Kaddressbook fails to correctly import utf8 csv-list

[Adrien Cordonnier]

Whereas the default encoding in Kubuntu is utf8, Kaddressbook is expecting imported cvs file to
be in latin1 or latin15 (I don't know which one). Either a option menu should allow to choose
the charset encoding or it should try to detect the encoding automatically or at least, default
to utf8.

[Jonathan Riddell (jr)]

Can you attach an example file?

[Adrien Cordonnier (adrien-cordonnier)]

In Breezy, the bug is not relevant anymore with csv files (a window dialog allows to set encoding) but the same bug
exists when importing VCards.

I attach two files with fake entries encoded in UTF8 (with accents): a csv file and a VCard (v3.0) file.

The same bug also appears when the VCard Addressbook preview is used inside Konqueror (I get chinese characters).

[Adrien Cordonnier (adrien-cordonnier)]

Created an attachment (id=4646)
fake csv address book in UTF8

The bug is not shown anymore with this file.

[Adrien Cordonnier (adrien-cordonnier)]

Created an attachment (id=4647)
fake VCard v3.0 address book in UTF8

The very same address book in VCard in UTF8, I have the same bug as previously
with UTF8 csv files. Strangely enough, with this example, the é are
transformed in chinese (or japanese?) character whereas when importing my
normal addressbook from .kde/share/apps/kabc/std.vcf, é are transformed in
question marks within diamonds (I checked with Kate: both encoding are UTF8).

[Aurel Schwarzentruber (zed-balcab)]

Works for me with the latest KDE packages from Dapper.

[Adrien Cordonnier (adrien-cordonnier)]

Works indeed in Dapper (KAddressBook 3.5).

Csv import lets now choose the encoding with utf8 as default. Vcards are imported as utf8.

[Aurel Schwarzentruber (zed-balcab)]

Fixed in KDE 3.5.x