16:01 jnthn The first time
16:01 jnthn 'cus I wasn't watching
16:01 jnthn Second time I was, to see which test did it
16:02 jnthn And it hit 11GB in a second or two
16:02 nwc10 is that a live test that I'm currently running that will go boom?
16:02 jnthn Then hit the swap
16:02 jnthn Uh...yes
16:02 nwc10 OK. I'll kill that
16:02 jnthn Oh wait
16:02 jnthn No
16:02 jnthn You're safe
16:03 jnthn It's a Moar bug, but an unpushed Rakudo patch that exposed it
16:03 nwc10 have already killed. will rebuild with f601e242ef
16:52 hoelzro_ joined #moarvm
16:52 JimmyZ_ joined #moarvm
16:53 lue joined #moarvm
16:54 arnsholt_ joined #moarvm
16:57 stmuk_ joined #moarvm
17:07 arnsholt_ joined #moarvm
17:52 domidumont joined #moarvm
18:24 leont joined #moarvm
18:33 FROGGS joined #moarvm
18:34 zakharyas joined #moarvm
18:46 arnsholt joined #moarvm
20:20 timotimo nwc10: i imagine that 11GB would have been allocated from a somewhat tight loop in C-land, so ... performance wise that doesn't say much about moar, i expect
21:18 flussence joined #moarvm
21:36 timotimo jnthn: do you have a little roadmap what the NFA generator needs to learn to make "\r\n" work properly?
21:37 timotimo i expect since it works on a Str, it'll want to know about the special grapheme that \r\n turns into?
21:37 timotimo and that that is supposed to match \r, \n and \r\n?
21:40 timotimo i haven't looked into the API we expose to take apart graphemes yet
21:48 cygx joined #moarvm
21:48 cygx is there any way to dump the serialized objects from a bytecode file?
21:50 cygx context: a simple `sub foo is native {*}` compiles to about 1m of bytecode
23:55 timotimo cygx: i had a script that does that!
23:56 timotimo er ... well, it grabbed all serialized object from a given SC, i think i gave the SC id in the code via a string?
23:56 timotimo anyway, let me dig it up for you
23:57 timotimo oof, so many unversioned junk files in all my repositories
23:58 timotimo oh, huh
23:59 timotimo apparently there's a patch to moarvm that'd print out the scid and index and such into a logfile and then this script i had prints out what those things are

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