github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
MasterDuke off to sleep. hopefully you see something useful. catch you tomorrow... 00:11
MasterDuke timotimo: see anything in the spesh log? 07:45
timotimo i somehow forgot after downloading it 08:52
i blame firefox hiding away downloaded things behind a button
oh my 08:53
at first i thought you might have typo'd push-all's cuuid
but there's literally one with 1468 and one with 1568 08:54
MasterDuke ha 08:56
timotimo i see that it's OSR'd 10:04
timotimo so, anyway, it's guarding for an SF, but then it doesn't even fastinvoke, and perhaps that's odd 10:24
it doesn't know anything about either of the arguments, though 10:25
could be another instance of "OSR makes registers unknowable" 10:26
MasterDuke but only the push-all is getting OSR'ed, not the push? 10:27
but i guess the push is inlined into the push-all
timotimo not as i'm seeing 10:28
timotimo i've been wondering if we should perhaps compile a non-osr version of the same frame when spesh gets a moment of calm 11:10
but in this case it looks like push-all is only hit a single time throughout the entire program's run time, so it wouldn't make a difference 11:11
Geth MoarVM/new-disp: cd15d7c84f | (Jonathan Worthington)++ | 2 files
Sketch out some bits relating to callstack changes

No functional changes yet; this just sketches out some data structure definitions.
21:26