[Python-Dev] Python build trouble with the new gcc/binutils
Fri, 16 Aug 2002 17:49:13 -0400 (EDT)
Martin> Andrew Koenig <firstname.lastname@example.org> writes:
>> Seems unlikely -- the system linker isn't in my search path
>> and "ls -ltu" shows that it hasn't been executed in 10 days.
Martin> Absence in the search path is irrelevant - gcc just knows that the
Martin> system linker is in /usr/ccs/bin/ld (see gcc/config/svr4.h).
But the system linker has not been executed in 10 days. So whatever
gcc might or might not know, that proves that gcc did not execute
/usr/ccs/bin/ld during any of this testing.
Martin> It would help enourmously if you'ld focus on a single failing
Martin> scenario, and, for this scenario, would confirm that the
Martin> binutils that you had at configuration time of your compiler
Martin> are also the ones that it uses at run-time.
Martin> To confirm the latter, it would help if you would report the
Martin> output that you get from adding "-v" to one of the linker
Martin> lines (e.g. for struct.so).
which I do how, exactly?