GNOME Bugzilla – Bug 55144
dbf import don't convert charset
Last modified: 2004-12-22 21:47:04 UTC
It would be good enhancment for non-latin1 ppl (for example cyrillic) to have possibility select charset for imported dbf file and autoconvert it to user charset. (for example, cyrillic dfbs are in cp1251 but linux users as a rule use koi8-r, so they want to set "cp1251" for imported dbf and see it converted to koi8-r). Possibly you can to use e-charset-picker widget from GAL for this.
dbf don't store information about charset (same as XL95). Atm gnumeric convert imported XL95 on the base of user locale (?). I.e. if user locale is ru_RU.KOI8-R, then text strings will be converted from cp1251 (charset for cyrillic XL95) to koi8-r at opening and vise versa at saving. dbf can be produced under any locale and autoconverting can't be made (or we also need autodetect charset for example on the base of characters or character pairs frequency ;-). So any way for selectable converting of textual content will be good. (It can be common plugin, coz text importer has same trouble /can be solved with pre-importing conversion in a shell/)
Created attachment 6280 [details] [review] patch against 0.99
ohh... patch was made by Albert Sultanov
Created attachment 6286 [details] [review] prewious patch had some drawbacks and bugs, this one is fix for them (made by Albert)
Created attachment 6301 [details] [review] translatable strings was marked for gettext (made by Albert Sultanov)
I've used a somewhat simpler patch in cvs. Now that our internals are all utf8 and glib has lots of iconv support routines, this is much easier. Fix is only in 1.1.x
Just to clarify 2 things 1) dbf _does_ store the locale (codepage) both your patch and mine presume this 2) Do you happen to have a test document that I could add to the test suite ? PS Hi frob. Its been a while. How is the baby treating you ?