github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
timotimo i notice i haven't used my rss feed reader in a while 00:03
00:35 lizmat left
samcv i'm making an atkey for unsigned native int. should i name it atkey_u or atkey_u64 03:29
Geth MoarVM: 4eda279b5f | (Samantha McVey)++ | src/6model/serialization.c
Fix three memory leaks when serialization fails

In some instances when calling fail_deserialize we pass through c strings malloc'd by a different function but never free it. Pass these through and use MVM_exception_throw_adhoc_free_va instead of the non-free version.
Also ensure that reader->root.dependent_scs is free'd if it exists.
03:37
MoarVM: samcv++ created pull request #876:
Add support for all native int types to getlexref_ni and new atkey_i/u* ops
04:04
04:11 huggable left, huggable joined 04:50 AlexDaniel left 06:00 AlexDaniel joined 06:09 brrt joined 06:12 domidumont joined 06:18 domidumont left, domidumont joined 06:28 robertle joined 07:11 lizmat joined 08:23 zakharyas joined 08:31 MasterDuke left 08:46 lizmat left 09:50 brrt left 09:51 AlexDaniel left 10:02 AlexDaniel joined 10:14 squashable6 left 10:38 squashable6 joined 10:54 zakharyas left 10:55 zakharyas joined 11:05 AlexDaniel left 11:10 AlexDaniel joined 11:14 zakharyas left 11:20 brrt joined 12:11 AlexDaniel left
brrt \o 12:12
12:30 AlexDaniel joined
timotimo samcv: but jnthn already refused to accept atkey_iEverySize ops? 12:41
jnthn Indeed. :S 12:43
A _u one is fine
But not all the sized variants
Should just do the required introspection inside of the MVMContext 12:44
jnthn left a note on the PR suggesting what to do 12:48
12:49 zakharyas joined 12:57 brrt left 13:03 brrt joined 13:07 AlexDaniel left 13:08 scovit left 13:13 scovit joined, scovit left 13:14 scovit joined 13:31 jnthn is now known as KingWhenceless 13:32 lizmat joined 13:46 lizmat left 13:51 zakharyas1 joined, KingWhenceless is now known as jnthn 13:53 zakharyas left 14:35 dogbert2_ left 14:43 brrt left 15:00 domidumont left 15:26 robertle left 15:49 zakharyas1 left 16:09 AlexDaniel joined 16:46 AlexDani` joined 16:47 AlexDaniel left 17:20 robertle joined 19:05 MasterDuke joined 19:45 MasterDuke left 19:50 MasterDuke joined 19:54 MasterDuke left 20:02 MasterDuke joined
MasterDuke i still get a warning: src/profiler/telemeh.c: In function ‘MVM_telemetry_interval_annotate_dynamic’: src/profiler/telemeh.c:203:5: warning: ‘strncpy’ specified bound depends on the length of the source argument [-Wstringop-overflow=] strncpy(temp, description, temp_size); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ src/profiler/telemeh.c:201:17: note: length computed here temp_size = strlen(description) + 1; 20:18
the problem is that `char* description` is passed into MVM_telemetry_interval_annotate_dynamic, but its length isn't 20:19
so i guess gcc thinks if it isn't terminated, the strlen could give bad results, and then the strncpy would be bad 20:20
however, the only string that's ever passed to MVM_telemetry_interval_annotate_dynamic is created by MVM_string_utf8_c8_encode_C_string (which does put a null at the end of the string it returns) 20:22
we could just pass the size that MVM_string_utf8_c8_encode_C_string knows about along to MVM_telemetry_interval_annotate_dynamic and use that in the strncpy, and hopefully that would quite the warning
or is there some way to mark that that particular occurance is safe? 20:23
i haven't found anything searching around yet
timotimo i should probably look into that ... 20:30
i'm actually doing grant work right now and webstorm is giving me some major grief
it finds that some of my code is unreachable, though clearly i see it being executed 20:31
MasterDuke never heard of webstorm? 20:32
timotimo it's one of the JetBrains IDEs 20:42
i run mine with the Comma plugin ;)
MasterDuke ah, nice 20:43