github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
nwc10 good *, #moarvm 05:36
MasterDuke how do we want to handle aborts from gmp? (try to) detect the potential error conditions before calling the gmp function? trap any errors? 09:00
nwc10 how does it abort? call abort()? 09:01
MasterDuke well, i don't think it's any one thing. sometimes it's floating point exceptions, sometimes divide by 0, no mem available 09:02
nwc10 er, erk. 09:03
OK, I don't know. But if it's being called from gcc as part of an optimiser, surely that has to handle all this reliably. So how does it avoid crapping out, and instead returning useful messages? 09:04
I really don't know. It's sounding like "doctor doctor, it hurts when I do this" "Well, don't do that then"
as in - it sounds like we don't want to even call it if we know that it's going to bomb out 09:05
but I don't know whehter all the cases can be spotted before it's called.
I've not even looked into this code
MasterDuke you've not looked at the moarvm code that calls the gmp functions? or the gmp implementation? 09:08
nwc10 neither part 09:17
MasterDuke gmplib.org/repo/gmp/file/tip/errno.c#l44 09:19
nwc10 oh my. For that I trade you this: github.com/JuliaLang/julia/issues/37327 09:22
which looks kind of bonkers
but seems to start to answer "how does Julia do it"? 09:23
nine MasterDuke: linux.die.net/man/3/feenableexcept 12:48
MasterDuke nine: thanks, i'll give that a try 12:51
nine Huh.... the GMP docs tell you to use MPFR instead in new projects. And MPFR actually documents how to handle exceptions: www.mpfr.org/mpfr-current/mpfr.html#Exceptions 12:54
Geth MoarVM: patrickbkr++ created pull request #1358:
Minor optimization of MVM_file_isexecutable on Windows
17:09
brrt \o 17:50
nwc10 o/ 17:57
pamplemousse o/ 18:23
nine \o 18:34
brrt ohai pamplemousse 18:48
and nine, and nwc10 18:49
pamplemousse Hello! It's been a hot minute. Hope everyone's doing well! 18:54
lizmat is doing fine, pamplemousse ! 18:55