fonts, faces and the like

fonts, faces and the like

[Jens Lautenbacher]

Hi everybody,

I have some questions concerning the use of faces a it is now in
sgnus.

I strongly belive it is necessary to re-think the way of highlighting
certain parts of the summary lines.

Currently there seem to be two points where one can specify which face
to use where and in which situation.

1) In the gnus-summary-line-format using the %n{xxxx%} mechanism. This
   is some sort of global preference, but can act on parts of the
   line.

2) with the  gnus-summary-highlight list. This one combines certain
   conditions into certain faces (article read -> custom-face-?-?...)
   This list acts only on the whole line (XEmacs at least). 

2 takes precedence over 1. 

I belive that this should better be the other way around, say if  I
would like to have all subjects be in 'bold face, 2) will still do
it's job on the authors and the rest of the line.

But maybe one should do things right once and forever. The main
drawback from the scheme that's currently used is that one cannot
easily merge properties of faces. Say, I would like to have the root
of all threads be in a bigger font (which can be done in XEmacs, or
BOLD (not the face 'bold!) in Emacs) _and_ still have all lines
coloured with respect to their status. By now this is only possible if
I define _all_ permutations of <header is root> with all other
conditions. 

I am currently thinking of some kind of face-server which could have
.e.g the following calling syntax

(face-server-make-extent BEGIN END &rest KEYWORD VALUE)

This call would inspect all extents in the newly to generate extent
from BEGIN to END, would look at all faces defined in these extents
and changes them according to KEYWORDS witch could be 'family 'bold
'italic, 'bold-italic 'foreground 'background 'underline and 'size

The server would probably generate font names similar to custom-faces,
but with additional size info, so every generated font name would
uniquely determine it's face.

I don't think it should be to difficult to implement and hope I can
start coding over christmas, if noone tells me that it's impossible or
not useful. 

Any comments?

Re: fonts, faces and the like

[Lars Magne Ingebrigtsen]

Jens Lautenbacher <jtl@tkm.physik.uni-karlsruhe.de> writes:

> I am currently thinking of some kind of face-server which could have
> .e.g the following calling syntax
> 
> (face-server-make-extent BEGIN END &rest KEYWORD VALUE)
> 
> This call would inspect all extents in the newly to generate extent
> from BEGIN to END, would look at all faces defined in these extents
> and changes them according to KEYWORDS witch could be 'family 'bold
> 'italic, 'bold-italic 'foreground 'background 'underline and 'size

That sounds very useful, and would definitely solve lots of problems.

-- 
Home is where the cat is.

Re: fonts, faces and the like

[William Perry]

Lars Magne Ingebrigtsen writes:
> Jens Lautenbacher <jtl@tkm.physik.uni-karlsruhe.de> writes:
> 
> > I am currently thinking of some kind of face-server which could have
> > .e.g the following calling syntax
> > 
> > (face-server-make-extent BEGIN END &rest KEYWORD VALUE)
> > 
> > This call would inspect all extents in the newly to generate extent
> > from BEGIN to END, would look at all faces defined in these extents
> > and changes them according to KEYWORDS witch could be 'family 'bold
> > 'italic, 'bold-italic 'foreground 'background 'underline and 'size
> 
> That sounds very useful, and would definitely solve lots of problems.

  I've written a fair bit of this for Emacs-w3's stylesheet implementation,
and am probably going to put it in the C code eventually, just have to find
the time and expertise.  Check out font.el in the Emacs-W3 distribution.
It handles this fairly well on X and TTY (under XEmacs, etc)

-Bill P.