github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
timotimo MasterDuke: i've updated the userscript 07:27
MasterDuke timotimo: huh, the gen/* links are working, but e.g., src/Perl6/Compiler.nqp:38 is linking to github.com/MoarVM/MoarVM/blob/mast...er.nqp#L38 10:25
timotimo ah, that has to be nqp, right? 10:26
oh
no, rakudo. but it is moarvm instead
example link please
MasterDuke github.com/rakudo/rakudo/issues/2921 10:27
timotimo updated 10:31
MasterDuke works, timotimo++ 10:40
timotimo my first moarvm that asks for confprogs when hitting the profiler is now compiled 11:48
MasterDuke cool 12:01
timotimo: btw, do you remember where i/we left off with trying to time the profiler entering and leaving a block (so we could subtract that time from the profile afterward)? 12:30
timotimo sorry, i don't remember
MasterDuke hm, and i don't think i committed anything. maybe i stashed some changes though... 12:34
do you know a good way to preview the changes for each stash? 12:36
ah, found my stash 12:51
so `MVM_profile_log_enter(tc, tc->cur_frame->static_info, MVM_PROFILE_ENTER_NORMAL);` is wrong, but anything other than `tc->cur_frame->static_info` as the second option causes segfaults iirc 12:58
timotimo git stash show -u should give the diff 13:05
dogbert17 timotimo: you around? 14:01
timotimo now am 14:03
dogbert17 have time for a question 14:04
?
I would like to send a PR fro MoarVM, updating libuv to 1.29.1 the latest version 14:05
timotimo sounds good
dogbert17 is it enough to clone my fork of MoarVM, sync with upstream followed by 'perl Configure.pl' to get at the submodules ... 14:06
go to libuv dir and checkout v1.29.1, go up to 3rdparty and 'git add' the libuv dir? 14:07
have I forgotten anything important? I should add that I have spectested it on my 64 bit Linux and it was clean 14:08
timotimo i think that's it 14:09
dogbert17 I'll give it a go then, thx 14:11
Geth MoarVM: dogbert17++ created pull request #1106:
Update libuv to version 1.29.1
14:13
dogbert17 I hope I did that correctly 14:14
I guess we should update libtommath as well ... 14:16
although I suspect that might be trickier
timotimo great idea, calculate the end address of the bytecode to be bytecode_size after the chosen entrypoint 19:22
no wonder it's been finding lots of NYI ops at the end
MasterDuke timotimo: think there's going to be a way to call the existing profiling functions as is to get the timing? or think either some sort of flag or new function that's almost identical be added? 23:38