Re: bbdb and searching

Re: bbdb and searching

[Kai Großjohann]

remi.letot@easynet.be (Rémi Letot) writes:

> Now I don't understand the bits about 'using the X11 compose key
> (<compose> " a)'. Could you be more specific ? And is there a generic
> way to solve this problem ?

One of the keys on my kbd generates the Multi_key keysym.  I'm used
to calling this the Compose key, because that's what's printed on it
on a Sun keyboard.  X11R6 provides a general mechanism for typing
nonascii characters via the Multi_key.  You type the Multi_key, then
a sequence of other keys, and the application sees some nonascii
character.  For example, the Multi_key followed by " followed by a
inserts the ä character in my programs.

Some time ago, I was running in a Latin-1 environment, so the above
key sequence produced ä in Latin-1 encoding.  Now I switched to
LC_CTYPE=de_DE@euro, so Emacs uses a Latin-9 environment, and hence
the same key sequence produces ä in Latin-9 encoding.

But actually it does not matter how the characters get into Emacs.
If you have the following line:

    bäh bäh

and the two ä come from different charsets, then C-s C-w C-s when at
the beginning of the above line does NOT find the second word.  Of
course, users would like C-s C-w C-s to find the second word.

In Emacs from CVS, you can turn on unify-8859-on-decoding-mode and
this will help somewhat because it reduces the number of times that ä
from different charsets appears in your file.  (But it is also bad for
editing the file lisp/ChangeLog that comes with Emacs because it
changes the encoding of Pavel Janík's name -- the í should be from
Latin-2 but is reencoded into Latin-1 by unify-8859-on-decoding-mode.
Probably this special case is not very important for you.)

Work is currently in progress to change the internal encoding of
Emacs from Mule to something based on Unicode (ie, Unicode with some
extensions).  When this happens, all ä characters will be the same
inside of Emacs.  Then the searching problem will be gone.  (It also
means that Pavel's name will be affected, one way or another.)

kai
-- 
Silence is foo!

Re: bbdb and searching

[Fredrik Staxeng]

[-- Attachment #1: Type: text/plain, Size: 60 bytes --]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

[-- Attachment #2: Type: text/plain, Size: 973 bytes --]


> Fredrik Staxeng <fstx+u@update.uu.se> writes:
> 
> > Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:
> >
> >> In Emacs from CVS, you can turn on unify-8859-on-decoding-mode and
> >
> > Will this also solve the problem with 8859 variants in emacs -nw?
> > In the Swedish groups some people post with 8859-1, and some people
> > withh 8859-15 (I think), and one or the other get their åäö 
> > displayed as ?.
> 
> I don't know if unify-8859-on-decoding-mode will help.  But there is
> a file latin1-disp.el (or latin-1-disp.el) that comes with Emacs that
> might help.  (Possibly it's not in 21.1 or 21.2.)

Yes, but customize won't save the setting for me. I'll figure out how
to do it with lisp later.

> > Is there information somewhere about the CVS version of Emacs somewhere?
> 
> http://savannah.gnu.org/  Then search for the emacs project.  Is this
> the information you had in mind?

Yes. Thank you.

-- 
Fredrik Stax\"ang | rot13: sfgk@hcqngr.hh.fr

Re: bbdb and searching

[Kai Großjohann]

remi.letot@easynet.be (Rémi Letot) writes:

> Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:
>
>> In Emacs from CVS, you can turn on unify-8859-on-decoding-mode and
>> this will help somewhat because it reduces the number of times that ä
>> from different charsets appears in your file.  
>
> Herm, unify-8859-on-decoding-mode and unify-8859-on-encoding-mode are
> both on on my machine. And yet I still have different charsets in my
> bbdb it seems.

Hm.

> Is there some way to convert everything in it to Latin-9 ?

C-x C-f ~/.bbdb RET C-x RET f latin-9 RET C-x C-s

>> Work is currently in progress to change the internal encoding of
>> Emacs from Mule to something based on Unicode (ie, Unicode with some
>> extensions).  When this happens, all ä characters will be the same
>> inside of Emacs.  Then the searching problem will be gone.  (It also
>> means that Pavel's name will be affected, one way or another.)
>
> When will that be in a usable state ?

