bbdb acting up with Kai's name

bbdb acting up with Kai's name

[Josh Huber]

Every time I view a message from kai, the BBDB asks me if I want to
switch his name to "Kai Gro".  I think this is due to
bbdb-extract-address-components.  What do people get from:

(bbdb-extract-address-components
 "Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann)")

What is your value of bbdb-extract-address-components-func?  Is it the
default bbdb-rfc822-addresses?  Perhaps I have an old (broken?)
version of bbdb?

;;; $Id: bbdb-snarf.el,v 1.6 2002/02/12 22:56:38 youngs Exp $

Switching the *-func variable to bbdb-extract-address-components fixes
the problem for me...which is good ;)

Just FYI, if you're having issues, be sure to:

(setq bbdb-extract-address-components-func
      'bbdb-extract-address-components)

have a nice day,

-- 
Josh Huber

Re: bbdb acting up with Kai's name

[Josh Huber]

Josh Huber <huber@alum.wpi.edu> writes:

> Switching the *-func variable to bbdb-extract-address-components
> fixes the problem for me...which is good ;)

I spoke too soon -- something else must have fixed it.  It must have
been when I was playing around with
gnus-extract-address-components. (which DOES do the right thing)

Perhaps this has something to do with my encoding setup?

thanks,

-- 
Josh Huber

Re: bbdb acting up with Kai's name

[Kai Großjohann]

Josh Huber <huber@alum.wpi.edu> writes:

> Every time I view a message from kai, the BBDB asks me if I want to
> switch his name to "Kai Gro".

Maybe this is a symptom of your XEmacs not knowing Latin-9?  The
Latin-Unity package teaches XEmacs about Latin-9.

Hm.  But then, you're using Ben's new unicodified version.  It might
grok Latin-9 out of the box and it might do 8859 unification out of
the box.  I don't know enough about XEmacs to be sure.

kai
-- 
Silence is foo!

Re: bbdb acting up with Kai's name

[Josh Huber]

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

> Maybe this is a symptom of your XEmacs not knowing Latin-9?  The
> Latin-Unity package teaches XEmacs about Latin-9.

Perhaps, but I'm using the latin-unity package, and I've tested saving
files with different encodings, and everything seems sane.  Of course,
it's very suspicious ;)  I'll try and figure out which regexp isn't
matching right -- perhaps xemacs isn't treating that character as a
word char because of the different encoding.

> Hm.  But then, you're using Ben's new unicodified version.  It might
> grok Latin-9 out of the box and it might do 8859 unification out of
> the box.  I don't know enough about XEmacs to be sure.

Right...actually, I'm not sure if it does do Latin-9 out of the box,
but I don't think so.  AFAIK latin-unity is appropriate for both
released and unreleased versions...but I could be wrong here ;)

Thanks for the tips...

-- 
Josh Huber

Re: bbdb acting up with Kai's name

[Kai Großjohann]

John S Cooper <John.Cooper@eu.citrix.com> writes:

> BBDB prompts me to change "Kai Großjohann" to "Kai Gro<X>johann" where <X> is
> an outline box character (to denote a char not available?).
>
> I've attached a couple of screenshots that show how the character is encoded
> and that char value 0xf5f is what ends up getting displayed.
>
> If anyone can tell me how to avoid this it would be much appreciated.

This is Latin-9-ß.  (You could have done C-u C-x = instead of C-x = to
see this, I think.)

Isn't it enough to just install a Latin-9 font?  (Of course, that will
help with the box only.)

You could also store .bbdb in an encoding which differentiates between
Latin-1 and Latin-9 (emacs-mule and the iso-2022 encodings come to
mind) and then store two versions of my name.  Newer bbdb versions
have a variable bbdb-file-coding-system.  You must make sure that the
file on disk agrees with this coding system, though:

C-x C-f ~/.bbdb RET
C-x RET f foo RET
C-x C-s

Then set bbdb-file-coding-system to foo.

kai
-- 
Silence is foo!

Re: bbdb acting up with Kai's name

[John S Cooper]

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

 > John S Cooper <John.Cooper@eu.citrix.com> writes:
 > 
 > > BBDB prompts me to change "Kai Großjohann" to "Kai Gro<X>johann" where <X> is
 > > an outline box character (to denote a char not available?).
 > >
 > > I've attached a couple of screenshots that show how the character is encoded
 > > and that char value 0xf5f is what ends up getting displayed.
 > >
 > > If anyone can tell me how to avoid this it would be much appreciated.
 > 
 > This is Latin-9-ß.  (You could have done C-u C-x = instead of C-x = to
 > see this, I think.)
 > 
 > Isn't it enough to just install a Latin-9 font?  (Of course, that will help
 > with the box only.)

