Have completion in filter to match unaccented names

[risto3]

On a side question, we’ve been using unaccent for a long while, and it seems okay.
But apparently there are cases not treated, namely translated field names.
How to get these also diacritical insensitive.
For example, in account plan, if I would like (in French) to search all reconciled or unreconciled entries, it would be nice to be able to type ‘re’ and let command completion work but it only works for ‘ré’ (Réconciliation) … curiously enough, it is case insensitive though. (similar with débit/crédit)
Would it be feasible to use unicodedata.normalize or unidecode here?

[pokoli (Sergi Almacellas Abellana)]

@risto3

For example, in account plan, if I would like (in French) to search all reconciled or unreconciled entries, it would be nice to be able to type ‘re’ and let command completion work but it only works for ‘ré’ (Réconciliation) … curiously enough, it is case insensitive though. (similar with débit/crédit)
Would it be feasible to use unicodedata.normalize or unidecode here?

Indeed thats a separate topic. You want to add unaccent feature to search filter autocompletion, but the feature was to add unaccent functionality on the database values.

[risto3]

@pokoli

Indeed thats a separate topic. You want to add unaccent feature to search filter autocompletion, but the feature was to add unaccent functionality on the database values.

I see no specific request for database values in the initial posting.

[ced (Cédric Krier)]

I think we could use Unidecode and unidecode to make string and value comparison for the completion in the domain parser. But I would make it an optional dependency.