In the last mail Philip Blundell said: > >BUG IN DYNAMIC LINKER ld.so: ../sysdeps/arm/dl-machine.h: 450: elf_machine_rel > >: Assertion `! "unexpected dynamic reloc type"' failed! > > You've probably got some library or shared object that wasn't compiled with > -fPIC. Use "objdump --dynamic-reloc" and go hunting for PC24 relocations. Aaargh. > >Any thoughts, anyone? Is anyone able to build perl5.005_03_MAINT_TRIAL_6 > >on a glibc2.1 non ARM linux box to show whether this is ARMLinux at fault? > >[If it is demonstrably a more general problem then it's not (just) our bug] > > I've got glibc-2.1 based Intel, Alpha and Sparc machines. If you point > me at the appropriate source for perl I'll give it a go. ftp://plum.flirble.org/pub/languages/perl/CPAN/authors/id/GBARR/perl5.005_03-MAINT_TRIAL_6.tar.gz or any other CPAN mirror :-) perl developers would be very happy if you make ok (or make nok) depending upon success or failure of all tests. Should the gdbm test 12 fail or pass on glibc2.1 ? > BTW I built perl5.005_02 (using the SRPM from Rawhide) on my NetWinder > yesterday and got two `make check' failures -- lib/odbm (looks like the PIC > thing) and one other. No further on my other 2. If it's not ARMLinux at fault it may be perl's use of f(re)open(). [On Unix, fclose(stdin) followed by fopen() gets you the FILE* of stdin. This is an implicit assumption perl makes. I keep saying this, everyone ignores me] Nick unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]
