Hi Marat,
On Thu, 9 Oct 2014, Marat Radchenko wrote:
> On Wed, Oct 08, 2014 at 12:26:52PM -0700, Junio C Hamano wrote:
> > Marat Radchenko <ma...@slonopotamus.org> writes:
> >
> > > When crosscompiling, one cannot rely on `uname` from host system.
> >
> > That may well be true, but is that limited to cross-compiling to
> > mingw? Would it be generally true for any cross compilation,
> > wouldn't it?
> >
> > What I am wondering is if it is a better solution to make it easier
> > to allow somebody who is cross compiling to express "Mr. Makefile,
> > we know better than you and want you to do a MINGW build for us
> > without checking with `uname -?` yourself", i.e.
> >
> > $ make uname_O=MINGW uname_S=MINGW
> >
> > which would hopefully allow cross-compilation into other
> > environments, not just MINGW.
>
> So, do you really want this patch to be changed from 5-liner into
> a full-blow system detection rewrite based on `cc -dumpmachine`
> instead of `uname`?
Junio's responsibility as a maintainer is to keep this project afloat for
a broader audience than this patch series tries to reach. Keeping that in
mind, a patch that fixes a broad issue (in this case, cross-compiling)
*just* for one small use case (in this case, cross-compiling for Windows
using MinGW) might need a very good argument to convince the maintainer
that a broader solution is either not achievable or so much more
complicated that it is completely outside of the scope of this patch
series.
FWIW this is *exactly* what I pointed out here:
https://github.com/dscho/git/commit/e31124d4e8ba4e48181306198648cdfe73ff9c30#comments
Hmm. This is a very narrow solution to the underlying problem.
Just think about cross-compiling for e.g. MacOSX. The solution
presented in this commit works for MinGW, only for MinGW, and for
MinGW alone. The commit message needs to do a much better job to
make that clear.
And Junio actually pointed out that a broader solution is neither
impossible nor complex: Junio hinted at the use of the conditional
assignment "FOO ?= bar":
https://www.gnu.org/software/make/manual/make.html#index-_003f_003d
IMHO it would make tons of sense to replace the := assignments in
https://github.com/dscho/git/blob/e31124d4e8ba4e48181306198648cdfe73ff9c30/config.mak.uname#L3-L8
by ?= assignments.
Of course it is very obvious where the idea for the patch we are
discussing here comes from: the current config.mak.uname has this:
ifdef MSVC
# avoid the MingW and Cygwin configuration sections
uname_S := Windows
uname_O := Windows
endif
This could be cleaned up in the same run: to be assigned *before* the
conditional assignments of uname_S. Likewise, the MINGW-specific part
should be done by searching for "mingw" in CROSS_COMPILE, also before the
block of the uname_* variable assignments.
Ciao,
Johannes
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
