2011/4/28 Eric Blake <eblake@xxxxxxxxxx>: > On 04/28/2011 01:33 AM, Matthias Bolte wrote: >> libvirt compiles fine but virsh has a warning about vfprintf: >> >> virsh.c: In function 'vshDebug': >> virsh.c:12105:5: warning: function might be possible candidate for >> 'ms_printf' format attribute [-Wmissing-format-attribute] >> >> I'm no sure where to fix that. > > In mingw's <stdio.h>. ÂBut we can't do that. > > vfprintf is NOT being provided by gnulib, so it does NOT understand %zu > or %llu. ÂYet we use it with %zu. > > So the only real solution is to use virAsprintf (which _does_ understand > %zu, thanks to gnulib) and make vshDebug go through an intermediate > buffer, then use fputs on that string, rather than relying directly on > the unsanitized vfprintf. ÂPatch coming up shortly. Yep, fixed now. >> >> make check shows two warnings in gnulib. Maybe Eric wants to take a look: >> >> getsockopt.c: In function 'rpl_getsockopt': >> getsockopt.c:51:7: warning: passing argument 4 of 'getsockopt' from >> incompatible pointer type >> c:\dev\msys_setup\mingw\bin\../lib/gcc/mingw32/4.5.0/../../../../include/winsock2.h:543:32: >> note: expected 'char *' but argument is of type 'int *' > > Yep, I can fix that one in gnulib; I'll have a counterpart patch to bump > .gnulib to the latest once I've got it done. > >> >> test-getaddrinfo.c:28:1: warning: initialization from incompatible pointer type > > Probably the same category of gnulib fix. Also fixed now. >> Aside from that the rest seems to work fine. > > So I fixed the non-blocking pipe issues that broke 0.9.0 virsh > out-of-the-box? ÂGood to hear. At least the call to virEventRegisterDefaultImpl in virsh doesn't fail anymore, so the pipe setup and stuff seems to work, or at least does not fail anymore. :) Current git + pending (but already ACK'ed) patches from Eric compiles and works fine on mingw, as far as I can tell. Matthias -- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list