[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: darwin library_names_spec
From: |
Albert Chin |
Subject: |
Re: darwin library_names_spec |
Date: |
Wed, 17 Dec 2003 23:03:12 -0600 |
User-agent: |
Mutt/1.4i |
On Mon, Dec 08, 2003 at 03:52:42PM +0000, Gary V. Vaughan wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Peter O'Gorman wrote:
> | Even if you keep the old versions around, the dynamic linker can not
> | find them, so they are never used unless you go playing with the
> | symlinks by hand, it will only load libfoo.5.dylib, which in the normal
> | scheme of things always points to the latest version of libfoo.5 that
> | you installed.
> |
> | It is quite likely that other systems also have unnecessary symlinks,
> | but I wouldn't have a clue as to which ones they are :)
>
> Quite. Now that you have pointed out the additional symlink on darwin, I
> can
> see the same problem on linux.
>
> Open questions to the list: What is the purpose of creating a
> libfoo.x.y.z.so
> when libfoo.so is enough for the compile time linker, and all but the newest
> libfoo.x.y.z.so are ignored by the runtime linker? Where does the
> libfoo.x.so
> symlink fit into this.
Depends on the SONAME. I think we only need libfoo.so and the file
that corresponds with SONAME.
Of course, I don't know if all linkers search for libraries by SONAME.
--
albert chin (address@hidden)