Bugzilla

Ignoring Klaus B's issues for now, this fixes the save-as problem I mentioned in comment 11. Whoever wrote the original code seemed a bit uncertain if the encoding is UTF8 and it turns out it's now not. It is more like single byte ISO-<whatever> where the character is encoded in a single non-ascii byte like octal 344 in this example:
```
Alt-d   \344   228   E4  U+00E4  &auml;     ä  small a, umlaut mark
```
The changes I made here were actually in my obsolete patches from when I added UTF8=ACCEPT but, for unknown reason, I took them out. (Probably because I didn't test save-as, copy and view source on a UTF8 named folder so I missed the effect this has.)

I tested this on top of Ping's patch so I assume it is also needed to fix all the problems.

Ignoring Klaus B's issues for now, this fixes the save-as problem I mentioned in comment 11. Whoever wrote the original code seemed a bit uncertain if the encoding is UTF8 and it turns out it's now not. It is more like single byte ISO-<whatever> where the character is encoded in a single non-ascii byte like octal 344 in this example:
```
Alt-d   \344   228   E4  U+00E4  &auml;     ä  small a, umlaut mark
```
The changes I made here were actually in my obsolete patches from when I added UTF8=ACCEPT but, for unknown reason, I took them out. (Probably because I didn't test save-as, copy and view source on a UTF8 named folder so I missed the effect this has.)

I tested this on top of Ping's patch so I assume it is also needed to fix all the problems.
P/S: This probably only works for Western chars that are encoded in a single byte. Not sure about other sets, but probably not.