Open Source Realtime Voice Changer
github.com
github.com
1) If this is not meant for just a Japanese audience, then at least have an obvious English translation.
Google translate is doing an OK job, so at least put that up (just indicate machine translation).
2) Directing people to download from Google Drive or Hugging Face seems less than ideal.
SourceForge (https://sourceforge.net/) might be another acceptable option. Both because it can be used to mirror what is at GitHub and can be used to host larger files (though still a depends type of thing). At the very least, they are more well known and trusted.
Clear labelling of anti-features (like “relies on non-Free network services”) really is a killer feature of F-Droid. I hope Flathub copies.
I'll frequently commit 40 disparate changes under "stuff" in my personal projects. It's never been a problem. No real need to LARP a big software project with many developers.
The commit message tells you what you meant it to do. When fixing problems in complex systems this can be vital.
Git based workflows were not the missing technology that finally unlocked the development of complex systems. We wrote complex code before we had git annotate, even before version control systems were widely adopted.
Source: I have had commits rejected on code review for insufficiently descriptive commit messages in RCS, SCCS, Visual SourceSafe and CVS and probably others before git was first invented.
What is a sane practice varies immensely with the scale and scope of the project. Many things that are nearly mandatory for large scale development are pointless obstacles in small scale development.
If anything, I think that's more symptomatic of poorly structured work and questionable planning.
Also, Google Translate works quite well on GitHub: https://github-com.translate.goog/w-okada/voice-changer/blob...