In the case of large binaries previously you could end up getting
relocation failures. This mirrors ld64's ordering of these sections. I'm
not sure this solves all cases but it should help in some.
Details
Details
- Reviewers
int3 - Group Reviewers
Restricted Project - Commits
- rG3a8c6a3039af: [lld-macho] Fix __objc_stubs ordering
Diff Detail
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
Comment Actions
Thanks!
lld/test/MachO/x86-64-objc-stubs.s | ||
---|---|---|
24–28 | Ideally the test would've been written using substitutions instead of hardcoded addresses so we don't have this churn. It would probably also make the test more comprehensive. See start-end.s line 102 for an example of how to test RIP-relative addresses. I'm fine with this being done in some future commit though :) |
lld/test/MachO/x86-64-objc-stubs.s | ||
---|---|---|
24–28 | Updated these and verified they passed before and after. because of that maybe they're a bit less strict than they could be |
Ideally the test would've been written using substitutions instead of hardcoded addresses so we don't have this churn. It would probably also make the test more comprehensive. See start-end.s line 102 for an example of how to test RIP-relative addresses. I'm fine with this being done in some future commit though :)