Perl 6 - the future is here, just unevenly distributed

IRC log for #moarvm, 2017-12-15

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

All times shown according to UTC.

Time Nick Message
00:19 Morfent joined #moarvm
00:40 TimToady joined #moarvm
01:15 MasterDuke joined #moarvm
03:01 ilbot3 joined #moarvm
03:01 Topic for #moarvm is now https://github.com/moarvm/moarvm | IRC logs at  http://irclog.perlgeek.de/moarvm/today
04:05 bart___ joined #moarvm
04:10 brrt good * #moarvm
04:10 brrt good night, for all intents and purposes
04:10 brrt ASAN is complaining loudly of memory leaks in spesh
04:10 brrt which is a pain because that was not what i wanted to research
04:11 brrt patches coming up, obviously
04:45 Geth ¦ MoarVM/memory-leak-fixes: 85ca55f968 | (Bart Wiegmans)++ | 5 files
04:45 Geth ¦ MoarVM/memory-leak-fixes: Fix memory leaks reported by ASAN
04:45 Geth ¦ MoarVM/memory-leak-fixes:
04:45 Geth ¦ MoarVM/memory-leak-fixes: There were memory leaks in:
04:45 Geth ¦ MoarVM/memory-leak-fixes: - the string stream decoder separator specification buffer(s)
04:45 Geth ¦ MoarVM/memory-leak-fixes: - the static frame spesh body statistics buffer
04:45 Geth ¦ MoarVM/memory-leak-fixes: - the spesh plan type stats buffer
04:45 Geth ¦ MoarVM/memory-leak-fixes: - the spesh plan, when the static frame could not be added
04:45 Geth ¦ MoarVM/memory-leak-fixes: review: https://github.com/MoarVM/MoarVM/commit/85ca55f968
05:20 brrt there's another leak in nativecall build of the jit code object, it seems
05:23 brrt but that's a bit more mysterious it seems
05:24 Geth ¦ MoarVM/memory-leak-fixes: 3a772a67c1 | (Bart Wiegmans)++ | src/strings/unicode_ops.c
05:24 Geth ¦ MoarVM/memory-leak-fixes: Another leak in unicode_ops.c
05:24 Geth ¦ MoarVM/memory-leak-fixes:
05:24 Geth ¦ MoarVM/memory-leak-fixes: Would decode string to ascii, which would allocate the result, which
05:24 Geth ¦ MoarVM/memory-leak-fixes: would then not be freed.
05:24 Geth ¦ MoarVM/memory-leak-fixes: review: https://github.com/MoarVM/MoarVM/commit/3a772a67c1
06:37 brrt joined #moarvm
06:54 domidumont joined #moarvm
07:01 domidumont joined #moarvm
07:15 brrt .tell nine we can apparently leak jitcode from MVM_nativecall_buidl
07:15 yoleaux brrt: I'll pass your message to nine.
07:15 brrt i'm not seeing how since i'd think that it be garbage-collected
07:16 Geth ¦ MoarVM: 509315503d | (Bart Wiegmans)++ | src/jit/linear_scan.c
07:16 Geth ¦ MoarVM: [JIT] lets not overengineer linear-scan spill loop
07:16 Geth ¦ MoarVM:
07:16 Geth ¦ MoarVM: The orignal pointer-pointer plus pointer was much simpler, so lets
07:16 Geth ¦ MoarVM: just make it explicit rather than muck about with && to force order
07:16 Geth ¦ MoarVM: and conditionality into assignment.
07:16 Geth ¦ MoarVM: review: https://github.com/MoarVM/MoarVM/commit/509315503d
07:36 Morfent sorry if this is the wrong place to ask, but would support for locales in moar be something that'd be worth contributing?
07:39 moritz Morfent: I think this is absolutely the right place to ask
07:40 moritz my experience with locales in Perl 5 isn't great, so I'm a bit biased against them, but it's worth waiting for more opinions :-)
07:40 Morfent perfect, thanks
07:46 brrt joined #moarvm
09:28 zakharyas joined #moarvm
09:44 Ven` joined #moarvm
10:37 brrt joined #moarvm
11:27 Ven` joined #moarvm
13:14 Ven` joined #moarvm
13:34 AlexDaniel joined #moarvm
14:20 Ven` joined #moarvm
14:43 nine .
14:43 yoleaux 07:15Z <brrt> nine: we can apparently leak jitcode from MVM_nativecall_buidl
14:43 brrt ohai nine
14:43 nine hi brrt
