Thinking of glib's GObject (and) Introspection API

Allen Winter winter at kde.org
Mon May 9 13:42:44 PDT 2016


Thanks, committed in 0850f0e185cb435bf52a73dd23bfb5208197907c

also thanks for you hint about XML_CATALOG_FILES.  I added a hack
to the buildsystem to handle that on Mac, so now travis builds
on Mac work again.

Miao: I'd really appreciate if you could deal with cleaning up all those compile warnings with clang.
ie.  -Wtypedef-redefinition and -Wenum-conversion

On Monday, May 09, 2016 08:55:32 AM Milan Crha wrote:
> On Fri, 2016-05-06 at 10:36 -0400, Allen Winter wrote:
> > we already have this code in config.h.cmake.  question is why it
> > isn't being used with mingw.  can you look at fixing this just once
> > inside config.h.cmake?
> 
> 	Hi,
> my bad, I didn't realize it's there already. The MinGW actually uses
> that code, the warning it shows here is that the members of
> the 'struct tm' can be used uninitialized. I thought it's due to the
> gmtime_r() not being known (similar as in the evolution-data-server),
> but I see now that I was wrong. When you compare my define and the
> define you've used, then the only difference is that I also memset()
> the tmp content with zeros, thus the warning about using uninitialized
> variables went away.
> 
> Thinking of it, it was a very poor idea of "solving" the warnings, I'm
> sorry for that. The correct solution would be to check whether
> gmtime_r() returned non-NULL in the code and react based on it. Please
> find the attached patch with this approach. Feel free to change the
> return values on error, I thought these would work fine.
> 	Bye,
> 	Milan




More information about the libical-devel mailing list