On Mar 6,2003 21:45:17 +0100, Martin Costabel <[EMAIL PROTECTED]> wrote :

>But I suspect that as soon as you start compiling things under 4.3.0, 
>there will be conflicts of the libpng/libpng3 variety (one version of 
>library loaded directly, another one indirectly -> crash)

Well, I hope not.  I still don't fully understand what is happening with the
libpng/libpng3 business, but I strongly suspect the following scenario.  You
ask fink to install "foo" and it also requires "bar" so fink is going to
install both at once.  Let's say foo depends on libpng3 and bar depends on
libpng.  What does fink do?
  1) First loads libpng3 because thats what foo wants
  2) Next loads libpng because thats what bar wants
  3) Successfully compiles bar, linked to libpng.
  4) Manages to compile foo, with libpng still installed, because the
dependecies are not checked a second time just prior to building.

As far as I can tell, you would be able to build foo with libpng headers
installed, just fine.  Everything will have libpng support compiled in, and
then at link time, you are linking against libpng3.dylib and all symbols are 
successfully resolved, so everything is still fine.  It is only at
execution time that the mismatch is detected.

Does this scenario make sense to anyone else?  Has anybody seen a symptom
which is not consistent with this?

  -- Dave





-------------------------------------------------------
This SF.net email is sponsored by: Etnus, makers of TotalView, The debugger 
for complex code. Debugging C/C++ programs can leave you feeling lost and 
disoriented. TotalView can help you find your way. Available on major UNIX 
and Linux platforms. Try it free. www.etnus.com
_______________________________________________
Fink-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-devel

Reply via email to