Dynamic string search algorithm selection

[dg-pb]

Feature or enhancement

Proposal:

There is still a bit of ground to cover, but current results indicate that this could turn out to be fruitful.

The main idea is to dynamically switch from Horspool to two-way algorithm.

Below is the performance comparison of current string search versus current state of proposed approach.

The average percentage difference is close to 0, but it is still work in progress.

Has this already been discussed elsewhere?

I have already discussed this feature proposal on Discourse

Links to previous discussion of this feature:

https://discuss.python.org/t/optimize-str-find-aka-fastsearch/54483/13

[sweeneyde]

Let's discuss the ongoing proposals for fastearch.h here: #119702