github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
00:10 lucasb left 05:31 robertle left 06:12 domidumont joined 06:52 brrt joined
brrt \o 06:59
nwc10 o/ 07:00
07:10 brrt left 07:14 patrickb joined 07:41 zakharyas joined 07:49 domidumont left 07:52 domidumont joined 08:20 brrt joined 08:33 brrt left, brrt joined
jnthn o/ 09:02
nwc10 \o 09:11
09:22 brrt left 09:33 domidumont left
timotimo o/ 11:05
11:05 patrickb left 11:15 brrt joined 11:19 zakharyas left
Geth MoarVM: ed1ce32f7c | MasterDuke17++ (committed using GitHub Web editor) | 5 files
Silence -Wlogical-op-parentheses warnings (#1100)

But adding parentheses around some expressions.		 11:41
11:42 sena_kun joined 12:15 domidumont joined 12:20 patrickb joined 13:00 sivoais left 13:01 sivoais joined 13:02 Kaiepi left, zakharyas joined, Kaiepi joined
MasterDuke what would be a good default return value for this? github.com/MoarVM/MoarVM/blob/mast....c#L83-L99 13:33
clang warns `control may reach end of non-void function [-Wreturn-type]` 13:34
timotimo throw an exception or MVM_panic or whatever 13:35
MasterDuke will that silence the warning?
timotimo probably? 13:36
MVM_panic is probably marked NO_RETURN or something
13:41 pamplemousse joined
MasterDuke yep, MVM_panic silences it 13:41
13:54 pamplemousse left
MasterDuke timotimo: what should be returned here github.com/MoarVM/MoarVM/blob/mast...2153-L2198 ? 14:00
0? 14:01
14:14 brrt left 14:15 brrt joined
timotimo ah, yeah, 0 would be for "no error" 14:15
Geth MoarVM: MasterDuke17++ created pull request #1102:
Silence clang dangling-else & return-type warnings		 14:16
MasterDuke timotimo: ^^^ look good? 14:18
timotimo i don't think check_requirements should explode at the end
in fact, i think it always reaches the end
if things are correct
so i think if it reaches the end it's successful 14:19
but the return value of parse_message_map is ignored; only the argument.parse_fail is used
so perhaps they both want to become void functions
MasterDuke so just run check_requirements at the end, don't return its value? 14:22
timotimo yes, and check_requirements shouldn't panic
MasterDuke just return 1? 14:23
timotimo yes
lizmat And another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2019/05/13/...pumpkings/ 14:24
timotimo or make it void, too
lizmat++ 14:25
timotimo also AFK
MasterDuke timotimo: can't make check_requirements and parse_message_map void without also changing a bunch of macros 14:27
e.g., FIELD_FOUND and CHECK have returns in them 14:28
brrt what are some of the low-hanging fruit PRs I can send my mentee? 14:51
eh, issues, not PRs 14:52
patrickb brrt: moar only? 14:53
brrt moar is most interesting, yes 14:56
github.com/MoarVM/MoarVM/issues/1068 this one is interesting, I think 14:58
jnthn brrt: Remind me what they'll work on?
Geth MoarVM: bdw assigned to pamplemoussecache Issue SEGV when using --full-cleanup github.com/MoarVM/MoarVM/issues/1068
bdw self-assigned SEGV when using --full-cleanup github.com/MoarVM/MoarVM/issues/1068

Silence clang dangling-else & return-type warnings (#1102)
With this, both clang 8.0.0-3 and gcc 8.3.0 don't show any warnings on my machine.
brrt jnthn: Linking into semi-native executables
jnthn ah, neat 14:59
brrt semi-native, because the bytecode will still be MoarVM
but loadable by the system loader
brrt dinner
jnthn Well, github.com/MoarVM/MoarVM/issues/1046 is probably an easy "get your first commit in" task :) 15:00
brrt or that, yes
MasterDuke github.com/MoarVM/MoarVM/issues/1022 maybe? 15:01
15:05 brrt left 15:22 domidumont left 15:38 patrickb left 15:42 pamplemousse joined 15:44 zakharyas left 16:01 brrt joined 16:12 brrt left 16:14 pamplemousse left 16:22 patrickb joined 17:00 robertle joined 17:18 brrt joined 17:33 brrt left
MasterDuke timotimo: did you see my comments about check_requirements and parse_message_map? 18:24
timotimo changing the "return 0" to "return" in those macros should be absolutely fine 18:25
MasterDuke k, i'll give that a try
timotimo thank you for taking care of my rubbish code!
AFK
MasterDuke timotimo: src/debug/debugserver.c:2392:13: error: non-void function 'skip_whole_object' should return a value [-Wreturn-type] CHECK(0, "could not skip object: unhandled type"); 18:33
hmm, making skip_whole_object void fixes all those, but now `src/debug/debugserver.c:2336:21: error: invalid argument type 'void' to unary expression if (!skip_whole_object(tc, ctx, data)) {` 18:34
huh, p6doc has a bunch of `bailed completely`: popcompsc, settypecache, setmethcache, speshguardtype, param_on_o, ctxouter (some of which i've never seen before) 19:35
timotimo MasterDuke: maybe i can fix all this up, but perhaps it'll be easiest to have check_requirements return 1 on success and just ignore it in the caller 19:44
for now 19:45
20:37 sena_kun left 20:56 brrt joined
brrt \o 21:20
Kaiepi: I'm thinking of closing this: github.com/MoarVM/MoarVM/issues/1091 21:21
I understand the problem on OpenBSD, but I'm not inclined to 'fix' it; the alternative of return-address manipluation by the JIT is really rather annoying, and I'm not convinced I want to take on that tech debt 21:22
Kaiepi i'm alright with that 21:23
brrt If you want, you maybe can research if it could be reasonably implemented with setjmp/longjmp
(but it's possible that OpenBSD frowns on that, too :-D) 21:24
MasterDuke timotimo: github.com/MoarVM/MoarVM/pull/1102 updated 21:43
timotimo OH! 21:45
github have finally split the "show more context" button into two buttons!
one for "expand below section upwards" and one for "expand upper section downwards"
looks good! 21:46
MasterDuke cool, gonna merge then 21:53
22:00 brrt left 22:47 patrickb left

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!