Re: writephonebook vcf, ldiff

[Helge Deller]

On Sunday 12 September 2004 19:51, Pawel Kot wrote:
> > Strange that kaddressbook try to start writing at phone position #1000 -
> > it seems libgnokii delivers wrong entries up to this number for this
> > phone. I suspect a bug in gnokii - after that it should work in
> > kaddressbook as well.
> 
> It *may* be a bug in gnokii, but I have no idea how kaddressbook gets this
> #1000 position. If you provide the information how is the location
> retrieved from gnokii I could probably fix it.

Citing Oliver's original mail here:
addressbook: WARNING: Try to write entry 'somename' at phone_entry_no=1000, 
phone_count=0"
1000 times (phone_entry_no counting from 1 to 1000).
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^

This is exactly what the export filter does if it's asked to append entries.
It starts with the first entry of the phone (phone_entry_no=1), 
reads this entry:
error = gn_sm_functions(GN_OP_ReadPhonebook, &data, &state);

and only tries to overwrite this location, either when the returned error code 
is GN_ERR_EMPTYLOCATION, or if the entry.empty field is true:
        if (error == GN_ERR_EMPTYLOCATION)
                return true;
        if (error == GN_ERR_NONE && entry.empty)
                return true;

So, it seems libgnokii reads all entries from the phone up to entry #1000 and 
all of them seem to contain valid data.

Hope that helps,
Helge