Who knows?  I'm sure even the people doing the work don't know.  Such
is (often) the nature of free software: people do the work when they
have time.

But I can say that 21.3 is planned as a bugfix release.  Then people
are working on 21.4.  And Emacs 22 will surely come after that.  So
don't hold your breath.

kai
-- 
Silence is foo!

Re: bbdb and searching

[Rémi Letot]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

> remi.letot@easynet.be (Rémi Letot) writes:
>
>> Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:
>>
>>> In Emacs from CVS, you can turn on unify-8859-on-decoding-mode and
>>> this will help somewhat because it reduces the number of times that ä
>>> from different charsets appears in your file.  
>>
>> Herm, unify-8859-on-decoding-mode and unify-8859-on-encoding-mode are
>> both on on my machine. And yet I still have different charsets in my
>> bbdb it seems.
>
> Hm.
>
>> Is there some way to convert everything in it to Latin-9 ?
>
> C-x C-f ~/.bbdb RET C-x RET f latin-9 RET C-x C-s

Didn't do the trick. Tried with latin-1 and latin-9, but Rémi doesn't
match my entry. I can't event find it with C-s in ~/.bbdb

Another idea ? :-)

With M-x bbdb I can search it with 'R.mi', but that's not very user
friendly :-)

>
>>> Work is currently in progress to change the internal encoding of
>>> Emacs from Mule to something based on Unicode (ie, Unicode with some
>>> extensions).  When this happens, all ä characters will be the same
>>> inside of Emacs.  Then the searching problem will be gone.  (It also
>>> means that Pavel's name will be affected, one way or another.)
>>
>> When will that be in a usable state ?
>
> Who knows?  I'm sure even the people doing the work don't know.  Such
> is (often) the nature of free software: people do the work when they
> have time.

Yep, I know that and in no way would criticize those people. I
actually would like to contribute something useful, but haven't found
it yet :-)

> But I can say that 21.3 is planned as a bugfix release.  Then people
>are working on 21.4.  And Emacs 22 will surely come after that.  So
>don't hold your breath.

That's the answer I wanted : imminent or not. Thanks, now I know I
have to find a way around that problem. *That* could be usefull :
write a howto for non us-ascii people when the problem is solved.

-- 
Rémi Letot

Re: bbdb and searching

[Rémi Letot]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

> remi.letot@easynet.be (Rémi Letot) writes:
>
>> Didn't do the trick. Tried with latin-1 and latin-9, but Rémi doesn't
>> match my entry. I can't event find it with C-s in ~/.bbdb
>
> Open a new buffer (C-x b foo RET), then type Rémi in that buffer.
> Then go to the é character and hit C-u C-x =.  What do you see?

charset: latin-iso8859-15

> Open ~/.bbdb RET and go to the é in Rémi there.  Type C-u C-x =.
> What do you see?

charset: latin-iso8859-1

Clear as water : C-x C-f ~/.bbdb RET C-x RET f latin-9 RET C-x C-s
didn't convert my file to latin-9. I just tried it after seeing that
my é was still latin-1, but that didn't change the result.

Any idea ?

-- 
Rémi Letot

Re: bbdb and searching

[Robert Marshall]

On Wed, 22 May 2002, Kai.Grossjohann@CS.Uni-Dortmund.DE wrote:

> remi.letot@easynet.be (Rémi Letot) writes:
> 
>> Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:
>>
>>> remi.letot@easynet.be (Rémi Letot) writes:
>>>
>>>> Didn't do the trick. Tried with latin-1 and latin-9, but Rémi
>>>> doesn't match my entry. I can't event find it with C-s in ~/.bbdb
>>>
>>> Open a new buffer (C-x b foo RET), then type Rémi in that buffer.
>>> Then go to the é character and hit C-u C-x =.  What do you see?
>>
>> charset: latin-iso8859-15
> 
> Hm.
> 
Double hm, on reading this message I get prompted

Change name Kai Großjohann to Kai Großjohann, is that due to a change
in charset?

