Perl 6 - the future is here, just unevenly distributed

IRC log for #moarvm, 2016-11-22

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

All times shown according to UTC.

Time Nick Message
02:29 tony-o joined #moarvm
02:29 tony-o moar(63281,0x700003cbc000) malloc: *** error for object 0x7fdd4e6765b0: pointer being freed was not allocated
02:29 tony-o *** set a breakpoint in malloc_error_break to debug
02:35 tony-o hey all.  i'm going to dig into that more but i believe it's coming from Proc::Async - if that's already a known thing please let me know otherwise i'll dig further into that
02:48 ilbot3 joined #moarvm
02:48 Topic for #moarvm is now https://github.com/moarvm/moarvm | IRC logs at  http://irclog.perlgeek.de/moarvm/today
03:38 ilmari joined #moarvm
03:40 geekosaur joined #moarvm
03:46 ilmari joined #moarvm
06:26 domidumont joined #moarvm
06:58 domidumont joined #moarvm
07:02 domidumont joined #moarvm
08:04 geekosaur joined #moarvm
08:24 zakharyas joined #moarvm
09:40 timotimo tony-o: well, that's just memory corruption. could be caused from anything at all. run it with valgrind or with a moarvm built with --asan to figure out more
09:42 timotimo or if it's very new, i could have f'ed up the code that sets strings to "do not free this" when they are set to share memory with compunits
09:42 timotimo i thought i made sure it's fine, though
09:48 geekosaur joined #moarvm
10:27 timotimo tony-o: anything i can help you with? code i can try out or something?
10:49 zakharyas joined #moarvm
12:34 brrt \o
12:34 nwc10 o/
13:10 brrt how are things? we are a release further again, I see
13:38 lizmat brrt: indeed
13:40 brrt time flies...
13:42 jnthn Indeed, couldn't believe it was our 35th release already o.O
13:43 brrt :-o
13:43 brrt almost 3 years?
13:44 jnthn Yeah
14:06 domidumont joined #moarvm
14:22 timotimo i put in my branches and directly afterwards someone complains about a non-malloced pointer getting free()'d
14:22 timotimo this is making me uneasy :)
14:23 timotimo other things making me uneasy: a tiny splinter of what i believe to be glass being stuck in my foot
14:33 brrt that sucks, yes
14:33 timotimo it doesn't seem to stick out, and i can't seem to figure out which way it went :|
14:35 brrt hhmmm
14:36 tony-o timotimo: i'm getting it with overwatch - i was in a hurry yesterday but i'm just sitting down to see if i can reproduce
14:37 timotimo i played overwatch a bit on the free weekend ... i don't think i'll get it, though ... :P
14:37 tony-o haha, how was it?
14:37 timotimo it was mildly fun
14:38 jnthn tony-o: Are you using Proc::Async with binary or char streams? My $dayjob project is tripping over something that I think I narrowed down to a long-standing phaser scoping bug :/
14:39 jnthn (The phaser in question being the LAST one inside of the whenever in the bin => chars supply thingy)
14:39 tony-o jnthn: https://github.com/tony-o/perl6-overwatch/blob/master/lib/overwatch.pm6#L106
14:40 tony-o it's using char streams
14:40 jnthn Yeah, could be that we're seeing the same kind of issue then
14:40 jnthn I'll see if I can look into it tomorrow or so
14:40 tony-o what day job do you have that is using perl6?  is it being used in production environments now?
14:41 * tony-o crawls out of his cave
14:42 jnthn A team at $dayjob will be starting on a new project from Jan which I'll very likely be architect on, and I wanted to explore the problem domain a decent bit beforehand. And there's various places stuff will want to be done concurrency.
14:42 jnthn *concurrently
14:43 jnthn So I've been working on a prototype.
14:43 jnthn Of course, prototypes have a history of escaping :P
14:43 tony-o nice
14:43 tony-o haha
14:43 jnthn Though I suspect it won't happen in this case.
14:43 jnthn Though I suspected that before and was wrong so... :P
14:44 timotimo \prototype