14:44 brrt how are you doing
14:44 brrt i'm not sure how exactly we leak jitocde, and to all extents it seems like we can be leaking expr trees
14:45 Zoffix joined #moarvm
14:45 Zoffix Reminder: release is in 2 days and we still have 3 release blockers, if anyone can figure out how to fix them: https://github.com/rakudo/rakudo/issues?q=is%3Aissue+is%3Aopen+label%3A%22%E2%9A%A0+blocker+%E2%9A%A0%22
14:48 Zoffix left #moarvm
14:56 lizmat S01-perl-5-integration/basic.t pretty much segfaults 50+% of the time for the "Passing a Perl 6 coderef to Perl 5" test
14:56 brrt blimey
14:56 lizmat oddly enough, this appears *less* if the system is under load
14:57 lizmat and I haven't been able to get it to crash with MVM_JIT_EXPR_DISABLE=1
15:00 nine brrt: all in all fine, though work right now requires pretty much all my energy. How're you?
15:01 nine brrt: do we leak jit code or jit graphs?
15:02 brrt jitcode
15:02 brrt also jit trees, no jit graphs
15:02 brrt i'm not sure how we leak them
15:03 nine brrt: that is only referenced by the NC body and freed in gc_cleanup
15:05 zakharyas joined #moarvm
15:06 brrt i k now
15:07 brrt so i'm not sure how the leak happens
15:09 geospeck joined #moarvm
16:08 brrt joined #moarvm
16:14 brrt well, i do have a hypothesis though
16:14 brrt this happens mostly in short-lived scripts
16:15 brrt so i think it might happen if the process is shut down before spesh/JIT are done
16:47 timotimo uh oh
16:48 timotimo i can't send data on a tcp socket while i'm waiting for data on another
16:48 timotimo that seems rather ... bad?
16:48 nine brrt: MVM_SPESH_BLOCKING=1 should help confirm that hypothesis
16:48 brrt yeah, it stops the issue from happening
16:49 brrt whcih doesn't *quite* prove there is nothing going on though
17:01 AlexDaniel joined #moarvm
17:05 domidumont joined #moarvm
18:22 timotimo any opinions on whether we should split the io lock on tcp sockets into one for reading and one for writing?
18:25 lizmat full duplex seems to be the norm, so a split seems appropriate ?
18:25 timotimo i'm impressed it took that long for someone to stumble over this
18:27 lizmat well, I'm glad someone did stumble  :-)
18:27 lizmat *and* reported it  :-)
18:31 timotimo oh, huh
18:32 timotimo there are operations that are neither read nor write. like close
18:32 timotimo or seek or tell ...
18:32 timotimo ugh, this is kind of hairy
18:32 lizmat seek / tell on sockets ?
18:33 lizmat tcp sockets ?
18:33 timotimo or isatty, or fileno
18:34 timotimo libuv does offer read-write-locks
18:35 lizmat afk&
18:38 timotimo though that may not be the right thing at all.
18:39 geospeck joined #moarvm
20:15 geospeck joined #moarvm
20:30 geospeck joined #moarvm
20:30 geospeck joined #moarvm
20:57 committable6 joined #moarvm
20:58 releasable6 joined #moarvm
20:58 nativecallable6 joined #moarvm
20:58 squashable6 joined #moarvm
20:58 bloatable6 joined #moarvm
20:58 greppable6 joined #moarvm
20:58 evalable6 joined #moarvm
20:58 quotable6 joined #moarvm
20:58 unicodable6 joined #moarvm
20:58 bisectable6 joined #moarvm
20:58 benchable6 joined #moarvm
20:58 reportable6 joined #moarvm
20:58 coverable6 joined #moarvm
20:58 statisfiable6 joined #moarvm
20:59 dogbert17 joined #moarvm
21:02 dogbert17 thought we'd gotten rid of the 'Must not GC when in the specializer/JIT' but I was wrong
22:26 ggoebel joined #moarvm
23:05 brrt joined #moarvm
23:07 synopsebot joined #moarvm
23:07 Geth joined #moarvm
23:10 Geth joined #moarvm
23:13 Geth joined #moarvm
23:15 MasterDuke joined #moarvm

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