Re: Can I add the charset parameter to the pilot device configuration?

["lsla post cz" <lsla post cz>]

> or add-ons applied. If there is a Klingon charset add-on, should we 
> support that in pilot-link?
> 

I studied the code in other packages which use pilot-link libraries. 

1) KPilot has an charset configuration option, but it does not use the
convert_* functions from pilot-link. Instead, KDE codecs are used.

2) Jpilot also has such a option. Again, convert_* functions are not
used, conversion tables are in the Jpilot source.

3) Pilot-link itself never calls convert_* .

4) convert_* routines were probably added to pilot-link just because
evolution plugins and gnome pilot.

5) Charset is  a pilot (GPilotPilot) dependent property, therefore the
code in pilot-link should be modified.

6) The code in pilot link can be easily replaced by the GLib function
g_convert() removing pilot-link dependency on iconv.

I think that the best solution is:

1) To move charset support from pilot-link to gnome-pilot, using
g_convert instead of libiconv. The charset property  will be a part of
GPilotPilot structure.

2) Pilot-link will remain charset independent. If I create my Klingon
pilot hack, I have to create also a Klingon conduit which will  do the
conversion. Conversions for most natural language charsets are 
available in most environments (g_convert, kde codecs, python and perl
packages ...).

3) This will remove the dependency on libiconv and also the evolution
configure script will be simpler.

I am just experimenting with this. I can set charset in config file and
copy from/to pilot works fine. I have problems with synchronizing
records, still. After I remove the bugs, I will send the patch here, and
also to pilot-link and evolution lists for considering.


> > We can check to see if it is supported by iconv and listed in the 
> > current PalmLocale.h.  There's certainly some work matching the palm 
> > encoding types to iconv charset identifiers.  How standard are they, 
> > anyway?
> 

Let the user choose his/her encoding.

Ladislav Sladecek