This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: rfc: limited shared libgcc widgetry


On Tue, Sep 26, 2000 at 04:53:14PM +0100, Joseph S. Myers wrote:
> The patch includes
> 
> > + SHLIB_LIBS = -lc
> 
> How does this work when the soname of libc changes, users have binaries
> linked against the old shared libc and a shared libgcc, and GCC is
> upgraded, the new shared libgcc being built against the new libc?

You have to keep the old shared libgcc around somewhere.

Libc soname bumps are extremely rare, as you have to recompile
*everything*.  I wouldn't worry about it too much.

> If libgcc will include symbols that require the C++ compiler to be built,
> then shared libgcc should probably be disabled if --enable-languages is
> used to disable the C++ compiler; I don't think the symbols contained in
> a shared libgcc should depend on the languages chosen as that would be a
> recipe for incompatibilities.

It'd be simpler and more reliable to take the damn C++ code out of
libgcc and put it in libstdc++ where it should have been since day
one.

Yes, I know the argument about not requiring people to use libstdc++,
and I don't buy it.  C, Fortran, Java, Chill, Ada, etc. all insist you
use the language runtime library.  Why should C++ be different?

zw

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]