mån 2009-03-16 klockan 10:49 -0600 skrev Alex Rousskov:
> I agree with your logic and you are right that we already have virtual
> functions that should be used from the main program. I am guessing that
> the random advice that inclusion of virtual functions will help things
> was either flawed or there is some other problem we need to fix first.
>From what I can tell virtual functions is a pretty much a non-issue and
not related to the linking problem. If there is code depending on the
class the virtuals will be brought in by reference from the class,
including any instantiation found in those compile units (.o files).
the problem is compile units with no references to needing it. With the
objects thrown into library archives the linker MAY select to only link
those objects having direct symbolic references, just as done for any
other library.
In the past GCC (or binutils) has differentiated between .a archives
specified by path (fully linked) and .a archives used as libraries (-l,
only needed parts linked). But this is no longer done.
> One potential problem I am investigating is that convenience libraries
> that are linked with executables behave differently from same
> convenience libraries linked with other libraries.
What do you mean?
Regards
Henrik
Received on Mon Mar 16 2009 - 23:29:01 MDT
This archive was generated by hypermail 2.2.0 : Tue Mar 17 2009 - 12:00:04 MDT