| IRC logs at
Set by AlexDaniel on 12 June 2018.
00:45 sena_kun left 01:01 sena_kun joined 02:09 Kaiepi left 02:12 Kaiepi joined 02:46 sena_kun left 03:02 sena_kun joined 04:07 angelds joined 04:46 sena_kun left 05:02 sena_kun joined, AlexDaniel left 06:46 sena_kun left 07:00 sena_kun joined 07:01 AlexDaniel joined, AlexDaniel left, AlexDaniel joined 08:01 evalable6 left, linkable6 left 08:02 linkable6 joined 08:03 evalable6 joined 08:09 domidumont joined 08:45 sena_kun left 08:58 sena_kun joined 09:22 Kaiepi left 09:27 Kaiepi joined 09:28 Kaiepi left 09:29 Kaiepi joined 09:32 Kaiepi left, Kaiepi joined 10:25 Kaiepi left 10:30 Kaiepi joined 10:31 Kaiepi left 10:32 Kaiepi joined 10:34 Kaiepi left, Kaiepi joined 10:46 sena_kun left 10:51 domidumont left 10:53 domidumont joined 10:57 Kaiepi left 11:02 Kaiepi joined, sena_kun joined 11:29 Kaiepi left 11:33 Kaiepi joined 11:36 Kaiepi left, Kaiepi joined 12:22 angelds left 12:32 Kaiepi left 12:36 Kaiepi joined 12:38 Kaiepi left 12:39 Kaiepi joined 12:43 AlexDaniel left, Kaiepi left 12:44 Kaiepi joined 12:46 sena_kun left 12:50 Kaeipi joined, Kaiepi left 13:01 sena_kun joined 13:03 lucasb joined
Geth MoarVM: MasterDuke17++ created pull request #1233:
Use new functions introduced in ltm v1.2.0...
14:47 sena_kun left 15:01 sena_kun joined 16:04 AlexDaniel joined, AlexDaniel left, AlexDaniel joined 16:19 zakharyas joined 16:45 sena_kun left 17:00 sena_kun joined 17:25 Kaeipi left 17:26 Kaeipi joined
MasterDuke hm. is failing because "mp_set_double implementation is only available on platforms with IEEE754 floating point format". however, it appears we're assuming IEEE754 anyway according to 17:43
i can always just back out the part of the change that uses mp_set_double(), but are we still susceptible to problems in our custom code? 17:44
17:48 Kaeipi left 17:49 Kaiepi joined 17:50 Kaiepi left 18:07 ilogger2 joined 19:00 sena_kun joined 19:10 domidumont joined 19:29 japhb left
MasterDuke isn't this dangerous? 20:32
shouldn't we be using mp_init_copy()? otherwise we're just copying the mp_int, not the stuff that may be inside it? 20:33
20:34 domidumont left 20:45 sena_kun left 20:46 japhb joined 20:59 sena_kun joined
timotimo MasterDuke: it doesn't get mp_copy'd here because the body would otherwise disappear when the function is left, due to being put onto the stack 21:23
i'm not sure why alloca is used here, i think it could just be "mp_int i" and use &i whenever it's mentioned 21:24
i think there's also an mp_clear missing before the second exception throw there
MasterDuke re mp_clear, yeah, i have a new commit i'm just about to push that adds mp_clears everywhere where missing 22:11
timotimo: but then shouldn't it be mp_copy+memcpy? 22:14
wait. you mean the MVMP6bigintBody *body would disappear? 22:26
i'm thinking the else block should be: mp_int *i_cpy = MVM_malloc(sizeof(mp_int)); mp_init_copy(i_cpy, i); body->u.bigint = i_cpy; 22:29
22:46 sena_kun left 23:00 sena_kun joined