Bug 13974

Description of problem:

Kaddressbook allows to export addressbook data to files with vcard format (.vcf suffix). When it generates long lines, it does not correctly split such lines as specified by the standards:
  - according to the standard, it should do the split at a white-space. replacing the white-space by a CR+white-space sequence;
  - but it does not do such a replacement, rather it INSERTS a CR+white-space sequence at apparently arbitrary places in the data line

In the attached example, the UTF sequence representing the ü character
    =C3=A9
has been replaced by
    =C3<CR> =A9

This evidently will create serious problems in the programm that uses the .vcf file for input - it will upset the UTF encoding mechanism (or simply insert arbitrary white spaces).

I have only observed this problem in lines that contain UTF-encoded strings (all my other data lines are short and not subject to splitting) - possibly, but not likely, the problem is limited to encoded strings.


Version-Release number of selected component (if applicable):
kaddressbook-4.13.97-1.mga5


How reproducible:
always (if the addressbook contains long lines that need to be splitting)


Steps to Reproduce:

1. Open kaddressbook on an address book that contains some sufficiently lines

2. Do File -> Export -> Export vCard 2.1
   ( my details were "all Contacts", standard options as proposed, "Export to One file")

3. Check the contents of the so generated file


Kaddressbook is an ideal tool to create and manage address data, particularly where format conversion is necessary for synchronising between Linux and Android platforms - this bug is annoying (and difficult to discern as the origin for synchronisation failure).

I join attachments with an example of a vcard item that has been incorrectly split, and references to documents that specify the line splitting procedure.


Reproducible: 

Steps to Reproduce: