[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: *-config programs and multi-arch

Simon McVittie <smcv@debian.org> writes:
> On Thu, 15 Sep 2011 at 16:53:26 +0900, Miles Bader wrote:
>> Tollef Fog Heen <tfheen@err.no> writes:
>> > Your cross-toolchain is supposed to set up a symlink from
>> > /usr/bin/$triplet-pkg-config to /usr/share/pkg-config-crosswrapper which
>> > will then DTRT.  That's the idea at least, I haven't actually tested
>> > it.
>> How is one supposed to make pkg-config use this feature?
>> I mean, say I do:
>>    ./configure --host=muckey-muck-muck --build=my-host-type
> With the symlink Tollef mentioned, that should be all you need. Recent
> versions of pkg.m4 use AC_PATH_TOOL, so they will try to run
> arm-linux-gnueabi-pkg-config (if arm-linux-gnueabi is the host triplet) in
> preference to pkg-config; if arm-linux-gnueabi-pkg-config in your PATH is a
> symlink to pkg-config-crosswrapper, the right thing is meant to happen
> automatically.
> If the depending package was last autoreconf'd with an older version of pkg.m4,
> it might need an autoreconf to pick up the newer, multiarch-aware version.

Hmm, I did autoreconf, and aclocal.m4 didn't change.  I looked at the
pkg.m4 part of that file, and it does indeed use "AC_PATH_TOOL" -- but
that doesn't seem to really do the proper thing for my
cross-compilation test.

Ok, to be more concrete, I was trying to use the mingw cross tools.
I did:

   ./configure --host=x86_64-w64-mingw32 --build=`config.guess`

of course since there isn't an "x86_64-w64-mingw32-pkg-config"
program, instead it just uses the normal pkg-config, and this does the
wrong thing (it _does_ emit a warning message, although it was rather
cryptic until I started writing this post).

It seems to me that the right thing in this case (no host-prefix tool
is found) is simply to act as if pkg-config was not present at all,
instead of using the non-host-prefixed version.


`...the Soviet Union was sliding in to an economic collapse so comprehensive
 that in the end its factories produced not goods but bads: finished products
 less valuable than the raw materials they were made from.'  [The Economist]

Reply to: