:reload should fail if unsaved changes

[vwkd]

I'd like to suggest the reload command to fail if there are unsaved changes and adding an explicit reload! that discards unsaved changes, similar to q or bc.

Currently, reload discards changes anyway.

Sometimes it's useful to reload a file only if there aren't any changes. For example, when changing the underlying file through another source I reload to not revert the changes on the next save in helix. However, if I have unsaved changes, I'd like that helix prevents me from reloading so I don't lose the changes. I know I can undo, but sometimes the changes are not visible in the current window. An explicit confirmation would be better.

[David-Else]

I think your request relates to: #4734

[ninevra]

I think your request relates to: #4734

I think it's kind of the opposite? #4734 is about preventing :write from overwriting changes made to disk by other programs, whereas this is about preventing :reload from overwriting changes made in helix' memory and not yet written to disk.

[ninevra]

This would be doubly useful with :reload-all.

I end up using :reload and :reload-all a lot, e.g. after git checkout, to avoid accidentally later working on a buffer that's out of date. After git checkout, if there's a buffer with unsaved changes, it's either an untracked file or something I forgot to save & commit, so I really don't want to accidentally lose that work; but at the same time, there may be many open buffers, and I really don't want to leave one open with another branch's state in it.

:reload and :reload-all failing or prompting instead of overwriting unsaved buffers would be very useful.

[David-Else]

I think your request relates to: #4734

I think it's kind of the opposite? #4734 is about preventing :write from overwriting changes made to disk by other programs, whereas this is about preventing :reload from overwriting changes made in helix' memory and not yet written to disk.

I meant it is in the same genre of feature, preventing unwanted data loss. I was probably being too general and not much help, soz!

[ninevra]

I meant it is in the same genre of feature, preventing unwanted data loss. I was probably being too general and not much help, soz!

Np! Cross-referencing is great.

[Pablu23]

Ill try to do this

[Pablu23]

I opened a pr for this

[askreet]

@Pablu23 Can you mention this Issue # in your PR (as #5680)? That helps the GitHub UI show which issues have related PRs.

[Pablu23]

Ahh yes I just did it, this is my first time contributing so I didnt know that. Thanks :)

[gozes]

hmm, I think this may be related to something that bit me today where I put something up for PR, committed some suggestions from the GitHub UI, and then pulled so that I could address other issues but Helix didn't reload the file/buffer with the local changes so I ended up losing some work