On Sat, 12 Feb 2005 22:38:24 +0100, Kai Gro�johann <[EMAIL PROTECTED]> wrote: > I mark a region, I invoke M-%. This will perform the query-replace on > the marked region only. (I forget whether that's always the case, or > whether I have turned on something to make it so. I think it works > like this because I use transient-mark-mode.) > > The problem is that the highlighting of the search string is performed > in the whole buffer, not only in the region where replacing is going > to happen.
Yes, I find that very disturbing as well. A particularly bad case is where you're replacing a very common string (like a small space) in a small region, in which case the bogus highlights can completely overwhelm the cursor display and region highlighting! Oh, also another problem with the "query-replace highlighting": The highlighting seems to be erased and re-done from scratch everytime you hit "y" to replace an instance, which can be very annoying. I suppose the intent of this is to keep any highlighting up-to-date in the face of buffer-modification, but it could be made much smarter (e.g., only "refresh" the buffer region actually modified by the last replacement). Thanks, -Miles -- Do not taunt Happy Fun Ball. _______________________________________________ Emacs-devel mailing list [email protected] http://lists.gnu.org/mailman/listinfo/emacs-devel
