Mark Polyakov

@luin can you confirm that all which would need to be changed is the build scripts? No js changes necessary?

You may need to install an older version of distorm3 for python 2

I believe section 5.4.2 of the paper directly addresses your point (1) @ecpeterson .

I ultimately agree that it doesn't seem to be a major issue -- usually the ZXZ decomposition is not applicable.

For anyone interested in how the compiler chooses which Euler decompositions to support: At the time the chip-spec is constructed, the compiler tries to find the best way to compile...

After thinking more, adding a `Y-to-ZXZ` compiler as Erik suggested seems the best solution in my eyes. However, since `Y-to-ZXZ` is actually a special case of the Euler ZXZ transform,...

While the output is just as short as the output from the original `COMPRESS-INSTRUCTIONS`, my implementation is not exactly functionally identical. The main difference is that the original requires all...

The purpose of this PR is not to change the paradigm of compression in quilc from chip-specific to generic. In fact, my goal is not to change the behavior of...

Thanks for the feedback @ecpeterson. I've been looking a little bit more into the speed of my implementation. Without a global queue, it does tend to create a greater number...

I just pushed a change to restrict the recursive queue sizes to 2 as discussed in my previous comment. IMO the PR is now ready to merge.