[Mlir-commits] [mlir] [mlir][bytecode] Add support for deferred attribute/type parsing. (PR #170993)

[Jacques Pienaar]

jpienaar wrote:

> Seems fine overall, but could deserve some more documentation as this is all non-trivial. Any place you can write the high-level process / logic?

I've been thinking about it ... its not really related to the file format so the bytecode doc not good place, it is rather opaque to user (just happens under the hood and nobody should ever need to know about it) so perhaps in the reader file is best place. Perhaps as part of the depth documentation.

https://github.com/llvm/llvm-project/pull/170993