Oscar Benjamin

I suggest opening a pull request that can use DomainMatrix for the case of rational coefficients only. I think that it would be better to write new separate functions rather...

It is not clear to me what the intended purpose is here. Is it just to display the polynomial? Or is there some expectation of doing algebra with the polynomials?...

I think that the `2` should reduce to zero. Otherwise in general I don't know what is expected here because I have never used Zhegalkin polynomials. Someone needs to spell...

Since Zhegalkin polynomials have a finite number of possible monomials and all coefficients are either 1 or 0, one way to represent them for calculations would be as an integer...

Alternatively a sparser representation could just be as a set of terms. Then and is intersection and xor is the symmetric difference.

Bisected to aa54a252165881dbb6405423369092465d6746d7 from gh-27693. CC @krishnavbajoria02 @TiloRC

> end users trying to install snakemake via pip etc. cannot be expected to dig up issues, apply PRs and then install software from source. Indeed. I was evaluating snakemake...

It looks like datrie is used in these two places: https://github.com/snakemake/snakemake/blob/4baebfb9e212070a182a53ac363464ab1676d281/snakemake/output_index.py#L11 https://github.com/snakemake/snakemake/blob/4baebfb9e212070a182a53ac363464ab1676d281/snakemake/path_modifier.py#L28 As far as I can tell in both places the trie is constructed once and then later searched...

> I used `[::-1]` to iterate a Python list in reverse, but this created a copy As long as you avoid something like this I don't see how using bisect...

Looking at the changes in gh-28337 it seems that most are just error messages in exceptions. From a performance perspective it is problematic to format exception messages eagerly especially if...