[6.0] All strings must be XML compatible: Unicode or ASCII, no NULL bytes; patch provided

[Raphaël Valyi - http://www.akretion.com]

Hello folks,

on v6 (rev #2613) at least, there is an issue if you have some non ASCII chars inside your view. This can be tested, for instance, by installing the l10_br brazilian localization module from here https://code.launchpad.net/~openerp-brazil-team/openerp.pt-br-localiz/trunk

Then click in Localizaçao Brasil > Configuraçao > Fiscal > Codigo de Situaçao Tributaria
you'll get:

[...]
  File "/home/rvalyi/DEV/openerp/server/bin/osv/orm.py", line 2817, in read
    result = self._read_flat(cr, user, select, fields, context, load)
  File "/home/rvalyi/DEV/openerp/server/bin/osv/orm.py", line 2953, in _read_flat
    res2 = self._columns[f].get(cr, self, ids, f, user, context=context, values=res)
  File "/home/rvalyi/DEV/openerp/server/bin/osv/fields.py", line 697, in get
    res = self._fnct(obj, cr, user, ids, name, self._arg, context)
  File "/home/rvalyi/DEV/openerp/server/bin/addons/base/ir/ir_actions.py", line 199, in _search_view
    form_arch = self.pool.get(act.res_model).fields_view_get(cr, uid, False, 'form', context)
  File "/home/rvalyi/DEV/openerp/server/bin/osv/orm.py", line 1635, in fields_view_get
    xarch, xfields = self.__view_look_dom_arch(cr, user, result['arch'], view_id, context=context)
  File "/home/rvalyi/DEV/openerp/server/bin/osv/orm.py", line 1297, in __view_look_dom_arch
    fields_def = self.__view_look_dom(cr, user, node, view_id, context=context)
  File "/home/rvalyi/DEV/openerp/server/bin/osv/orm.py", line 1284, in __view_look_dom
    node.set('string', trans)
  File "lxml.etree.pyx", line 634, in lxml.etree._Element.set (src/lxml/lxml.etree.c:31548)
  File "apihelpers.pxi", line 487, in lxml.etree._setAttributeValue (src/lxml/lxml.etree.c:13896)
  File "apihelpers.pxi", line 1242, in lxml.etree._utf8 (src/lxml/lxml.etree.c:19848)
ValueError: All strings must be XML compatible: Unicode or ASCII, no NULL bytes

In bin/osv/orm.py, around line 1284, the error occurs because the self.pool.get('ir.translation')._get_source is passing us an unicode object, but because you call .encode('utf8') upon it, you transform it back to a string (read http://evanjones.ca/python-utf8.html if uncomfortable with Python + Unicode) and the Etree node.set API doesn't accept string with non ASCII chars.

Actually all you had to do is just use the unicode value that _get_source passed as in my attached patch. It just works and will also make OpenERP a bit faster here.

Hope this helps.

[Jay Vora (Serpent Consulting Services) (jayvora)]

There already exists such a bug for stable.
Read at the comment #15 as it clears about the issues.

Thanks.