[FIX] models: avoid Unicode error when translating that error message

The patch fa492d8 has been backported
from Odoo 12.0, that runs on Python 3.

The string '\n\n({} {}, {} {})' to be formatted is a byte-string in
Python 2, while the return value of _() is always a unicode-string.

As format() is (too?) nice, it attempts to convert the unicode-strings
into ascii in order to inject them in the format pattern.
With some languages that are written in ascii, this works -- by chance.
When you use non-ascii languages like Japanese, it fails.

We then fix that issue by using unicode-strings in the formatting
pattern.

#OneCharacterPatch B-)

opw-2032016

-----------------------------

For full technical understanding:

Python 2.7.16 (default, Mar 11 2019, 18:59:25)
[GCC 8.2.1 20181127] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> '{}'.format('test')
'test'
>>> '{}'.format(u'test')
'test'
>>> '{}'.format(u'エ')
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
UnicodeEncodeError: 'ascii' codec can't encode character u'\u30a8' in position 0: ordinal not in range(128)
>>> u'{}'.format(u'エ')
u'\u30a8'

closes #34698

Signed-off-by: Richard Mathot (rim) <rim@openerp.com>

odoo/models.py

@@ -4474,7 +4474,7 @@ def exists(self):
             # mark missing records in cache with a failed value
             exc = MissingError(
                 _("Record does not exist or has been deleted.")
-                + '\n\n({} {}, {} {})'.format(_('Records:'), (self - existing).ids[:6], _('User:'), self._uid)
+                + u'\n\n({} {}, {} {})'.format(_('Records:'), (self - existing).ids[:6], _('User:'), self._uid)
             )
             (self - existing)._cache.update(FailedValue(exc))
         return existing