Bug when working backwards in multiple-cursors

[obar]

This is a brilliant interface! Thanks for the great package.

It even works with multiple-cursors… except when extending backward. Not a deal-breaker as one can move the cursors to do it if need be, but thought I should record this bug.

Here's a demo of what happens with three lines, and three cursors at the start of 'over'.

The quick brown fox jumps ^over the lazy dog.
The quick brown fox jumps ^over the lazy dog.
The quick brown fox jumps ^over the lazy dog.

Then type this sequence:
M-u M-u M-u C-b

The quick brown fox jumps OVER THE LAZY^ dog.
The quick brown fox jumps OVER THE LAZY^ dog.
The quick brown fox jumps OVER THE LAZY^ dog.

M-l

The quick brown fox jumps OVER THE lazy^ dog.
The quick brown fox jumps OVER THE lazy^ dog.
The quick brown fox jumps OVER THE lazy^ dog.

M-l

The quick brown fox jumps OVER the lazy^ dog.
The quick brown fox jumps over THE lazy^ dog.
The quick brown fox jumps OVER THE lazy^ dog.

It looks to be caused by the use of fix-word--quick-fix-times as a global variable. Perhaps there would be a way to increment that count only once, or divide by the number of active cursors, or look into how smartparens handles sp--mc/cursor-specific-vars… haven't dug into this too much, as I said any case switching can be performed in other ways right now.

[mrkkrp]

I'd accept a PR fixing this, but I myself do not use multiple cursors and do not have time to experiment with this.