umlauts garbled up

[Derek Atkins]

Eildert Groeneveld <Eildert.Groeneveld at fal.de> writes:

> On Friday 10 March 2006 18:41, Derek Atkins wrote:
>> Was this a datafile created with 1.8?  Or a file created with 1.9?
>> Data files created with 1.8 are potentially completely screwed up..
> yes, it came all the way from 1.1 ... 1.8 :)
>> It's VERY possible that they have a combination of iso-8859 AND utf8
>> encodings in the same XML file!
> that was apparently the case. I fixed this now with a editor (set to utf8) and 
> it seems to be ok again. (I like the xml format!)
>
> This is then probably worth a line in the migration docs?

Probably, yes.  It would be nice if we had a good, automated
upgrade path, but unfortunately the problem is gnucash-1.8.  :(
It screwed up.

What we probably need to do is write a program that goes through the
file and finds every non-ascii "character" and asks the user what
charset the character is from, perhaps giving them a choice of
different charsets and what the character would be..  (assuming of
course we can figure that out).  Then of course we can attempt to
remember those choices..  And then perform the various necessary
character conversions to produce a utf8 file.

Someone want to work on a small program to do that?

-derek

-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord at MIT.EDU                        PGP key available