> 
> What is the value of the variable bbdb-file-coding-system?  

or maybe I need a version of bbdb with this variable
2.34 doesn't appear to have it

R
-- 
All the slow fish of ignorance
turned toward  the sound
-- Chase Twichell

Re: bbdb and searching

[Kai Großjohann]

remi.letot@easynet.be (Rémi Letot) writes:

> Still no match :-(

Then it's probably the new BBDB version.  Sorry for wasting your time.

> <couic>

What does this stand for?

kai
-- 
Silence is foo!

Re: bbdb and searching

[Josh Huber]

remi.letot@easynet.be (Rémi Letot) writes:

> By the way, is there a way to make gnus automatically insert that
> when I cut text in a post ?

See message-elide-region (C-c C-e) and message-elide-ellipsis.

ttyl,

-- 
Josh Huber

Re: bbdb and searching

[Kai Großjohann]

remi.letot@easynet.be (Rémi Letot) writes:

> Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai writes:
>
>> remi.letot@easynet.be (Rémi Letot) writes:
>>
>>> Still no match :-(
>>
>> Then it's probably the new BBDB version.  Sorry for wasting your time.
>
> You didn't waste my time. Is the new bbdb version usable without
> problems ? If you don't experience the problem I describe, I might
> consider switching.

Well, I encountered some problems related to bbdb-file-coding-system,
but other than that, 2.35 seems worthwhile.

I'm not exactly sure what were the "problems" I alluded to above.
Might as well have been a mistake on my part.  I noticed something
strange.  Maybe the strangeness was that BBDB asked me again and
again to change Kai Großjohann to Kai Großjohann (from Latin-1 to
Latin-9, I guess), even after I had accepted the change and saved
BBDB.  Then I started to mess with the encoding myself, and of course
now I'm responsible for any damage this might have caused.

kai
-- 
Silence is foo!

Re: bbdb and searching

[Rémi Letot]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai writes:

> remi.letot@easynet.be (Rémi Letot) writes:
>
>> Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai writes:
>>
>>> remi.letot@easynet.be (Rémi Letot) writes:
>>>
>>>> Still no match :-(
>>>
>>> Then it's probably the new BBDB version.  Sorry for wasting your time.
>>
>> You didn't waste my time. Is the new bbdb version usable without
>> problems ? If you don't experience the problem I describe, I might
>> consider switching.
>
> Well, I encountered some problems related to bbdb-file-coding-system,
> but other than that, 2.35 seems worthwhile.
>
> I'm not exactly sure what were the "problems" I alluded to above.
> Might as well have been a mistake on my part.  I noticed something
> strange.  Maybe the strangeness was that BBDB asked me again and
> again to change Kai Großjohann to Kai Großjohann (from Latin-1 to
> Latin-9, I guess), even after I had accepted the change and saved
> BBDB.  Then I started to mess with the encoding myself, and of course
> now I'm responsible for any damage this might have caused.

That seems to be related to what I experience : my .bbdb file is
encoded in Latin-1 whatever I do, which prevents me from searching it
with Latin-9 characters.

You problem seems to be the same kind : your name always reverted to
Latin-1. 

Have you found a way around that ?

-- 
Rémi Letot

Re: bbdb and searching

[Erwan David]

remi.letot@easynet.be (Rémi Letot) writes:

> That seems to be related to what I experience : my .bbdb file is
> encoded in Latin-1 whatever I do, which prevents me from searching it
> with Latin-9 characters.
>
> You problem seems to be the same kind : your name always reverted to
> Latin-1. 
>
> Have you found a way around that ?

I've got a similar problem with ispell which does not regonized
characters with accents in the words when I use iso-8859-15.

-- 
 MB un troll ? Meuh non ! Un vrai troll est imprévisible :-)
 -+- LD in: Guide du Cabaliste Usenet - Contrôle de routine -+-