Characters appearing as '?' (Re: Good Unicode serif font)

On Sun, 10 Sep 2006 01:58:51 GMT, Rich Walsh <spamyourself@xxxxxxxxx> wrote:
Does anyone use/know of a nice Unicode serif font, which eliminates *all* of
those "?????" f.e. on http://en.wikipedia.org/wiki/Zeta_%28letter%29 ?

There's something "wrong" with one of the stylesheets. Using PmW-Fx-os2-1.5,
if I view page source, or if I select View->Page Style->No Style, all of the
Greek appears OK. With the style *** enabled, I see the same question marks
that you do.

I don't know if it's the style*** or not, but this looks suspiciously
like the same bug I've been running up against for the past couple of
months. (See my postings in various groups, particularly the mozilla one,
complaining about rendering of UTF-8 characters.)

I mostly encounter it in reading UTF-8 encoded emails in a Japanese
newgroup I read from time to time. However, I've seen something similar
on certain web pages as well. I haven't yet identified the pattern that
causes web pages to exhibit the problem, but it seems to be 100%
reproducible with messages.

In every case, the problem appears the same way: some or all characters
that don't exist in the default encoding (ISO-8859-1 in my case) simply
show up as question marks when viewed in the main program. HOWEVER, they
appear correctly in the "view source" window.

Usually, on web pages the problem goes away when you turn off the checkbox
"allow pages to use their own fonts". (This doesn't work with emails,
however.) I notice on the web site discussed here, this doesn't appear to
help.

I believe the problem is that Mozilla is _supposed_ to automatically load
the character glyphs (that don't exist in the current encoding+font) from
whatever font is specified under 'Other Languages'... but, for whatever
reason, it isn't doing so.

I'm currently working on a submission to Bugzilla, but it may be another
day or two before I can get it in.


The problem occurs with Firefox, with Thunderbird, with SeaMonkey, with
Mozilla 1.7.x -- both official and PW builds -- and even with IBM Web
Browser 2.x. I've only started to notice it this year, because I've
started reading sci.lang.japan which contains a fair number of UTF-8
messages.


--
Alex Taylor
http://www.cs-club.org/~alex

Remove hat to reply (reply-to address).
.