00:27
llfourn joined
01:28
llfourn joined
02:10
ggoebel7 joined
02:24
llfourn joined
02:51
FROGGS joined
|
|||
dalek | arVM: 7e4711a | timotimo++ | tools/graph_spesh.p6: add missing "flat" to graph_spesh to fix output |
02:59 | |
arVM: 9686530 | timotimo++ | tools/graph_spesh.p6: give a Str to "use lib" instead of an IO::Path. |
|||
arVM: b3b4e41 | timotimo++ | tools/graph_spesh.p6: lines() :eager is no longer a thing. |
03:09 | ||
flussence | .tell brrt libuv-1.8 branch builds and spectests all clear here, except advent2013-day14 but that was already flapping | 03:22 | |
03:25
FROGGS joined
|
|||
dalek | arVM: a3fba3d | timotimo++ | tools/graph_spesh.p6: understand coderefs |
03:31 | |
03:46
FROGGS joined
04:12
FROGGS joined
08:49
FROGGS joined
09:13
FROGGS joined
09:58
domidumont joined
10:04
domidumont joined
11:02
virtualsue joined
11:17
domidumont1 joined
12:20
flaviusb joined
12:29
geekosaur joined
13:42
FROGGS joined
14:11
ilbot3 joined
14:27
ilbot3 joined
14:39
virtualsue joined
|
|||
timotimo | cool | 14:57 | |
dalek | arVM: 651a1b2 | timotimo++ | tools/graph_spesh.p6: also recognize callsites |
15:10 | |
timotimo | so i'm wondering how a "heap explorer" would function | 15:19 | |
it'd be easy enough to write something that walks the nursery and gen2 and collects a whole bunch of stats, that's for sure | |||
what would a user expect to do with such a tool? | 15:22 | ||
15:51
brrt joined
|
|||
timotimo | hmm. the visualvm heap explorer lets you get a list of all instances for a given class, see all its fields and also find a "nearest GC root" | 15:59 | |
16:32
zakharyas joined
16:34
FROGGS joined
|
|||
timotimo | i wonder if i should let users give an environment variable to choose the type of profile to generate and implement a memory profiler that gives more detailed statistics about allocations and GC runs, but doesn't do the call graph ... or something like that | 16:57 | |
adding more details to the GC page would be kind of nice, but the profile json blobs are usually already too big in many cases | 16:58 | ||
flussence | I'm looking at bugs.gentoo.org/buglist.cgi?quicksearch=moarvm -- most of them are long since fixed on this site, but 504720 isn't | 17:33 | |
s/site/side | |||
18:39
virtualsue_ joined
18:57
FROGGS joined
21:29
vendethiel joined
22:26
vendethiel joined
23:07
leont joined
23:58
ilmari_ joined
|