Details
Diff Detail
- Repository
- rL LLVM
Event Timeline
llvm/test/tools/llvm-objdump/X86/disassemble-implied-by-disassemble-functions.test | ||
---|---|---|
6 ↗ | (On Diff #199981) | Nit: I'm trying to fight against unnecessary whitespace in these blobs. Reduce the space between the tag and value to the minimum needed for the values to line up within the block i.e: Class: ELFCLASS64 Data: ELFDATA2LSB Type: ET_REL Machine: EM_X86_64 |
14 ↗ | (On Diff #199981) | Get rid of this, as it's not necessary for the test. |
15 ↗ | (On Diff #199981) | Let's reduce this right down to just a single nop instruction (0x90) or ret (0xc3). |
20–24 ↗ | (On Diff #199981) | You probably only need one symbol and section for this test case. The rest of the disassembly/disassemble-functions functionality (such as --disassemble-functions dumps only the specified symbols) should be tested separately (I actually am currently working on those tests, and hope to have a patch up ready by the end of the day). |
33 ↗ | (On Diff #199981) | I don't think you need to check this first bit. |
38–42 ↗ | (On Diff #199981) | You can probably afford to get away with just checking that the add1: and main: lines. |
llvm/test/tools/llvm-objdump/X86/disassemble-implied-by-disassemble-functions.test | ||
---|---|---|
6 ↗ | (On Diff #199981) | Should we modify obj2yaml to eliminate unnecessary whitespace? I'm an ELF noob so when I made this test case I started with a source (.c) file, compiled to an object (.o) file, and finally ran obj2yaml to generate the yaml file, which seems to always include extra whitespace. |
LGTM, with one minor comment outstanding.
Do you need somebody to commit it for you?
llvm/test/tools/llvm-objdump/X86/disassemble-implied-by-disassemble-functions.test | ||
---|---|---|
22 ↗ | (On Diff #200163) | You can use CHECK-NEXT here. |
6 ↗ | (On Diff #199981) | It might be good to do so. I don't know the background behind why it does that. For the record, I quite often copy an existing yaml blob and work from there. You may find that useful going forwards. |
Thanks @mmpozulp! I'll try to commit this later, but it might not happen until tomorrow. I've subscribed some others who have worked on the binutils, who might have time to land it before then.