Sebastian Thiel

> With the current implementation performance will slightly drop with each supported language. But last time I compared tokei and gengo on a clean clone of the [linux repo](https://github.com/torvalds/linux), they...

> Thanks for the research! I hadn't put much thought into memory usage. Although I was initially surprised, it makes sense. While I was trying to avoid generating code, memory...

There is now [a PR](https://github.com/spenserblack/gengo/pull/157) over at `gengo` which fixes all performance issues to the point where `onefetch` simply cannot get around using `gengo` anymore. As an example, `WebKit` takes...

> Love this application. :) Thanks :)! As of now, there is no way to configure the keybindings of `dua i/interactive`. However, while that is the case, you could probably...

This might happen sooner than later as I plan to make a few performance improvements.

Thanks a lot for reporting, a very interesting case! I agree that more threads at merely 500 entries/s shouldn't be able to hog a whole CPU. Profiling would be needed...

For posterity, here is a profile run of a `dua -t2` run. Most of the time spent is in the kernel, and if it is to be believed it's mostly...

That's amazing! > I can tell it is really building from commits more recent than the most recent release because, for example, [the PDF](https://gitpython.readthedocs.io/_/downloads/en/latest/pdf/) has a section on "Top-Level" in...

Sounds good to me, thank you!

Thanks for reporting. I presume that trying to set `allow_unsafe_protocols=True` when making the call to GitPython, whichever it is, isn't possible? It seems you have control over the GitPython dependency,...