Re: font problem persists

[Stefan Monnier]

Stefan> So instead, you should only specify the font family, and
Stefan> size, while the charset is automatically selected by Emacs
Stefan> (which will use the iso10646-1 font for the Unicode chars,
Stefan> ...).

>>> emacs -font  "-adobe-courier-bold-o-normal--0-0-100-100-m-0-iso10646-1"

> Ok then I tried

>>> emacs -font
>    "-adobe-courier-bold-o-normal--0-0-100-100-m-0-*-*"

> And the behaviour is the  same, that is 
>  C-h h displays Hebrew and Greek correctly.
> However when I open a file, which I recently have edited with Greek
> and Hebrew symbols and saved as UTF-8, those symbols are displayed as
> empty boxes.

The internal representation of non-ascii chars in MULE (later inherited by
Emacs and XEmacs) is based on iso-2022 which, contrary to Unicode which
didn't really exist yet at that time, does not unify "identical" characters.

So what you're seeing is that the Hebrew and Greek chars used in C-h
H are not the same chars as the ones used in utf-8 files (try C-u C-x =
on each of those chars to see which charset they are part of).

The Hebrew chars you see in C-h H are displayed using a Hebrew font.
The Greek chars you see in C-h H are displayed using a Greek font.
The Hebrew and Greek chars you don't see in your utf-8 file are displayed
with a Unicode font.

With Emacs-CVS ("soon" to become Emacs-22), you can set
utf-fragment-on-decoding, which should help for some of those chars
(probably the Greek chars, maybe not the Hebrew).


        Stefan