[PATCH] D154958: [RISCV][MC] Relax conditional branches to unresolved symbols

[Alex Bradbury via Phabricator via llvm-commits]

asb added a comment.

Just to note that @jobnoorman and I spent some time discussing this earlier in the week and I was pretty convinced the current behaviour isn't what is desired (especially for the second bullet point in the patch description). In D108961 <https://reviews.llvm.org/D108961>, @reames made a good argument for supporting this relaxation from the perspective of improving toolchain resilience (in particular, where we fail to accurately estimate the upper bounds of of instruction sizes). I think that applies here just as well.

Flagging as putting the code review aspect aside, it would be useful to figure out early if people feel strongly that the intent of this patch is wrong.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D154958/new/

https://reviews.llvm.org/D154958