old-cross-binutils/ld/testsuite/ld-elf/endsym.d

#source: start.s
#source: endsym.s
#ld: --sort-common
#nm: -n
#notarget: hppa*-*-hpux*
#xfail: m68hc1*-* xgate-*

#...
.* end
#...
.* end2
#...
.* _?_end
#pass
Don't PROVIDE over top of common symbols This: int end[100000]; int main(void) { end[99999] = 0; return 0; } should not segfault. ld/ * ldexp.c (exp_fold_tree_1 <etree_provide>): Leave bfd_link_hash_common symbols alone. ld/testsuite/ * ld-elf/endsym.s, *ld-elf/endsym.d: New test. 2014-12-22 00:19:23 +00:00			`#source: start.s`
			`#source: endsym.s`
			`#ld: --sort-common`
			`#nm: -n`
			`#notarget: hppa--hpux*`
m68hc11/12 and xgate config.sub weirdness Oddly, config.sub converts a duple ending in -elf for these target to -unknown-none, which means they aren't seen as elf targets by binutils. So, counter that. This exposes a number of testsuite issues (ones you would have seen if configuring with a full triple, say m68hc11-unknown-elf). binutils/ * testsuite/lib/binutils-common.exp (is_elf_format): Return true for m68hc11/12 and xgate triples. gas/ * testsuite/gas/cfi/cfi.exp: Exclude m68hc11/12 from m68k test. ld/ * testsuite/lib/ld-lib.exp (check_shared_lib_support): Exclude xgate. * testsuite/ld-elf/endsym.d: xfail m68hc11/12 and xgate. * testsuite/ld-elf/pr14156a.d: Likewise. * testsuite/ld-elf/pr14926.d: Don't run for m68hc11/12 and xgate. * testsuite/ld-elf/sec64k.exp: Likewise. 2016-01-17 01:43:43 +00:00			`#xfail: m68hc1- xgate-*`
Don't PROVIDE over top of common symbols This: int end[100000]; int main(void) { end[99999] = 0; return 0; } should not segfault. ld/ * ldexp.c (exp_fold_tree_1 <etree_provide>): Leave bfd_link_hash_common symbols alone. ld/testsuite/ * ld-elf/endsym.s, *ld-elf/endsym.d: New test. 2014-12-22 00:19:23 +00:00
			`#...`
			`.* end`
			`#...`
			`.* end2`
			`#...`
			`.* _?_end`
			`#pass`