58412 - i18n: using different fonts for latin script

Status: RESOLVED FIXED

(Core :: Internationalization, defect, P3)

Description

[Richard]

When viewing document  with CSS like
http://www.unicode.org/unicode/standard/translations/lithuanian.html
 you can see accented letters in smaller size than non-accented.
 I do have arial and other fonts and have set all 'language encoding'
 fonts to the same font in different encodings (arial, times new roman,
misc-fixed) and size (16 prop. & 14 fixed).
 In any case Mozilla shouldn't use different fonts for accented and non-accented
 characters because it looks very ugly; or at least allow selection of
 iso10646 fonts for 'language encoding's other than Unicode if that can help.

 Mozilla/5.0 (X11; U; FreeBSD 4.1.1-STABLE i386; en-US; m18) Gecko/20001029

Comment 1

[nhottanscp]

Reassign to erik.

Comment 2

[Erik van der Poel]

Hello Richard,

Thanks for the bug report. Which "language group" would you say Lithuanian
belongs to:

ar
el
he
ja
ko
th
tr
x-baltic
x-central-euro
x-cyrillic
x-unicode
x-user-def
x-western
zh-CN
zh-TW

Would it be Baltic? Also, which X font charsets support Lithuanian completely?
iso8859-2? iso8859-4? iso10646-1?

Comment 3

[Richard]

        Hello,
It is x-baltic. iso 8859-2 lacks some letters. iso 8859-4 lacks quotes
0x201e 0x201c but
quotes are not frequently used on web pages.  iso 8859-13 also has baltic
letters though some rarely used windows-125x specific signs may be missing for
windows-1257 webpages.  But there are few 'original' iso 8859-4 or iso 8859-13
fonts, most likely these fonts are simply recoded from iso 10646 font
by X or by other means.  IMHO it would good to have iso10646 fonts available in
'Baltic' language encoding menu if this is possible.
        There are also some rarely used accents 0x300,0x301,0x303 but AFAIK
mozilla
doesn't support combining characters anyway.

        Thank you,

Comment 4

[Erik van der Poel]

Thanks for the info, Richard. As you know, Mozilla has a font prefs dialog, and
we try to show only the fonts that are suitable for the selected language group.
Would you say that iso8859-2, iso8859-4 and iso8859-13 fonts are not suitable
for the Baltic group? Or should we show one or more of those 3?

For iso10646-1 fonts, I suppose we could load the font and check to see whether
it contains enough glyphs to support Baltic. In that case, which Unicodes do you
think we need to check for? Just the Unicodes corresponding to windows-1257? Or
the union of windows-1257 with some other (e.g. iso8859-4)? Or is 1257 a
superset of iso8859-{2,4,13}? Thanks again.

Comment 5

[Richard]

iso 8859-2 is definitely not suitable.  iso8859-4 and iso8859-13 are suitable -
 I'm assuming if there is some random rare character ouside symbols covered by
 these fonts, mozilla will show it anyway.  I think iso8859-4, iso8859-13 and
 iso10646 can be shown.  As far as heard Estonian pages may as well use
 iso8859-15/iso8859-1 but this is 'Western' script not 'Baltic'.

 iso8859-13 is almost subset of cp1257, lacking 'standard' windows cp125x
 0x80-0x9F range.  iso8859-4 is different.  I think characters common both for
iso8859-4 and for iso8859-13
 should be enough.  Both these encodings have characters belonging to other
 languages besides Baltic ones.  For Lithuanian you must have us-ascii and
 104 10c 118 116 12e 160 172 16a 17d 105 10d 119 117 12f 161 173 16b 17e
 letters.   201e 201c 300 301 303 are not essential, thought 201e and 201c
 ideally should be converted to simple us-ascii quotes in that unlikely case
 they can't be shown.  Sorry I can't list Latvian and Estonian characters -
maybe
 it is possible to get them from
 http://www.unicode.org/Public/UNIDATA/NamesList.txt
 or from some Windows 'script' list.

Currently I can't see any iso10646 fonts except 'Unicode' font menu though I
have
common Windows fonts and Cyberbit Unicode font.

Comment 6

[Erik van der Poel]

iso-8859-4 is already set to x-baltic in

  mozilla/intl/uconv/src/charsetData.properties

but iso-8859-13 is missing. Is 8859-13 only suitable for Baltic, or also others?

Also, we don't list iso10646-1 fonts under Baltic in the font prefs. To support
this, we would need to load one or more of the 10646 fonts to see which glyphs
they support, and if they are suitable for Baltic, add those fonts to the list.

Brian, is this something you could take on? Please reassign to yourself if so.

Comment 7

[Richard]

In theory iso8859-10 is similiar to iso8859-4 and AFAIK has the same letters for
Baltic languages.  But I didn't saw any iso8859-10 fonts except these autogenerated
from iso10646 altogether with iso8859-4 and other iso8859 encodings.
Also there are vendor encodings as windows-1257 and MacCE but I don't think this
should be supported on Unix.

Comment 8

[Frank Tang]

reassign to bstell and keep it as future

Comment 9

[kill this account]

--> ftang

Comment 10

[Frank Tang]

bulk move NEW FUTURE bug to ASSIGN

Comment 11

[Jungshik Shin]

This was fixed a long time ago. Both iso8859-13 and iso8859-4 are listed for
Batlic 'langGroup'. I can add Windows-1257 as well.