[PATCH] D76903: Add a flag on the context to protect against creation of operations in unregistered dialects

[Uday Bondhugula via Phabricator via llvm-commits]

bondhugula added a comment.

In D76903#1948918 <https://reviews.llvm.org/D76903#1948918>, @mehdi_amini wrote:

> (I don't even know why we allow such operation name in the first place to be honest, other than saving a few chars when writing tests there isn't much interest)


One advantage has been in writing test cases where you need an op with a side-effect so that it just doesn't become dead code (when canonicalization is being run as part of the pass/utility you are testing). These unknown ops provided the easiest/shortest IR for it - we would have otherwise had to think of another side-effecting op (like creating a memref type and a load/store on it). Any other quick shortcut replacements?

Another use case of allowing such unknown ops is a need to continue / allow transformation in their presence before the output goes to another tool that is able to finally deal with those ops. Why force a dialect to be registered when an mlir-opt transform can meaningfully work in the presence of a mix of registered and unregistered ops?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D76903/new/

https://reviews.llvm.org/D76903