tbielawa / bitmath

Python module for representing file sizes with different prefix notations
http://bitmath.readthedocs.org/en/latest/
MIT License
93 stars 25 forks source link

Project isn't dead! bitmath 2.0.0 coming soon! #99

Open tbielawa opened 1 year ago

tbielawa commented 1 year ago

2.0.0 is coming and it will be python 3 only. Keep up with progress on this branch as I work up to the release.

Watch the news file for specifics. https://github.com/tbielawa/bitmath/compare/master...2023-01-26-no-more-py2#diff-61d113525e9cc93565799a4bb8b34a68e2945b8a3f7d90c81380614a4ea39542R25

PR's welcome. Existing PR's will get reviewed to see how they fit into the future direction of the project.

tbielawa commented 1 year ago

Docs are building on read the docs now: https://bitmath.readthedocs.io/en/2023-01-26-no-more-py2/index.html

FelixSchwarz commented 1 year ago

This is great to hear. One thing I was recently confused about is parse_string_unsafe: "unsafe" means to me that should not use this on untrusted inputs (security) but it seems the function just uses some heuristics/there is some ambiguity. It would be nice if the name reflects that in 2.0. Using an "unsafe" function will be flagged in every code review and will cause the reviewer to dig into the implementation to verify that this is really "safe" to use.

tbielawa commented 1 year ago

This is great to hear. One thing I was recently confused about is parse_string_unsafe: "unsafe" means to me that should not use this on untrusted inputs (security) but it seems the function just uses some heuristics/there is some ambiguity. It would be nice if the name reflects that in 2.0. Using an "unsafe" function will be flagged in every code review and will cause the reviewer to dig into the implementation to verify that this is really "safe" to use.

That's a very smart idea. I'll add that to the list

tbielawa commented 1 year ago

This is great to hear. One thing I was recently confused about is parse_string_unsafe: "unsafe" means to me that should not use this on untrusted inputs (security) but it seems the function just uses some heuristics/there is some ambiguity. It would be nice if the name reflects that in 2.0. Using an "unsafe" function will be flagged in every code review and will cause the reviewer to dig into the implementation to verify that this is really "safe" to use.

Can I have your thoughts on this?

https://github.com/tbielawa/bitmath/issues/101

tbielawa commented 5 months ago

Thanks for your patience everyone. 2023 was especially busy for me and I am just now getting settled into a place where I can get back to work here.