00:36 vendethiel joined 01:00 vendethiel joined 02:47 ilbot3 joined 03:56 vendethiel joined 04:46 ggoebel111111117 joined 05:46 oetiker joined 06:16 FROGGS joined 07:40 FROGGS joined 07:51 JimmyZ joined 08:14 zakharyas joined 08:35 kjs_ joined, JimmyZ_ joined 08:42 tgt joined 08:45 brrt joined
brrt \o 09:10
let me backlog and see if anybody else saw big jit failures
wait a minute
[Coke] - what kind of make do you have?
and more importantly, is it possible that your dynasm-compiled file may be out of sync 09:11
because that would certainly explain why the failures that you report tend to be out of sync with the repository :-)
and it would be a fixable error
09:26 kjs_ joined
timotimo that would certainly be an interesting slash horrifying kind of bug 10:38
brrt hmm 10:41
12:18 ggoebel111111117 joined 12:59 woolfy joined
[Coke] I use the xcode-supplied tools. 13:00
GNU Make 3.81 13:01
This program built for i386-apple-darwin11.3.0
timotimo 3.81 is quite recent
[Coke] also, I do a fresh build every day, clean checkouts. 13:02
timotimo OK, so if it didn't do the .dasc to .c thing, it'd explode altogether
[Coke] I am doing parallel build wherever possible, though. (not currently possible to build moar that way, no passthrough make options)
just started today's run, will let you know what happens. This is the second time I've had big failures with no reproducability elsewhere. 13:03
timotimo that's pretty bad
thank you once again for your continued test runs
it's hard to express how helpful they have been; even though sometimes people seem to not care about test failures for a couple of days ;( 13:04
[Coke] heh.
when I have time, I don't mind also fudging (and even occasionally fixing) things. hard to rememer everyone is at least as time crunched as I am. :) 13:05
thanks to everyone for all the hard work!
brrt we try :-) 13:24
[Coke] looks like at least one of moar's segfaults from yesterday is gone. 13:25
so, flapping segfaults, maybe? blah. 13:26
brrt ah.. that's nasty 13:30
how much memory do you have? could potentially be a gc issue with JIT
timotimo huh? but we use MVM_malloc everwhere that ought to do error checking, no? 13:32
brrt i say potentially 13:49
i can be wrong
very wrong
timotimo mhm 13:51
15:33 FROGGS joined 16:17 tgt joined 16:19 tgt joined 16:26 tgt joined
[Coke] brrt: this machine has 16G 16:54
so that's probably not it.
17:01 woolfy left 17:08 woolfy joined 17:30 tgt joined 17:50 FROGGS_ joined 19:39 lizmat joined 20:16 Ven joined
dalek arVM: 69f322c | (Tobias Leich)++ | src/io/syncpipe.c:
fetch exit code of a closed pipe on windows
timotimo can we get something like a supply that emits whenever a Proc::Async that's open for writing is "ready to read from stdin"? 23:14
23:16 kjs_ joined
timotimo hm. it seems like wanting to detect when a subprocess is waiting for input on stdin requires the use of a pty :\ 23:22
i will have to do something "smarter" in that case. 23:25