encoding of font-name with unicode branch new font-backend

encoding of font-name with unicode branch new font-backend

[Miles Bader]

If I do "C-u C-x =" on a character which is displayed with a
font-backend font that has a non-latin name, the font-name is not
displayed correctly:  It's displayed as raw binary characters, which are
actually the bytes of a UTF-8 encoded font name.

It seems that somewhere (decode-coding-string ... 'utf-8) has to be
called on the font name, but I'm not sure which layer has the knowledge
that the string is actually utf-8...

-miles
-- 
Fast, small, soon; pick any 2.

Re: encoding of font-name with unicode branch new font-backend

[Kenichi Handa]

In article <buor6vcuo2y.fsf@dhapc248.dev.necel.com>, Miles Bader <miles.bader@necel.com> writes:

> If I do "C-u C-x =" on a character which is displayed with a
> font-backend font that has a non-latin name, the font-name is not
> displayed correctly:  It's displayed as raw binary characters, which are
> actually the bytes of a UTF-8 encoded font name.

> It seems that somewhere (decode-coding-string ... 'utf-8) has to be
> called on the font name, 

I agree.

> but I'm not sure which layer has the knowledge
> that the string is actually utf-8...

Me neither. :-(
It seems that fontconfig can contains FC_FAMILYLANG property
in FcPattern, but there's no FC_FAMILY_ENCODING.
---
Kenichi Handa
handa@m17n.org

Re: encoding of font-name with unicode branch new font-backend

[James Cloos]

>>>>> "Kenichi" == Kenichi Handa <handa@m17n.org> writes:

Miles> but I'm not sure which layer has the knowledge
Miles> that the string is actually utf-8...

Kenichi> Me neither. :-(
Kenichi> It seems that fontconfig can contains FC_FAMILYLANG property
Kenichi> in FcPattern, but there's no FC_FAMILY_ENCODING.

My understanding is that it is always utf-8.

I'll look through the code to be sure and write back if I am wrong.

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

Re: encoding of font-name with unicode branch new font-backend

[James Cloos]

>>>>> "|" == James Cloos <cloos@jhcloos.com> writes:

|> My understanding is that it is always utf-8.

If I am reading the src correctly it does transcode everything to utf8.

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

Re: encoding of font-name with unicode branch new font-backend

[Kenichi Handa]

In article <m3veknan3r.fsf@lugabout.jhcloos.org>, James Cloos <cloos@jhcloos.com> writes:

>>>>>> "|" == James Cloos <cloos@jhcloos.com> writes:
>>> My understanding is that it is always utf-8.

> If I am reading the src correctly it does transcode everything to utf8.

Thank you for checking that.  Ok, I'll improve the code.

---
Kenichi Handa
handa@m17n.org

Re: encoding of font-name with unicode branch new font-backend

[Juliusz Chroboczek]

>> but I'm not sure which layer has the knowledge that the string is
>> actually utf-8...

Theory: it is always UTF-8.

Practice: it will always be UTF-8 for fonts in the TrueType/OpenType
format.  For fonts in weird formats, all bets are off.

I suggest you treat it as UTF-8, but you should be robust if it turns
out not to be well-formed UTF-8.

                                        Juliusz