14:45 tony-o i should try to find a gig using perl6 rather than node
14:45 timotimo don't worry, soon we'll have perl6 code running on node, too :P
14:45 tony-o excellent, the bay area isn't huge on perl
14:59 zakharyas joined #moarvm
15:33 FROGGS joined #moarvm
15:48 brrt the bay area has a culture all of its own
15:49 FROGGS o/
15:49 yoleaux2 21 Nov 2016 22:58Z <|Tux|> FROGGS: http://hpux.connect.org.uk only ships 32bit objects :( :( - I will install libffi by hand
15:59 * [Coke] has been using perl 6 for getting-stuff-done work at $dayjob, very nice.
16:01 FROGGS did so too just lately :o)
16:18 zakharyas joined #moarvm
17:02 domidumont joined #moarvm
18:02 vendethiel joined #moarvm
18:22 zakharyas joined #moarvm
20:11 TimToady joined #moarvm
20:37 pyrimidine joined #moarvm
21:03 zakharyas joined #moarvm
21:34 [Coke] anyone have instructions on how to regen the unicode_db? trying to find how to put the files into UNIDATA/ that the tool expects.
21:34 diakopter I did it
21:34 diakopter (originally)
21:35 diakopter [Coke]: you unzip it from http://unicode.org/Public/zipped/
21:35 [Coke] trying to track down if we're including the information http://unicode.org/reports/tr14/#DescriptionOfProperties, but I think having the UNIDATA/ files would be very helpful answering that question.
21:36 [Coke] diakopter: awesome, thanks.
21:36 diakopter is the latest built from 8.0?
21:36 diakopter might as well put that link in the generation script
21:37 [Coke] diakopter: no, we're on 9 already
21:37 diakopter oh cool
21:37 [Coke] jnthn++
21:41 jnthn Yeah, you still just unzip it into UNIDATA in your MoarVM build directory, nothing's changed there.
21:47 domidumont joined #moarvm
21:58 [Coke] jnthn: it's just not written down anywhere - if you don't mind, I'll throw a comment into the p5 script.
21:59 [Coke] also, I think we're not processing LineBreak.txt, which is making dmaestro sad.
22:00 [Coke] (it says that 0x2014 is a B2 (Break_Both), but we're not processing that yet, I don't think.
22:00 diakopter a couple of the files actually duplicate values from the other property files, or did at one point
22:00 diakopter they might have de-duped it
22:01 [Coke] I don't see that particular bit of info elsewhere, but am just n00b.
22:02 diakopter can you still query a property directly? did you try asking it about Break_Both
22:03 dalek MoarVM: c2af75c | coke++ | tools/ucd2c.pl:
22:03 dalek MoarVM: Add comment about where to find UNIDATA
22:03 dalek MoarVM: review: https://github.com/MoarVM/MoarVM/commit/c2af75c5ac
22:04 [Coke] m: "\C[EM DASH].uniprop('Line_Break');
22:04 camelia rakudo-moar 689da7: OUTPUT«===SORRY!===␤Unrecognized backslash sequence: '\C'␤at <tmp>:1␤------> "\⏏C[EM DASH].uniprop('Line_Break');␤Unable to parse expression in double quotes; couldn't find final '"' ␤at <tmp>:1␤------> "\C[EM DASH].uniprop('Line_Brea…»
22:04 [Coke] m: say "\c[EM DASH]".uniprop('Line_Break');
22:04 camelia rakudo-moar 689da7: OUTPUT«BK␤»
22:05 [Coke] m: say "\c[EM DASH]".uniprop('Break_Both');
22:05 camelia rakudo-moar 689da7: OUTPUT«BK␤»
22:05 [Coke] Break_Both ~~ B2
22:06 diakopter [Coke]: https://github.com/MoarVM/MoarVM/blob/master/tools/ucd2c.pl#L1580
22:06 diakopter looks like it needs B2 added
22:06 diakopter actually it needs several added
22:07 [Coke] AH. I was looking for the name of the file, didn't realize it was not a literal.
22:08 diakopter heh, there are only 3 characters with B2
22:08 [Coke] there's a list at the top of the LineBreak file, I assume we could just pull all of them?
22:09 diakopter I think so, but some of those have some interpreted meaning
22:10 diakopter for instance, I don't think it's currently handling ID and XX correctly as defaults
22:11 [Coke] Heading out now, thanks.
23:15 pyrimidine joined #moarvm
23:46 pyrimidine joined #moarvm
23:56 pyrimidine joined #moarvm

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