Re: Wgcc status report

[Ralf Wildenhues]

Hello Markus,

I would prefer if you did not send messages to the libtool list twice,
did not send messages including large patches to the libtool list
(that's what the libtool-patches list is for), and I would also
personally prefer if your announcement frequency were more that of an,
err, announcement.  Users interested in every next step can still read
your interix-wgcc-developer and/or the libtool-patches list.

* Duft Markus wrote on Tue, Sep 26, 2006 at 09:44:10AM CEST:

> FAIL: build-relink.test
> FAIL: build-relink2.test
> FAIL: mdemo2-exec.test

> This looks *very* cool ;o) For the tests failed: hardcode will never
> work, thats clear (fails with gcc too).

Wrong.  The hardcode test tests that the settings in libtool.m4 conform
to expected behavior.  The settings are false, thus the failure.

> I now changed the hardcode.test to skip on all interix platforms for
> now.

So you paper over the issue.

> The other three, i really have
> no idea there. Mdemo2 fails only half, as far as i've seen the static
> version executes OK. I think the error is, that libltld allways tries to
> dlopen the static library, not the shared one, but i can't figure it out
> (i hate dlpreopening ;o)) (and i can't read my own changed code
> anymore... *argl* ;o))
> 
> I also tested gcc, to see if my changes influenced it somehow. Only
> hardcode.test failed, and thats ok (it does without my patch too) ;o).

If you desire to get your changes into GNU Libtool, you should strive to
- keep system-specific changes out of ltmain.sh, and out of all the
  tests as much as is sanely possible.  Remember the tests should also
  demonstrate what users are expected to do, so work hard to keep the
  number of preprocessor conditionals down to a minimum; and to
- not break any other system with your patches; that is a show-stopper.

Using `uname' in ltmain.sh is wrong, that breaks for cross-compilation.
Some of the comments in the patch you posted are wrong (they are copy
and pasted from elsewhere, but do not match the pasted code).

I would like to encourage you to port your patch to CVS Libtool, as it
has a better testsuite.  In case that is too much work for you, you may
anyway want to try its testsuite with
  ./tests/testsuite LIBTOOL=/path/to/1.5.22/build/libtool -k libtool \
                    -k stresstest

Hope that helps.

Cheers,
Ralf