Alexander Viand

> I _think_ all we need to do is clone Tensorflow's patch and apply it during our bazel workspace instantiation. It would also be prudent to include the patch application...

Hey @asraa, just checking if there's any update on this (that maybe hasn't made it to the public/open-source side yet)? :)

The current draft already looks very nice (modulo the TODO's you identified) - did you get a chance to work on those? Also, what do you think about adding a...

for future reference: merged as https://github.com/google/heir/commit/2fb64740a1c3478b2e77a9374e3b0d9c0bfd4af2

> This also reminds me that we should extend the pass that adds client helper encrypt/decrypt functions to support the rust codegen path. _Originally posted by @j2kun in https://github.com/google/heir/issues/897#issuecomment-2282367968_

I think some of the early compilers[^1][^2] looked at these kinds of rewrites, mostly from the perspective of reducing multiplicative depth rather than just number of multiplications, though. We evaluated...

> The intention here is to write some rewrite patterns for the intern project before doing larger optimizations, so I don't think we should hold off until we get optimizations...

Sorry, I didn't see this before replying on #837 so I'll just leave a link to my comment there: https://github.com/google/heir/issues/837#issuecomment-2257586402

@asraa In today's meeting, I said this only needed a small bugfix, because I only remembered that I still need to swap `bgv.negate %c` to `bgv.sub_plain %one, %c`, However, I...

I've cleaned up the diff and the commit history a bit (should be free of unrelated changes now) + updated the PR description above. Below, I'll try and summarize some...