[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [lmi] What is --host for a native build? [Was: Compiling takes longe
Re: [lmi] What is --host for a native build? [Was: Compiling takes longer with gcc-4.9.2]
Tue, 19 Jan 2016 17:54:24 +0100
On Tue, 19 Jan 2016 16:31:56 +0000 Greg Chicares <address@hidden> wrote:
GC> Okay, we've established that '--host=i686-w64-mingw32' is correct for
GC> Cygwin's MinGW-w64 cross compiler, and that '--host=i686-pc-mingw32',
GC> though wrong on the face of it, does in fact work for that cross
GC> compiler because we fully specify each tool.
FWIW I tried to find if the --host option value was used for anything else
but it doesn't seem so: the documentation at
When the --host is used, configure will search for the
cross-compiling suite for this platform. Cross-compilation tools
commonly have their target architecture as prefix of their name.
and looking at the generated configure itself I can indeed only see it used
for setting "ac_tool_prefix", which is then used as you'd expect, and also
setting "cross_compiling" to "yes" which is, again, used later to check for
things that it doesn't make sense to do when cross-compiling (like running
GC> (but not all--there's no prefixed preprocessor, e.g.),
Surely because the preprocessor is architecture-independent anyhow.
GC> which seems to suggest that '--host=i686-w64-mingw32', which is in HEAD
GC> now, is the best choice. I suppose it doesn't matter; I don't plan to
GC> change it unless it becomes necessary.
I agree that it's the best choice and I also think that it won't become
necessary to change it. OTOH I still think that we should stop setting all
tools manually and just let configure do it for us.
[lmi] cgicc changes [Was: Compiling takes longer with gcc-4.9.2], Greg Chicares, 2016/01/18