| IRC logs at
Set by AlexDaniel on 12 June 2018.
00:05 brrt joined 00:29 brrt left 04:08 evalable6 left, linkable6 left 04:10 linkable6 joined, evalable6 joined
nine MasterDuke: patching gmp to fix issues means that we can't use an already installed gmp 06:57
07:04 SmokeMachine left 07:22 SmokeMachine joined 07:29 sena_kun joined
MasterDuke nine: true. we have patches on a couple of our 3rdparty libraries, some of them are for building, which obviously doesn't matter for --has-foo, but i don't think they all are. we don't test --has-* in CI do we? 07:49
nwc10 good *, #moarvm 07:51
MasterDuke i think i'm going to go with just '*' today 07:53
nwc10 OK. I hope that tomorrow is better
MasterDuke likewise, fighting a cold for the past week. they usually only last a day or two 07:54
ok, all nqp tests pass (quick hack of ucing the string if the base is > 10) 07:57
down to two failed tests in rakudo's tests and 20 files with fails in a spectest 08:02
huh. two TODOs passed in t/spec/S02-types/int-uint.rakudo.moar
09:40 Altai-man joined 09:43 sena_kun left 10:10 chansen_ left 10:11 leont joined 10:12 chansen_ joined 10:16 brrt joined 11:30 brrt left
MasterDuke ugh. i'm not sure i can fix any of the other problems until i figure out why some bigints are going to zero. e.g., `say +(42 xx 2**99)` gives 0, no idea why 13:08
lizmat xx ? 13:34
aren't we limited to 64bIt wrt to lengths of nqp::lists ? 13:35
shouldn't that just horribly die? or is that your point ?
13:41 sena_kun joined 13:43 Altai-man left
MasterDuke m: say +(42 xx 2**99) 13:48
camelia Cannot .elems a lazy list
in block <unit> at <tmp> line 1
MasterDuke yeah, returning 0 is just all kinds of wrong
the 2**99 is being calculated correctly, but then it's actually getting into elems of src/6model/reprs/VMArray.c and body->elems is 0 (that part isn't really surprising) 13:49
i just haven't figured out what's going wrong in the middle 13:50
m: say so 42 xx 2⁹⁹⁹⁹⁹ # though is this really correct? 13:54
camelia True
MasterDuke i guess cause it's a sequence it's ok 13:55
14:02 Kaiepi joined 14:25 dogbert17 joined
dogbert17 if the MVM_SPESH_CHECK_PRESELECTION is set here,, and it triggers a warning 15:19
when running e.g. t/spec/S06-operator-overloading/circumfix.t is that a bug or something which can be safely ignored? 15:20
MasterDuke the current diff if anybody wants to look 15:42
tobs MasterDuke: that comment seems outdated now? 15:46
MasterDuke tobs: ha, right. though i hope that code will change again 15:54
if that is the way we end up dealing with gmp returning lowercase values then it'd probably be faster to just uppercase the c str directly, instead of doing it to the MVMString 16:00
17:40 Altai-man joined 17:43 sena_kun left 18:20 zakharyas joined 19:39 brrt joined
nwc10 good *, brrt 19:57
20:01 brrt left
timotimo nwc10: does your DISPLAY_MODE_NONE have a typo in it %) 20:20
nwc10 oh yes. oops. 20:22
but I need to go to bed.
well spotted.
20:48 zakharyas left 21:01 ggoebel__ left 21:20 Altai-man left 21:34 codesections left 22:20 leont left 23:42 ggoebel joined