An analysis of return-in-finally in the wild

[iritkatriel]

@guido

I do agree that maybe doing the same (hopefully mostly automated) analysis on a similar amount of random GitHub repos using Python would be nice to confirm what the “average” Python coder does.

I wrote a script to select 20K random packages from @sethmlarson’s db, download them and analyse.

There were in total 77,398,892 lines of code with 316 instances of return/break/continue in finally. So about 4 per million lines of code.

I also got this comment on twitter, which seems relevant:

ᅟ