From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20021218 Description of problem: When I input Japanese string into html form with Mozilla, it shows very dirty fonts. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Input Japanese string into http://www.google.com/ form with mozilla 2. Japanese 'aaaaa' 3. Dirty Additional info:
Created attachment 89418 [details] mozdirty.png
If Hiragana is in the string, it gets very dirty. If not, the string gets antialiased. That's very silly behavior.
That's all Xft.
Look like it is getting two different fonts. Is mozilla just getting the first "suitable" font on the list?
Yes, that's what it does via the xft setup.
A) Perhaps Mozilla isn't setting the language tags for controls based on the language of the page? B) Does removing the fonts in /usr/share/fonts/ja/misc improve the situation? C) Does removing bitmap-fonts-cjk improve the situation (if you have that installed?)
> B) Does removing the fonts in /usr/share/fonts/ja/misc improve the situation? mv /usr/share/fonts/ja/misc /tmp; still the same > C) Does removing bitmap-fonts-cjk improve the situation (if you have that installed?) Same- However I have tested around on which fonts that effect the selections: *if* i remove console* in /usr/share/fonts/bitmap-fonts/ then the display and font selection will look much better (will only select ttf font so that we could not see the rendering problem of console glyphs). We may have two folds the problem - one may related to ways of only return a font with nearest to fullset of the charset. Another one is the rendering problem on the console font(?)
> A) Perhaps Mozilla isn't setting the language tags for controls > based on the language of the page? As far as I know, it is.
Seems to be fixed in Phoebe4 + mozilla 1.2.1-15