github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
jnthn Wow...compiling HEAD is very warning-y 14:08
lizmat I only see lots and lots of "loop not vectorized" 14:18
and some format specifies type 'long' but
the argument has type 'MVMuint64'
jnthn gist.github.com/jnthn/d2c700e9d192...a2e5a9bed3 fwiw 14:25
I don't have time to go digging at the moment.
Kaiepi can someone review github.com/MoarVM/MoarVM/pull/1222 ? 15:19
Guest38485 jnthn, lizmat: a few more warning options have been enabled recently 15:39
nine: after some more debugging it turns out that the 'sc' is somehow cleared in this routine github.com/MoarVM/MoarVM/blob/mast...on.c#L2843 16:03
nine jnthn: we compile with -Wall and -Wextra now. I've fixed almost all the warnings I got but different compiler (versions) will be more noisy still 18:36
jnthn nine: Yeah, mine is a bit noisy. I'm all for more warnings in principle, just know it's easy to not spot new ones among the noise when there's a lot of them triggered. 18:38
nine I know. That's why I've fixed several hundreds of them. I only left ~ 3 that I hope will be tackled by domain specialists. And of course those I simply do not get on my gcc 18:39
jnthn nine++ 18:44
Will hopefully get chance to look at the ones I see soonish 18:45
The fallthrough ones are very odd, it's a warning on something that looks like it's meant to prevent one...
Anyways, home time... :)
nine Oh, seems like that fallthrough attribute is only supported on gcc >= 7 18:46
Same for #pragma GCC diagnostic ignored "-Wimplicit-fallthrough" 18:48
nine dogbert17: I think I got it! 22:00
or still not :( 22:01
Or maybe there's just more issues with this code. 22:03
One thing's certain: this line github.com/MoarVM/MoarVM/blob/mast...all.c#L189 must be moved 2 down to be outside the if just like the pushes. Otherwise we may end up with outdated pointers on the temp root stack 22:04
dogbert17 nine: cool 22:21
the only thing I have found is that the sc, which get trashed has the following description: site#sources/0F50B574622C7F9833911E194CCEAFFE83348161 (Inline::Perl5::Interpreter) 22:22
it shows up many times before the crash occurs 22:23
Native call expected return type with CPointer representation, but got a P6opaque (Scalar) 22:34
in method p5_call_parent_gv
on the positive side, the SEGV's seems to have vanished 22:43

Powered by Cro and the Raku® Programming Language.

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

Thank you!