diff options
| author | Alan Modra <amodra@gmail.com> | 2024-07-27 09:13:27 +0930 |
|---|---|---|
| committer | Alan Modra <amodra@gmail.com> | 2024-07-29 20:01:30 +0930 |
| commit | 6c41f7a812d843d6ba76b65ec7e66074fbfed041 (patch) | |
| tree | 00d3c979221247192aa4904b39c555f27b187b15 /ld/ldfile.c | |
| parent | fb2a924990d3595c4b39cbcd37ebc36df761a3ad (diff) | |
| download | gdb-6c41f7a812d843d6ba76b65ec7e66074fbfed041.zip gdb-6c41f7a812d843d6ba76b65ec7e66074fbfed041.tar.gz gdb-6c41f7a812d843d6ba76b65ec7e66074fbfed041.tar.bz2 | |
Fixes to "PR 31728 testcases"
This brings us down to just these fails for the set of targets I
usually test when making testsuite changes.
aarch64-pe +FAIL: ld-pe/symbols-ordinals-hints-imports-ld
arm-pe +FAIL: ld-pe/symbols-ordinals-hints-exports-dlltool
arm-pe +FAIL: ld-pe/symbols-ordinals-hints-imports-dlltool
The aarch64 one is likely due to the target missing support somewhere.
It is fairly new, I haven't investigated. The arm-pe fails are due to
arm-pe being a target that adds underscores to symbol names (see
config.bfd) whereas dlltool thinks it does not (see
dlltool.c:asm_prefix). arm-wince-pe on the other hand doesn't add
underscores. I would guess the right fix for dlltool is to get this
symbol info from bfd using bfd_get_target_info.
Note I'm not very happy about the creative use of ld_after_inputfile
in symbols-ordinals-hints-imports-ld.d, which is likely to break with
some future run_dump_test change.
Diffstat (limited to 'ld/ldfile.c')
0 files changed, 0 insertions, 0 deletions