Help: Unwinding the C++ stack...throw, longjmp & threads
George T. Talbot
george@moberg.com
Tue Aug 24 13:57:00 GMT 1999
Ulrich Drepper wrote:
>
> "George T. Talbot" <george@moberg.com> writes:
>
> > I don't know. Mr. Drepper, can you tell us what was measured? Was this
> > with the new exception model or the old setjmp()/longjmp() exception
> > model?
>
> I don't remember. One of the things which was measured for sure is
> code/data size and this went up.
OK...I just patched glibc 2.1.1pre2 (the one that comes with Mandrake
Linux)
so that -fexceptions would be added to C compiles. The size went from
1.2MB
to 1.4MB. Is that an acceptable size increase? I don't have enough
experience
with glibc customers to guage that. The additional size increase
appears to
be from .eh_frame sections.
As to performance, do you have a standard method of measuring overall
performance? I noticed that the testsuite directory was empty, with a
notice
about copyright problems. Is the testsuite used to measure
performance? If
not, what would I use to measure performance and whether/how much
-fexceptions
changes performance?
By the way, I'm using egcs 1.1.2 for now. Eventually, when I have a
version of
glibc that can be compiled with 2.95.1, I intend to recompile with that
for
comparison.
--
George T. Talbot
<george@moberg.com>
More information about the Libc-alpha
mailing list