[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: detect identical deplibs
From: |
Peter O'Gorman |
Subject: |
Re: detect identical deplibs |
Date: |
Fri, 30 Sep 2005 23:28:27 +0900 |
User-agent: |
Mozilla Thunderbird 1.0.2 (Macintosh/20050317) |
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Ralf Wildenhues wrote:
| Current libtool (both branch-1-5 and CVS HEAD) does not detect identical
| deplibs in all cases:
| sub/liba.la
| ../sub/liba.la
| ../../sub/liba.la
| $top_builddir/sub/liba.la
| $abs_top_builddir/sub/liba.la
| -Lsub -la
| /sub/.libs/liba.a
|
| or even referenced through symlinks, may all refer to the same library.
| I'm not certain that we're required to detect all of these cases. We
| not not recommend to use the last two for in-tree deplibs and in fact,
| we recommend using relative paths in all cases, but they can still break
| down (I believe one of them is converted to absolute and the other is
| not, while comparing).
|
| The patch below does not change libtool behavior but adds a test
| exposing the issue. Fixing the actual failure may be a bit hairy, and
| I'm not sure we should fix the `-Lsub -la' case, so I'm not going for it
| right away.
|
| OK to apply?
I thought we'd agreed to avoid having to use autotools if possible in the
new test suite. I don't see a requirement in these tests for
automake/autoconf...
Peter
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (Darwin)
iQCVAwUBQz1Li7iDAg3OZTLPAQIfpgP/WdUfnF3UFtYUNrnUUojU/C0fSGFEs6Vi
Ox1C4lZfdVxhkeHB5ZFapN4lmx2fCxw2x9hLaM0GfvHHOvz7HzMNsM6Y+hfODeGB
K8D9iMFKJsbjZm64nuVCSZMZ65OTl34J5xxq09aG7tJqCTYnuWskLtTMI+ocFeqv
Oc7qMUt6AOc=
=4RzB
-----END PGP SIGNATURE-----