Characterset Problems

[Silvan Scherrer]

Following up to the above... the WPSU header reports support for the specific Greek codepages 1253 and 869. That is actually correct, because it /does/ support all the Greek characters in those codepages.

Something in QT4 (possibly Pango?) is somehow drawing the conclusion that WPSU supports the entire Greek Unicode block, which is wrong. Whether it's basing that conclusion on the fact that those two codepages are supported, or because it checks the font for some particular character (or range of characters), I don't know.

see also ​http://svn.netlabs.org/lucide/ticket/141 comment:7 and later

[Silvan Scherrer]

i think this has to do with r227. Probably this hack needs a adjustment

[Alex Taylor]

I've run into a more serious problem. Japanese text is completely undisplayable because QT4 apparently assumes that Workplace Sans supports it (it doesn't).

I'm guessing this is also related to changeset 227.

// Note that "Workplace Sans" TTF by Alex Taylor doesn't provide correct 
// unicode and codepage ranges in os2_table for some reason but versions 
// 0.4 and earlier are known to support many popular scripts so hack it. 

Um... do you care to elaborate on this? The values SHOULD be correct -- I'm pretty confident the codepage ranges are correct, and the unicode fields should be as well.

[Dmitry A. Kuminov]

Yes, this comment is from r227. I remember adding this hack because I found what it says in the commit message:

gui: Added a hack for "Workplace Sans" to make Qt think it supports all scripts
(Unicode and codepage ranges in the header indicate that it only supports
Latin1 but it's much less than the actual number of scripts which include Greek,
Cyrillic and even Japanese, which are supported at least partially).

I don't remember though how exactly I found this. But judging rom the message it would not even display Cyrillic (which is surely there). This needs to be rechecked indeed.

[Alex Taylor]

Was somebody testing against the incomplete WPSU version 0.5 maybe? That was more of a 'proof of concept' release to test some design changes, and may indeed have only supported Latin-1.

Since version 0.8 at least, Workplace Sans supports Latin-1, the Baltic range of Latin-2 and a few others (about half of Latin-2 in total), Cyrillic, Greek, various glyphs in the U+20xx - U+26xx range, and a handful of ligatures. It does not (yet?) support Hebrew or Hangul, nor Arabic, and will probably never support Japanese to any usable degree.

[Dmitry A. Kuminov]

Yes, this is possible. This is among the first things I will check when I switch back to Qt (I hope really soon). Thank you for your feedback, Alex.

[Silvan Scherrer]

this needs to be checked in Qt5. Please open a ticket at ​https://github.com/bitwiseworks/qtbase-os2