Rather than depending on the hex dump from obj2yaml. Now the test shows the
expected function body in a human readable format.
Details
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
lld/test/wasm/data-segments.ll | ||
---|---|---|
33 | It looks like most of the tests in test/ELF use -d --no-show-raw-insn .. but I guess --no-leading-addr is good too. Maybe both ? |
- Add --no-show-raw-insn
lld/test/wasm/data-segments.ll | ||
---|---|---|
33 | Oh nice, I had missed that option. This is much better now. |
lld/test/wasm/data-segments.ll | ||
---|---|---|
171 | Kind of shame there are so many differences here between the PIC/no-PIC and 64/32-bit. Do you think this more useful than just having completely separate blocks (with duplication)? I wonder if you could do -DINT=i32 / -DINT=i64 to remove at least one level of splitting here? | |
227 | We weren’t able to create the releas |
lld/test/wasm/data-segments.ll | ||
---|---|---|
227 | Oops, somehow this much have been in my paste buffer.. please ignore. |
lld/test/wasm/data-segments.ll | ||
---|---|---|
171 | I removed one level of splitting, as you suggested. Let me know if you think it looks reasonable, otherwise I can fully duplicate the function between PIC and non-PIC code as well. |
It looks like llvm-objdump doesn't yet handle --disassemble-all (it seems to think all the different sections contain functions), so the YAML still seems good for making sure the rest of the file structure is correct.
It looks like most of the tests in test/ELF use -d --no-show-raw-insn .. but I guess --no-leading-addr is good too. Maybe both ?