Remco Bloemen

> Are people really writing tests for their tests? > I also definitely don't think people are writing tests for their tests (or at least I hope not...) I can...

This line will also fail in the official Python docker image `python:3.6-alpine`. (See [issue](https://github.com/docker-library/python/issues/111)). The implementation of `find_library` is a hack. It calls `ldconfig -p` and parses the output, hoping...

I'm assuming this issue tries to solve the non-determinism problem like so: https://github.com/WebAssembly/design/issues/582#issuecomment-191318866 > For NaN bits, it'd be straightforward for a preprocessor to inject NaN canonicalization code after every...

> @barakman Another point to improve in this method - this time an actual performance improvement (and not just coding or documentation) - is that following this: I am aware...

> It is worth mentioning that briefly, for example, "Always >= 1, because denominator > 0". It is there in my blog post, which is linked from the code. Not...

> As for "If twos is zero, then it becomes one", it should actually be the opposite: "If twos is one, then it becomes zero". It should not, the comment...

> I would go further and use a new variable for this - something like twosInv, because there doesn't seem to be any good reason to "reuse" it (which just...

> Finally, a minor coding error is that the function is declared as returns (uint256 result), and then ends with return result. Matter of style; I strongly dislike implicit returns...

It's moved to a final `FROM scratch` container. Capabilities are preserved during this move since they are extended attributes. I need to do this in the build container because `setcap`...

I had not thought of adding an extra stage, but that would work too. Just `apt-get` installing it like above is also fine. I'm only proposing including it as a...