[Top][All Lists]

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

Re: Interix Shared Libraries

From: Ralf Wildenhues
Subject: Re: Interix Shared Libraries
Date: Wed, 5 Jul 2006 09:26:52 +0200
User-agent: Mutt/1.5.11-2006-06-19

* Duft Markus wrote on Tue, Jul 04, 2006 at 05:14:50PM CEST:
> I ran some tests now ;o) The thing is, that i left libtool as it was
> (my patches applied) when i built it using wgcc (i normally use gcc to
> build it, since i don't need libltdl, so it doesn't really matter), so
> the configure used hadn't got
> My changes in it (it tried updating by itself, but it seems i'm
> missing the right autotools versions ;o)). Some tests failed, the
> first some with a segfault, and the latter ones, which try to exec
> something simply don't find their .dll file, because the path (really
> PATH on windows) has to be set to point to the dll. Then again some
> segfaults, but i'll look into this ;o)
> I'll tomorrow manually bootstrap libtool with itsself, so i get wgcc
> in... Maybe then more tests succeed.

Please test libtool with all your changes applied in any case; after
all, we are interested in how all cases work with those changes.  The
README file has a section "The Test Suite" which tells you how to report
failed test group output verbosely, please do that.

> Testing wgcc on f77 is a little unfair ;o) wgcc is C and C++
> (everything M$'s cl.exe can do)...

Fine, then we don't care if the Fortran tests fail.

If you post large output, please pack it with bzip2 or gzip.  Thanks.

> Heres the output for wgcc:

> ====================================
> 37 of 95 tests failed
> (17 tests were not run)

This needs more work.  The verbose output will hint as to what kind of

> And here goes the output for gcc, which still did a little better ;o):
> ======================================================================

> FAIL: hardcode.test

> 1 of 112 tests failed

Probably fixed by setting hardcode_shlibpath_var to unsupported (the
default).  But again I can tell this only by seeing the verbose output.
And for changes to this I'd need to look at system documentation.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]