MinGW devel rpms

[Yaakov Nemoy]

On Tue, Aug 12, 2008 at 5:37 PM, Richard W.M. Jones <rjones at redhat.com> wrote:
> On Tue, Aug 12, 2008 at 05:21:20PM +0200, Erik van Pienbroek wrote:
>> Do you have a list of the current technical issues you're coping with?
>> Maybe I can help.
>
> Yes, I did ... and in fact I was looking for it earlier today, but
> couldn't find it.  So let's make a list again:
>
> (1) At the moment rpm runs programs like 'strip' on the Windows
> libraries, which actually corrupts them.  To avoid that we have hacked
> the __os_install_post RPM variable so it basically doesn't do
> anything.
>
> However a better solution would be to run the correct strip binary
> depending on the type of binary/library (ie. ordinary strip or
> i686-pc-mingw32-strip as appropriate).
>
> If you have a look at this file (not written by me) and search down
> for __os_install_post, you'll see one working but rather ugly solution
> to this:
>
>  http://www.annexia.org/tmp/i686-pc-mingw32-binutils.spec
>
> Is it possible to do any better?  Perhaps by patching RPM itself?

Or just modifying a) the macro or b) the strip binary itself in the
environment you compile the packages in.

Namely
a) there is a macro __strip that you can try replacing with your own value.
b) rpm2git does some interesting things with patch, where it just has
rpmbuild run through a srpm, and then captures all the patch commands
with an alternate binary.  In order to use it, you put a special
version of patch somewhere, put it at the top of the path, run
rpmbuild, and then revert your path.  Seems like a 'support' rpm to
build mingw packages would be easier than patching the RPM source
itself.

-Yaakov