Comment 24 for bug 67844

This is definitely bug of this peace of software, that's why:
if you doing some action with user data - do it best possible way OR warn user about low quality / no guaranty of this action.

In this case if gedit so stupid, that can't use already existing library from kedit to guess file encoding, it should't try to do it at all, just show to user dialog with encoding selection, instead of give dummy result, that definitely useless at all.

As Alexander A. suggested workaround is:
open in gconf-editor:
/apps/gedit-2/preferences/encodings/auto_detected
and put cp1251 or other your native most common encoding to the top.
This stops works with retarded KOI8-R, but UTF-8 will continue work as usual.

PS: I'm not using linux for a wile exactly for this reason, usability level here sometimes good, sometimes terrible and nobody cares - instead of fixing usability bugs before release new features with same mistaces (or simpler just bugs) it storing in long-term wish-list, es example - Evolution.