Re: Characters displayed small

At 19:22 on Feb 2 2006, Haines Brown said :

Sébastien Kirche <sebastien.kirche.no@xxxxxxxxxxxxxxxxxxxx> writes:

- there is some helpers like ucs to merge latin-1 and latin-9 for
decoding or encoding (see unify-8859-on-decoding-mode and
unify-8859-on-encoding-mode)

I confess my ignorance. What do you mean by "helper"? What is
"unify-8859-on-decoding"? The closest thing I could discover was the
lisp program rs-gnus-unify.el, but probably I need utf-8.el and
ucs-tables.el. Is installing all of them what you are suggesting?

Well, ucs stands for «Universal Character Set» as defined by the
iso-10646 standard (perhaps that the Utf-8 and Unicode FAQ[1] could give
you some more answers about it).

In its ucs-tables.el module, Emacs contains some tables definitions and
routines (what I called helpers) to convert and ease handling of
characters between different sets.

unify-8859-on-decoding-mode and unify-8859-on-encoding-mode are 2
functions that make emacs consider equally the characters from latin-1
set that are the same as the ones from latin-9 (with the exception of
euro char, oe and some other few).

I don't know your skills level with emacs, but if you ask emacs to give
some help about unify-8859-on-encoding-mode (C-h f
unify-8859-on-encoding-mode RET), the docstring gives some interesting
infos.

To give it a try, you can put the following in your .emacs file :
(unify-8859-on-encoding-mode t)
(unify-8859-on-decoding-mode t)

After that, while entering some latin-9 characters (as you keyboard
coding system seems to do) they should use the same font than those
already in latin-1.

And I don't understand why the font changed from the Lucida (which
surely has a representation for e-acute) and Schumacher.

I am pretty sure that is because when you display some already typed
text is was encoded (at the sent time for a message, or at recording
time for a file) in latin-1/iso-8859-1, but when you type some text the
keyboard generates some latin-9/iso-8859-15 characters. Each character
set using a different font as told by the fontset.

It does. Is the implication that
-b&h-lucidatypewriter-bold-r-normal-sans-12-120-75-75-m-70-iso8859-1
does not have accented characters? Does this mean that when I
configured emacs to use this font, I should have used another? A more
general question, wouldn't this font contain accented characters. I
can display the character map for LMTypewriter10 for Latin, and it
shows accented characters.

I am not sure. The use of a different font may be explicit in a way or
another, for example by defining a fontset in the .emacs, or maybe Emacs
did not find a suitable character in the latin-1 font and looked for a
capable font. I also might be a bad behavior fixed in the further
development.


It is possible to tailor a fontset to display the right characters
but I follow David Kastrup when he suggests you to try a developper
version of Emacs.

Is using a "helper" such as you mentioned above an alternative way to
reconcile character sets? Or should I really try a developer version
of emacs?

You could try with the unify-8859-xx functions I mentioned as a first
step.

Maybe some further investigation could be done to check that the coding
system settings in the emacs configuration suit your system locales. But
it could be difficult to point out. One have to compare the locales to
several settings (functions or variables) including :
- set-language-environment
- prefer-coding-system
- set-locale-environment
- unify-8859-xx
- set-keyboard-coding-system
- set-terminal-coding-system
- the fonts available and the fontset (the latter usually does not have
necessary to be altered)

To have some ideas about such a configuration, you could take a look to
mine[2] where i use almost all of them, but beware that I am not using
utf-8 locales but latin-9...

I'm still in the dark about "developer packages". As a non-developer,
or programmer, "twiddling" of some SID package sources (for emacs?)
seems intimidating. And I don't know where to locate emacs-snapshot
packages and choose among them. Are snapshots and developer packages
the same thing?

Generally speaking, the cvs version of Emacs is called the developer
version as it usually need developer skills to be able to build the
application from sources, and also because it is not in a finished
bug-free state (of course, a release version can also have bugs, but the
cvs version has no guarantee at all ;).

But you can find quite easily some pre-compiled packages as the
emacs-snapshot debian package. As you can guess from its name it is
built periodically from a snapshot of the source tree. At least the
debian package maintainer ensures that the build is stable enough to be
usable.

To be able to get that package that is only in the Debian SID (aka
Unstable) you have to alter /etc/apt/sources.list in order to add some
source for the unstable distribution. But I told 'twiddling' because if
you use the stable version you must be careful not to let apt install
too many of the unstable packages dependencies and compromise your
system stability.

I am myself totally with the unstable distribution, so I don't know
accurately how to proceed...

I'm currently using 21.4.1. Are you suggesting that I might do well to
downgrade emacs to 21.3 and not have to "twiddle" with the testing
version?

Mmh, no. 21.4.1 is not very different from the 21.3. It was released as
a security fix, and I don't know if there is more bugfixes, but 21.3
can't be better. I hope that you will manage to configure apt to get the
emacs-snapshot as you can have it together with the release version.

Footnotes:
[1] http://www.cl.cam.ac.uk/~mgk25/unicode.html
[2] http://sebastien.kirche.free.fr/emacs_stuff/emacs.html

--
Sébastien Kirche
.

Relevant Pages

  • CTAN package update: microtype
    ... A package has been updated on tug.ctan.org and should soon make its way ... new command \microtypecontext to change the configuration context; ... * New key `font' to add single fonts to the font sets ... to insert small spaces before certain characters (e. ...
    (comp.text.tex)
  • Re: Multiple coding systems, and filesystems
    ... The page has a style sheet that makes special suggestions on the font of such words, in a most confusing and tricky way. ... If there were just a few "special" characters, you could present them using entity references like ö or character references like ą, but this gets clumsy if you have full sentences that consist of "special" characters. ... It's not possible to switch the character encoding in the middle of an HTML document. ... Sometimes Emacs asked me for what coding ...
    (comp.infosystems.www.authoring.html)
  • Re: Characters displayed small
    ... But you have to know that internally Emacs is considering ... unicode will be used internally to handle all characters ... example an us-ascii only font without accentuated characters, ... I'm still in the dark about "developer packages". ...
    (comp.emacs)
  • CTAN Update: cyklop font
    ... Licence: GUST Font License ... Most characters have thin rectangles as additional counters giving the ... The font was generated using the Metatype1 package. ... For the CTAN Team ...
    (comp.text.tex)
  • Re: Chinese character & pinyin frequency analysis
    ... Chinese characters, you should be using one that does it ... Richard> in the font you are using, ... How does Emacs help with my telling where ZWJ and ZWNJ have been placed? ...
    (sci.lang)