108888 - UTF8 page rendering font improvement

Status: RESOLVED WORKSFORME

(Core :: Internationalization, defect)

Description

[Shanjian Li]

Original message from Markus Kuhn,

> > How good is Netscape 6.2 for Linux with regard to UTF-8 ?
>
> Somewhat better than Netscape 4, but not perfect. It uses now for UTF-8
> pages the same normal font sizes that it uses for ISO 8859-1 pages,
> therefore UTF-8 pages such as
>
>   http://www.columbia.edu/kermit/utf8.html
>   http://www.kernel.org/
>   http://www.unicode.org/
>
> are not shown any more in the ugly and huge single Unicode font size
> that Netscape 4 used.
>
> Unfortunately, Netscape 6.2 still picks for instance the Cyrillic fonts from
> (probably) JIS X 0208 encoded EastAsian fonts, which have square
> (double-width) glyph spacing designed for ideographs, therefore Russian
> UTF-8 text looks still pretty ugly by default. This is particularly
> silly, because Netscape obviously does open installed KOI8-R fonts in
> order to get suitably sized box drawing characters.
>
> Disappointing is also that it mixes glyphs from different font sizes
> where a monospaced font is used, even if suitable monostaced
> -misc-fixed-iso10646-1 fonts are installed, as you can see for example
> on fixed-width plain-text test pages such as
>
>   http://www.cl.cam.ac.uk/~mgk25/ucs/examples/TeX.txt
>   http://www.cl.cam.ac.uk/~mgk25/ucs/wgl4.txt
>
> I have not been able to fix this myself. I tried going to
> Preferences|Fonts|Unicode|Monospace and selected there the offered
> "misc-fixed-iso10646", but nothing changed.
>
> I understand if Netscape doesn't use ISO10646-1 fonts everywhere by
> default (due to the unfortunately huge XFontStruct allocations that they
> cause), but I don't understand why explicitly selecting a single
> "misc-fixed-iso10646" size doesn't work. Looks like a bug to me.
>
> Summary: With Netscape 6.2, the ugly UTF-8 fonts used by Netscape 4 are
> now no longer a reason for web page authors to avoid the UTF-8 encoding.
> On the other side, I have failed to use proper ISO10646-1 fonts and as a
> result, UTF-8 test pages with content bejond Latin-1 still do not look
> as good as they could easily.
>
> Markus
>
> --
> Markus G. Kuhn, Computer Laboratory, University of Cambridge, UK
> Email: mkuhn at acm.org,  WWW: <http://www.cl.cam.ac.uk/~mgk25/>

Comment 1

[Frank Tang]

change os to Linux

Comment 2

[Shanjian Li]

accepting. 

Comment 3

[Alexey Chernyak]

I think this is fonts availability issue, rather than Mozilla's fault.
Initially I did have Cyrillic looking very ugly as well. But, I have sent email
to administrator with a request to install a proper Cyrillic font set on the
system, and after it was done, Mozilla displays it beautifully.

Comment 4

[Alexey Chernyak]

hmmmmm though actually the pages I experienced this were KOI8-R, but UTF-8 pages
show with no such problems as well now.

Comment 5

[kill this account]

so should we close this bug?

Comment 6

[Markus Kuhn]

I don't think this is a fonts availability issue. The underlying problem is that
glyphs from CJK fonts are preferred over glyphs from ISO10646-1 fonts. So even
if all glyphs in a text are available in a single consistent style from the
ISO10646-1 currently being used and configured, many glyphs are unnecessarily
being taken from fonts in other encodings, leading to very unpleasant
presentation of text pages such as

http://www.cl.cam.ac.uk/~mgk25/ucs/wgl4.txt
http://www.cl.cam.ac.uk/~mgk25/ucs/examples/TeX.txt

This is definitely an unpleasant problem with

  Mozilla 1.0
  Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.0) Gecko/20020529

on any system with all the standard XFree86 4.2 fonts being available.

I haven't tried any more recent built yet (has anything significant been fixed
since then?).

Comment 7

[kill this account]

To my knowledge there has been no change in this area. There is no efficient 
way to find out what is in a iso10646 font and new font development is mostly 
focusing on Truetype at the moment.

Comment 8

[Alexey Chernyak]

Cyrillic double-width problem is bug 107217.
Monospace problems bug 105026.
Some more monospace problems bug 182738.

1.1alpha was released ages ago, resetting target milestone.

Comment 9

[Frank Tang]

shanjian is no longer working on mozilla for 2 years and these bugs are still
here. Mark them won't fix. If you want to reopen it, find a good owner first. 

Comment 10

[Alexey Chernyak]

VERIFIED
All testcases linked here look good on Firefox 1.0.1 Linux with plenty of TTF
fonts installed and without manual font selection in settings.

Otherwise, the bugs I listed above should cover pretty much all causes of this
bug report.

Comment 11

[Travis Chase]

Marking a proper resolution of worksforme based on comment #10

Comment 12

[Markus Kuhn]

This is just to confirm that this bug report was a problem of an ancient release
and is not an issue with recent Firefox versions. Has worked fine for me for
quite a while on SuSE Linux 9.1 or newer.