[Evolution-hackers] Handle Spanish names and surnames properly

[David Marín Carreño <dmarin dyr es>]

Hi all.

After two years, a bug submitted by me (#2982) continues in the Ximian
bugzilla waiting for a bit of love...

The bug is related with the way Evolution splits the whole name into
parts (prefix, name, surname, middlename, suffix...). Also, it's related
to bug #37297.

I don't know if this division works perfectly for English names, but it
doesn't with Spanish names, because of two main problems:

* Here, in Spain, we often use our two surnames. For example, my own
complete name: David Marin Carreño -> Name: David, Surname: Marín
Carreño.

Now, Evolution thinks that my first surname is my middle name, and
indexes my name with my second surname (which is pretty annoying...)

* Other people, in commercial activities, only uses their first surname,
but they have a compound first name, that shouldn't be divided into
First + Middle names. 
For example: Francisco Javier López -> Name: Francisco Javier, Surname:
López.

Evolution now saves this name as "Francisco López", which is
particularly wrong if we know that "Francisco Javier" is usually called
"Javier", and not "Francisco"...


Someone could say that this situation can be solutioned using the "Full
Name" button... But this is bad when I want to quickly introduce new
Spanish contacts into my addressbook... I don't want to press "Full
name" button for make my Spanish contacts to get indexed well.


After seeking the related module, I've found that the bug is related
with the evolution-data-server/addressbook/libebook/e-name-western.c
module, and its extremely rigid way for parsing names.

According to this module, a given name-string is automatically divided
into prefix, first, middle, nick, last and suffix. And this division
doesn't depend on anything else.

I think this is a problem, and there should be a way for changing the
behaviour of e_name_western_parse function, and making it to depend on
some settings, such as the current locale or a given configuration
string extracted from a combo box, or a selection box

For example, for English names, the configuration string could be:

PARSE_STRING = "[PREFIX] FIRST [MIDDLE] LAST [SUFFIX]"
with PREFIX = "Mr.", "Miss", "Mrs.", "Ms.", "Dr"
and SUFFIX = "Jr", "Sr"

and always indexing with LAST and FIRST.

But, for Spanish names, this configuration string could be:

PARSE_STRING = "[PREFIX] FIRST LAST+"
or "[PREFIX] FIRST+ LAST"

That is, a given string "Word1 Word2 Word3" could be parsed as: 
Name: Word1, Surname: Word2 Word3 
or
Name: Word1 Word2, Surname: Word3 

The user should be able to change among the different parsing rules when
adding the contact into the addressbook, in both new-contact and
quick-new-contact dialogs.

What do you think about this solution?


I'm offering help and time to program this question in my weekends, but
I need some help for knowing which other modules should be changed.


Greetings

-- 
David Marín Carreño <dmarin dyr es>
Desarrollo y Recursos, S.L.