Ticket #2410 (closed defect: duplicate)
Position calculation is byte-related not character-related
| Reported by: | gotar | Owned by: | |
|---|---|---|---|
| Priority: | trivial | Milestone: | |
| Component: | mcview | Version: | 4.7.4 |
| Keywords: | Cc: | ||
| Blocked By: | Blocking: | ||
| Branch state: | no branch | Votes for changeset: |
Description
Movement functions are using constant widths of 1 byte ignoring character conversion (especially for UTF-8 encoding which shrinks 2 byte representation to display, but also tab character which expands 1 byte to 8 cells).
This results in some glitches in wrap mode on top line, when wrap point is missed (function takes view->data_area.width of bytes not characters, i.e. on a left side) or skipping an extra row if wrapping occours 'after' EOL (row expanded to the right by tab characters).
This is mostly harmless for files containing mostly 1-byte encoded characters, and resolving it would require either converting entire file to internal representation, or doing this in-place for every line moved. Anyway this would have negative impact on performance, so unless it affects users of non-latin alphabets (like russian or chinese) I'd suggest to close this ticket as wontfix and leave it just for the record.
This is easy to reproduce with UTF-8 encoded files on 8-bit terminal at least and with tabbed lines having $COLUMNS length on screen.
