bfredl

bfredl.bsky.social

27 September 2026 22:23

Back in the day, I ASKED for clarification in their github issue if the situation was just annoying (regrettable, but needed to fix other annoyances) or if we actually corrupted user data. They soundly said, no, no corruption only annoyance. But whatever I guess.

bfredl

bfredl.bsky.social

27 September 2026 22:23

Emacs and vim do not read each other's state files, and they use different locations to not mess with each others state files. but maybe they are also violating someone's tenth law something.

bfredl

bfredl.bsky.social

27 September 2026 22:23

#Neovim obviously uses a different LOCATION for undofiles than vim by default. You only get in trouble if you SPECIFICALLY ASK vim and neovim to mess with each other by using the same cache directory. Even then vim and neovim shares the same code which SAFELY refuses to touch incompatible files.

bfredl

bfredl.bsky.social

27 September 2026 22:22

Ugh, a thing I don't like with working in FOSS is when people just casually lie about things I have said as a maintainer. No, #neovim does not delete #vim undofiles No, neovim does not overwrite vim undofiles, No, neovim does not "corrupt" vim undofiles I confirmed all of this with the user.

Neovim

neovim.hachyderm.io.ap.brid.gy

27 September 2026 22:21

There has been the need to clarify some rumors that have been circulated on social media: Yes, #neovim uses an different undofile format than #vim. In an ideal world we would not need this, but the way neovim implements buffer properties is completely different than vim, and this state is part […]