You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I had refactored PG's behavior w.r.t. C-_ a long time ago (which was fairly broken before my patches)
There is a limitation of the current algorithm of "pg-protected-undo":
one idea of this feature is that PG retracts by defaults, except when one edits just inside a comment
this idea applies as well if we kill a region within a single comment
but the implementation of this idea was done so that it just checks whether the first char and the last char are within comments
as a result, the limitation is that "if we kill a region that spans over two distinct comments", then it doesn't retract (while it should, as we removed text from the comments and some *) (* chars...)
My suggestion in Issue#800 was probably naive. I think the problem is more subtle because
when we undo a deletion of "foo ) bar ( baz" it means we'll add this string, but we can't figure out whether
this undo will change non-comment code without a much finer analysis (in the general case we'll need to compare the code after the undo).
I had refactored PG's behavior w.r.t.
C-_
a long time ago (which was fairly broken before my patches)There is a limitation of the current algorithm of "pg-protected-undo":
FTR: see also #800
The text was updated successfully, but these errors were encountered: