abbrev_defs encoding issues

abbrev_defs encoding issues

[Andreas Röhler]

Hi,

when using Emacs 24.. and Emacs 23.. series parallel or
consecutive, I'm running into trouble due to different
encoding for ~/.abbrev_defs - 23 uses emacs-mule, while 24 uses utf-8.

Having a look into NEWS.23

"** The abbrevs file is now a file named abbrev_defs in
user-emacs-directory; but the old location, ~/.abbrev_defs, is used if
that file exists."

that might be avoided to a certain extend by requiring the new location 
for 24 series.

Maybe some warning should user try to change that might help too.

Best regards,

Andreas

--
https://code.launchpad.net/~a-roehler/python-mode/components-python-mode
https://code.launchpad.net/s-x-emacs-werkstatt/

Re: abbrev_defs encoding issues

[Stefan Monnier]

> when using Emacs 24.. and Emacs 23.. series parallel or
> consecutive, I'm running into trouble due to different
> encoding for ~/.abbrev_defs - 23 uses emacs-mule, while 24 uses utf-8.

Care to tell us what kind of trouble you bump into?


        Stefan

Re: abbrev_defs encoding issues

[Andreas Röhler]

Am 06.06.2011 18:34, schrieb Stefan Monnier:
>> when using Emacs 24.. and Emacs 23.. series parallel or
>> consecutive, I'm running into trouble due to different
>> encoding for ~/.abbrev_defs - 23 uses emacs-mule, while 24 uses utf-8.
>
> Care to tell us what kind of trouble you bump into?
>
>
>          Stefan
>

if an .abbrev_def is saved from v24, which has been edited from v23 
before, all german umlauts gets destroyed for example:

(the corresponding upper part gets displayed further below in the diff)

<     ("ürr" "Überraschung" nil 0)
<     ("üs" "überaus" nil 0)
<     ("ütr" "Übertragung" nil 0)
<     ("ütre" "übertreiben" nil 0)
<     ("ütro" "übertroffen" nil 0)
<     ("ütt" "erschüttert" nil 0)
<     ("üu" "Übung" nil 0)
<     ("üz" "Überzeugung" nil 0)
<     ("üzd" "überzeugend" nil 0)
<     ("üzn" "Überzeugungen" nil 0)
<     ("üzt" "überzeugt" nil 0)
---
 >     ("��" "da��" nil 0)
 >     ("��h" "Mi��handlung" nil 0)
 >     ("��dt" "��nderte" nil 0)
 >     ("��g" "��gypten" nil 0)
 >     ("��ge" "��gyptische" nil 0)
 >     ("��gen" "��gyptischen" nil 0)
 >     ("��n" "ausl��ndischen" nil 0)
 >     ("��nd" "��nderungen" nil 0)
 >     ("��r" "��lter" nil 0)
 >     ("��u��" "��u��erung" nil 0)
 >     ("��u\201\337" "��u\201\337erung" nil 0)
 >     ("����" "��u��erten" nil 0)
 >     ("��\201\337" "��u\201\337erten" nil 0)

Re: abbrev_defs encoding issues

[Andreas Röhler]

correcting my previous posting:

edits have been done lately in v23, v24 abbrev are destroyed:

RCS file: .abbrev_defs,v
retrieving revision 1.1
diff -r1.1 .abbrev_defs
1c1
< ;;-*-coding: utf-8;-*-
---
 > ;;-*-coding: emacs-mule;-*-
149c149
<     ("staf" "Stammesführer" nil 0)
---
 >     ("staf" "Stammesf��hrer" nil 0)
152c152
<     ("stap" "Stabilitätspakt" nil 0)
---
 >     ("stap" "Stabilit��tspakt" nil 0)
155c155
<     ("stb" "Stabilität" nil 0)
---
 >     ("stb" "Stabilit��t" nil 0)
159c159
<     ("stee" "Steuererhöhung" nil 0)
---
 >     ("stee" "Steuererh��hung" nil 0)

Re: abbrev_defs encoding issues

[Stefan Monnier]

> if an .abbrev_def is saved from v24, which has been edited from v23 before,
> all german umlauts gets destroyed for example:

This diff only shows that the byte-sequence has been changed, not that
the result will behave differently.  Since the encoding is different,
it's no wonder that non-ascii chars get represented differently.

Do you actually see ill-behavior?


        Stefan