Case insensetive search is not available anymore in non-unicode locale

[andrew_b]

This bug was introduced in [1a1496fc0dc13998803be9caf149ce3be8af9206].

How to reproduce.

1. Switch to the any 8-bit locale (KOI8-R as example).
2. Create text file with non-latin symbols in any charset (utf-8, cp1251, koi8-r, etc).

For example, cyrillic text:

ШШ
шш
шШ

where Ш is the upper-case letter (capital), ш is the lower-case one.

3. Open this file in internal editor or viewer.
4. Press F7 and enter Ш in the input line. Press Enter.

Result: only Ш is found, ш is not found at all.
Expected result: ш is as found as Ш.