This has been complicated on #12, scoping it out to this issue
A pretty good initial bandaid on #12 has been to never allow a registereditor that is currently on focus to autorefresh.
Remaining question on desired behavior:
- Let's say you're writing on
@"
- You temporarily switch back to main buffer to do some unrelated editing. However,
dw will overwrite the @" buffer, that had unfinished edits in it.
- Is this okay, or not okay?
My point of view has been that if a single undo can get the previous state cleanly on the registereditor buffer, then that's ok, because doing large edits in registers is a very niche case.
But speaking of that, we had a problem with undoing resulting in a recursive loop
I believe this was also fixed by disallowing edits to a buffer to refresh itself?
First step here would be to see how well undo works in registereditor buffers, and whether there's a problem or if we already arrived at what makes sense.
This has been complicated on #12, scoping it out to this issue
A pretty good initial bandaid on #12 has been to never allow a registereditor that is currently on focus to autorefresh.
Remaining question on desired behavior:
@"dwwill overwrite the@"buffer, that had unfinished edits in it.My point of view has been that if a single undo can get the previous state cleanly on the registereditor buffer, then that's ok, because doing large edits in registers is a very niche case.
But speaking of that, we had a problem with undoing resulting in a recursive loop
I believe this was also fixed by disallowing edits to a buffer to refresh itself?
First step here would be to see how well undo works in registereditor buffers, and whether there's a problem or if we already arrived at what makes sense.