Some console fonts (i.e.Lat2-VGA8.psf.gz) have bad Unicode maps

Bug #190746 reported by Michal Breznicky
6
Affects Status Importance Assigned to Milestone
console-setup (Ubuntu)
Invalid
Medium
Unassigned

Bug Description

Binary package hint: console-setup

It seems that the file Lat2-VGA8.psf.gz has a partially incorrect Unicode map:

a) The Unicodes for some box lines and capital letters are wrong; for example, a character number 80h, that has the following bitmap (8x8 matrix follows)
........
XXXXXXX.
.XX.....
.XXXX...
.XX.....
XXXXXXX.
..XX....
...XX...
has associated unicodes 2500 (BOX DRAWING LIGHT HORIZONTAL) and 2501 (BOX DRAWING HEAVY HORIZONTAL) instead of 0118 (LATIN CAPITAL LETTER E WITH OGONEK). Consequently, all the mentioned characters are not displayed correctly. Still, the majority of the map is OK (i.e. the lowercase accented characters are mapped correctly). The easiest way to reproduce this is probably to:

        1) Set /etc/default/console-setup this way:
        CHARMAP="UTF-8"
        CODESET="Lat2"
        FONTFACE="VGA"
        FONTSIZE="8"

        2) Run setupcon in the console

        3) Run midnight commander in the console and have a look at the boxes.

This also applies to some other fonts (I checked a handful of other Lat2* fonts).

b) I wonder if this is a bug (probably technically it is not): Actually the order of the characters in Lat2-VGA8.psf.gz (even those that are correctly mapped) is not the same as in the official ISO_8859-2 definition.

For example, the letter number a lowercase "r" with an acute (unicode 0155h) is number E0h in the ISO table, but number C8h in the Lat2-VGA8 file.

Revision history for this message
Colin Watson (cjwatson) wrote :

You're right, there does seem to be something wrong with the source BDF files. Apparently these were autogenerated from console fonts, and I don't really know my way around which BDF files are used for what - I'll have to talk to Anton to figure out what's going on here.

Changed in console-setup:
importance: Undecided → Medium
status: New → Triaged
Revision history for this message
Paul Sladen (sladen) wrote :

They were likely generated using 'otf2bdf'. I'd concur that because of the simple nature of 'bdf2psf' the actual errors are likely have been generated at the pre-BDF stage. They could relate in incorrect equivalency mappings.

Revision history for this message
Mörgæs (moergaes) wrote :

Closing due to age.
If a similar bug appears in 15.04 (development) please open a new report.

Changed in console-setup (Ubuntu):
status: Triaged → Invalid
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.