dalek arVM/line_based_coverage_5: 050cf02 | (Zoffix Znet)++ | tools/parse_coverage_report.p6:
Avoid division by zero error
06:34
arVM/line_based_coverage_5: 3643961 | (Jimmy Zhuo)++ | tools/parse_coverage_report.p6:
Merge pull request #460 from zoffixznet/patch-1

Avoid division by zero error
lizmat jnthn: is it a known issue that the Allocations tab in --profile doesn't show any information anymore ? 08:59
brrt thinks that the jit bisecter and jit comparifier are one of the best investments i've made in the jit 09:27
for instance, now I know that i broke WHEN, although it's not obvious why 10:03
brrt should lldb that... 13:52
lldb, fwiw, is not as polished for command line use as gdb is
timotimo probably not as bad as jdb 13:55
dalek MoarVM/even-moar-jit: 8a8a8e3 | brrt++ | docs/jit/plan.org: 16:58
MoarVM/even-moar-jit: Macro definitions for register names
MoarVM/even-moar-jit:
MoarVM/even-moar-jit: This will allow us to use shorthand register names in the tiler (like
MoarVM/even-moar-jit: rax, rcx) which will be expanded to fully-qualified identifiers on
brrt sorry dalek. that was 4 commits by the way, and MVMJitValue is now dead 16:59
this had become a long-standing goal, so, yay
japhb brrt: What is the fully qualified identifier that goes with 'rcx'? I mean, I think of that as being a full register identifier by itself .... 17:00
brrt next on the list: legitimise the 'pseudo' tile
MVM_JIT_ARCH_X64_RCX
which is an enum for '1' :-)
anyway, afk
japhb Ah, gotcha, thanks brrt. 17:06