https://bugzilla.novell.com/show_bug.cgi?id=246573 ------- Comment #55 from mfabian@novell.com 2007-02-23 17:40 MST ------- Here is a comment by Werner LEMBERG (Werner currently has no reliable internet access and asked me to paste his mail into bugzilla): WL> > Werner, is it correct that the width of strings changes if the WL> > autohinter is used? WL> WL> Yes, of course. This is the same as with TrueType. Just think of the WL> `hdmx' table which caches horizontal metrics for selected ppem values. WL> WL> > As you can see in comment #43 this seems to make it impossible to WL> > check whether "一二三四五六七八九零" is twice as wide as "1234567890" WL> > in xterm because the result may depend on the hinting settings. WL> WL> This is a difficult issue. Because real TrueType fonts have been WL> designed and hinted by humans, the CJK widths don't change even if WL> hinted (in the normal case). However, the autohinter doesn't know WL> this. WL> WL> I see two solutions; none of them is implemented currently. WL> WL> 1. Introduce a user-settable flag FT_FACE_FLAG_DOUBLE_WIDTH which WL> indicates that the font has single and double width characters WL> only. WL> WL> 2. Add some heuristics to both the latin and CJK autohinter (or WL> perhaps even to the general font loading engine of FreeType) WL> which sets the above flag automatically. WL> WL> Handling all CJK characters automatically as fixed-width is probably WL> not a good idea, at least not for the vertical case; there exist WL> Japanese variable-height fonts. However, this might be a workaround WL> until we find a real fix. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.