I suppose so - if I use the following font, the character is displayed
correctly:
  -b&h-lucida console-medium-r-semi condensed--17-120-100-100-m-100-iso8859-15

However, there doesn't seem to be an iso-8859-15 variant of the font I like for
my Gnus frame, namely:
  -etl-fixed-medium-r-normal-*-*-160-*-*-c-*-iso8859-1

 > You could also store .bbdb in an encoding which differentiates between
 > Latin-1 and Latin-9 (emacs-mule and the iso-2022 encodings come to
 > mind) and then store two versions of my name.  Newer bbdb versions
 > have a variable bbdb-file-coding-system.  You must make sure that the
 > file on disk agrees with this coding system, though:
 > 
 > C-x C-f ~/.bbdb RET
 > C-x RET f foo RET
 > C-x C-s
 > 
 > Then set bbdb-file-coding-system to foo.

Hmm, my bbdb doesn't have this variable - I'm using 2.34 which
bbdb.sourceforge.org claims to be the latest released version.

Perhaps it's easiest for me to just put up with the outline character and
figure out how to tell BBDB to just quietly accept duplicate names.

    --- John

Re: bbdb acting up with Kai's name

[Kai Großjohann]

John Cooper <john.cooper@eu.citrix.com> writes:

> I suppose so - if I use the following font, the character is displayed
> correctly:
>   -b&h-lucida console-medium-r-semi condensed--17-120-100-100-m-100-iso8859-15
>
> However, there doesn't seem to be an iso-8859-15 variant of the font I like for
> my Gnus frame, namely:
>   -etl-fixed-medium-r-normal-*-*-160-*-*-c-*-iso8859-1

Indeed.  I wonder if the GNU intlfonts package will be augmented?

>  > You could also store .bbdb in an encoding which differentiates between
>  > Latin-1 and Latin-9 (emacs-mule and the iso-2022 encodings come to
>  > mind) and then store two versions of my name.  Newer bbdb versions
>  > have a variable bbdb-file-coding-system.  You must make sure that the
>  > file on disk agrees with this coding system, though:
>  > 
>  > C-x C-f ~/.bbdb RET
>  > C-x RET f foo RET
>  > C-x C-s
>  > 
>  > Then set bbdb-file-coding-system to foo.
>
> Hmm, my bbdb doesn't have this variable - I'm using 2.34 which
> bbdb.sourceforge.org claims to be the latest released version.

Oh.  Why do I have 2.35, then?  Hm.

> Perhaps it's easiest for me to just put up with the outline
> character and figure out how to tell BBDB to just quietly accept
> duplicate names.

If you store that one variation of my name, then you will be fine.
(Until I move to Poland :-)

kai
-- 
Silence is foo!

Re: bbdb acting up with Kai's name

[Josh Huber]

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

> [snip]

I've also noticed that it's not bbdb's fault that things are screwing
up for me. (well, not entirely)

bbdb uses either mail-extract-address-components or rfc-822-addresses
which are both part of mail-lib.

Could someone (Kai?) using Emacs 21 which Latin-9 support try these:

(let ((hdr "Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann)"))
  (string-match (car (nth 3 bbdb-extract-address-component-regexps))
		hdr)
  (mail-extract-address-components (match-string 2 hdr)))

(let ((hdr "Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann)"))
  (bbdb-rfc822-addresses hdr))

rfc-822-addresses does this:

(rfc822-looking-at "[^][\000-\037 ()<>@,;:\\\".]+")

Which works for a Latin-1 ß (0xdf), but not a Latin-9 ß (0x6df)

Thanks for the help.

-- 
Josh Huber

Re: bbdb acting up with Kai's name

[Kai Großjohann]

Josh Huber <huber@alum.wpi.edu> writes:

> Could someone (Kai?) using Emacs 21 which Latin-9 support try these:
>
> (let ((hdr "Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann)"))
>   (string-match (car (nth 3 bbdb-extract-address-component-regexps))
> 		hdr)
>   (mail-extract-address-components (match-string 2 hdr)))

=> (nil "Kai Großjohann")

> (let ((hdr "Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann)"))
>   (bbdb-rfc822-addresses hdr))

=> (("Kai Großjohann" "Kai.Grossjohann@CS.Uni-Dortmund.DE"))

Does this help?  The ß are iso-8859-15 characters.  It's not clear to
me whether this means that it works, or that it doesn't work.

kai
-- 
Silence is foo!

Re: bbdb acting up with Kai's name

[Kai Großjohann]

Josh Huber <huber@alum.wpi.edu> writes:

> Could someone (Kai?) using Emacs 21 which Latin-9 support try these:
>
> (let ((hdr "Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann)"))
>   (string-match (car (nth 3 bbdb-extract-address-component-regexps))
> 		hdr)
>   (mail-extract-address-components (match-string 2 hdr)))

=> (nil "Kai Großjohann")

> (let ((hdr "Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann)"))
>   (bbdb-rfc822-addresses hdr))

=> (("Kai Großjohann" "Kai.Grossjohann@CS.Uni-Dortmund.DE"))

Does this help?  The ß are iso-8859-15 characters.  It's not clear to
me whether this means that it works, or that it doesn't work.

kai
-- 
Silence is foo!