rebase

The order is always the same of a given mapping and op, but the accumulation may indeed introduce instabilities if an exact order is assumed (i.e. acc feeds into the chain rather than being just added at the end). We can make semantics more strict if needed.

I like where this is going a lot. Some suggestions on the doc though

added TODOs as requested

addressed comments

addressed comments, rebased

trusting that you will fix both typos, good to land

rebased

updated description, one more verify test

less code is better

Okay, friends of concise code, PTAL

addressed comments

Is there a way to remove your blocking review but without approving yet? I don't see that option here.
Anyway, please look at the other comments that are still open.

rebased with head

address comments

args in patterns

sync back with latest pulled master

note that the title still needs "[mlir] [VectorOps]"

Utils -> VectorUtils typo

Thanks! I was not sure on the naming, and like your suggestions better indeed. All addressed.

addressed comments

rebased with master

\o/ yeah, it has been a long road, but PTAL

synced with latest update

Hmm. I would argue that the tuple discussion makes less sense in the context of this particular rewriting (since it really deals with getting rid of extract_slices/insert_slices using other slice ops). Cleaning up tuple uses is a nice bonus, but not a requirement, since tuples are really a part of the vector dialect to start with, as the leaking example shows. If you want something that blocks any tuple uses, I would argue that we need another pattern population for that, and run that at places where we need to guarantee they are gone (such as the lowering to LLVM, although the "legality" part takes care of it there already).