Re: [patch] compile: Fix MinGW build [Re: [mingw rfc] Add mkdtemp to gdb/gnulib/]

[Kai Tietz <ktietz@redhat.com>]

----- Ursprüngliche Mail -----
> > Date: Mon, 15 Dec 2014 23:28:01 +0100
> > From: Jan Kratochvil <jan.kratochvil@redhat.com>
> > Cc: brobecker@adacore.com, yao@codesourcery.com,
> > gdb-patches@sourceware.org,
> >         ktietz@redhat.com
> > 
> > Even if it managed to load libcc1.so (it needs host-dependent name
> > libcc1.dll)
> > then it would soon end up at least on:
> > 
> > default_infcall_mmap:
> >   error (_("This target does not support inferior memory allocation by
> >   mmap."));
> > 
> > As currently there is only:
> > 
> > linux-tdep.c:
> >   set_gdbarch_infcall_mmap (gdbarch, linux_infcall_mmap);
> 
> Why is mmap needed here?
> 
> > OK for check-in?
> > 
> > 
> > Thanks,
> > Jan
> > gdb/ChangeLog
> > 2014-12-15  Jan Kratochvil  <jan.kratochvil@redhat.com>
> > 
> > 	Fix MinGW compilation.
> > 	* compile/compile.c (get_compile_file_tempdir): Call error on _WIN32.
> > 
> > --- ./gdb/compile/compile.c	2014-12-14 02:48:38.000000000 +0100
> > +++ ./gdb/compile/compile.c	2014-12-15 23:21:28.788716340 +0100
> > @@ -191,7 +191,11 @@ get_compile_file_tempdir (void)
> >  
> >    strcpy (tname, TEMPLATE);
> >  #undef TEMPLATE
> > +#ifdef _WIN32
> > +  error (_("mkdtemp needs to be implemented for MS-Windows hosts"));
> > +#else
> >    tempdir_name = mkdtemp (tname);
> > +#endif
> >    if (tempdir_name == NULL)
> >      perror_with_name (_("Could not make temporary directory"));
> 
> I think _WIN32 will catch Cygwin as well, which is not what you want.
> 
> Why not just check HAVE_MKDTEMP?
> 

Well, _WIN32 isn't set necessarily for Cygwin.  For Cygwin this define is just set if the windows.h header got included AFAIK.  Anyway, there are better ways to check, and indeed HAVE_MKDTEMP seems to be the most correct check, isn't it?

Kai