288571 - Armenian text doesn't render correctly on web pages

Status: RESOLVED DUPLICATE of bug 246527

(Camino Graveyard :: General, defect)

Description

[Paul Davidson]

User-Agent:       Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.8b2) Gecko/20050321 Camino/0.8+
Build Identifier: Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.8b2) Gecko/20050321 Camino/0.8+

I was looking for multilingual pages to test the non-Latin script support of
both OS X and Camino. At the page in question, I discovered that neither Thai
nor Armenian text were correctly rendered. Note that I do have Thai and Armenian
fonts installed; moreover, this page renders correctly in Safari.

The Thai sample appears as random Latin characters. The Armenian text appears as
a string of question marks.

Reproducible: Always

Steps to Reproduce:
1. Navigate Camino to
http://meta.wikimedia.org/wiki/Complete_list_of_language_Wikipedias_available
2. Find Thai and Armenian in the list.

Actual Results:  
The rendering problem becomes apparent.

Expected Results:  
Camino should have rendered Thai characters in Thai script, and Armenian
characters in Armenian script, using appropriate fonts. See Safari for an
example of correct behaviour.

Firefox exhibits the same problem, so the gecko engine might need to be fixed.

Comment 1

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

There's a known bug where Gecko cannot properly interpret some Mac fonts and
their character repertoires.  

If you've set your Thai (and perhaps Armenian) font to anything other than the
default (Lucida Grande and Monaco), then text will display as gibberish. 

*** This bug has been marked as a duplicate of 246527 ***

Comment 2

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

I'm reopening this bug while we track down what's going on with Armenian...if it
turns out to be a bug, we want the discussion preserved here rather than lost in
email.  Removing the Thai stuff from the summary and URL, but keeping it in
Camino for the moment....

> Sorry about the duplicate, this was my first time submitting a Camino  
> bug.

No worries.  It's a pretty obscure bug with a bizarre work-around, and I only
knew about it because I ran into it (and had it explained to me) while working
on modernizing the default font prefs :-)

> Since Armenian (1) does not get its own font settings in Camino, and  
> (2) displays as question marks rather than random characters, could  
> that be a different bug than #246527? Could bug #237434 be the cause?  
> If not, should I submit a separate bug report?

Hmm.  There's not a UI in Firefox, either, so I have no idea where it's getting
the font settings from.  But both Camino and Firefox are displaying Armenian for me.

A couple of questions for you:

1) What version of Mac OS X?
2) Have you changed your "Unicode" or "User-Defined" font preferences
3) Did you install a limited set of fonts when installing Mac OS X, or have you
removed any OS-installed fonts?

I suppose it's also possible that I have a 3rd-party font that you don't and
that's why I'm seeing the Armenian, so I'll look into that, too.

Comment 3

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

Ah, I missed the part where you found the bug about adding the UI.  It could
very be the problem--if the font isn't set, Gecko could be guessing what font to
use, and it's font guessing code is very old and very bad.  But it could also be
the fact Gecko can't read certain Mac fonts (bug 246527 that I originally duped
this to).

There's only one standard Mac font that has Armenian glyphs, Mshtakan (whereas I
have 6 others on my system that have those glyphs).

I'm not sure how to check whether it is one or the other; I'll ask in the UI bug.

Comment 4

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

jshin explained in bug 237434 comment 11 that if the page contained a lang=xx
statement, setting the "Unicode" font prefs to the proper fonts should display
the page in those fonts if Gfx:Mac can handle the fonts.

The Armenian Wikipedia has
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="hy" lang="hy" dir="ltr">
which should cover the lang statement.

Changing my Unicode fonts to Mshtakan (in my testing account without those other
six fonts) didn't produce any change in display--still ????--and the work-around
we use for Arabic, Hebrew, and Thai (Lucida Grande and Monaco) didn't work,
either.  The clincher was the fact that Mshtakan doesn't even show up in the
Firefox prefs (which use the broken Gfx:Mac stuff directly; Camino uses the
Cocoa dialogue), so, while there may be other problems, the most immediate one
is still the original one I duped this against, bug 246527.

As a final note, if you have third-party fonts like TITUS Cyberbit, Everson
Mono, the GNU Free fonts series, etc., installed, the pages will display in
Armenian because Gfx:Mac seems to understand those fonts and font fallback is
working more or less properly.

*** This bug has been marked as a duplicate of 246527 ***