DavidSpickett wrote:

I will be able to approve this from a general perspective but want at least one 
RISC-V related person to approve too.

> One case we probably need to address is that the arch tag in the attributes 
> can contain conflicting extensions, as it is the union of the arch tags from 
> all objects. We probably want a test for that - and we need to be careful 
> that the resulting disassembly is not prioritising one of the conflicting 
> extensions over the other (which could be misleading). I don't mind if 
> addressing this is a follow-up.

I recommend adding a test that shows the problem (meaning: it passes when the 
problem happens). Then doing another PR to fix the issue.

This is done with code generation a lot, to cut down the diff in the PR that 
fixes the bug. It's somewhere in the developer guide but I can't find a link 
right now.

https://github.com/llvm/llvm-project/pull/173047
_______________________________________________
llvm-branch-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits

Reply via email to