On Fri, Aug 7, 2015 at 04:53, Peter Bex wrote:
> On Fri, Aug 07, 2015 at 09:54:22AM +0200, Christian Kellermann wrote:
> > * Claude Marinier [150806 20:15]:
> > > Note that I did not have this problem with the release candidates.
> >
> > Could you try with them again and tell me exactly which one worked?
> > Because there are no differences that touch code between the rc4 and
> > the 4.10.0 version in git. If there are differences in behaviour for
> > you this could mean we shipped a rc with broken pregenerated C code
> > for whatever reason and I'd like to investigate that.
> > I agree, this is somewhat worrying.
>
> One problem with MINGW is that it's badly maintained (unless I'm
> mistaken, the most recent binary release is from 2013). Mingw-w64
> is better maintained, and despite the name it also works on 32 bit
> systems. I think it does support getc_unlocked. You can find it
> here:
http://mingw-w64.orgHi,
Building both 4.10.0 and rc4 with the MinGW installation on the Windows 7
64-bit PC at work produces the same error : undefined reference to
`getc_unlocked'. This was from a "DOS box" (clink) without MSYS.
Building on rc4 on the same PC with mingw-w64 works without error (also
without MSYS). Check works as expected as does numbers egg installation.
I have had problems with MinGW and re-installed a couple of times. I am
growing less confident in these installations. This is a good time to move
to mingw-w64; I will try it at home later today.
Is it useful to dig deeper into the problem with the old MinGW or should
I just drop it?
Thanks.
--