Support of utf-8 database [7.8]
Currently YACS encodes every content to Unicode entities before saving it in the database. This is quite efficient and flexible, except that punctuation characters put around Unicode entities just break the MySQL indexing algorithm that support full-text search.
Therefore the necessity we have to natively support UTF-8 database where possible, and to allow for search requests that have accented characters, etc.
Several design questions are still open:
Database experts are welcome to help us...
Therefore the necessity we have to natively support UTF-8 database where possible, and to allow for search requests that have accented characters, etc.
Several design questions are still open:
- can we detect automatically the full support of UTF-8 from MySQL?
- can we ask for it when creating tables in the database ?
- can we change database encoding afterwards ?
- how to migrate previous database content ?
Database experts are welcome to help us...
2- Agnes 
on Sep. 12 2007 -
Hum, je ne suis pas sûre d'avoir absolument tout compris, mais... est-ce que se serait à cause de ça que mon identifiant "Agnès" s'est mis à ne plus fonctionner après le passage en 7.8 et qu'il a fallu changer en enlevant l'accent ?
-----
Agnès
Il n'y a pas de problèmes, que des solutions.
-----
Agnès
Il n'y a pas de problèmes, que des solutions.
3- Bernard 
on Sep. 14 2007 -
Agnes: le problème est effectivement lié au passage vers utf-8. Il faut éditer ton profil et le sauver pour que l'accent soit sauvé en binaire. Tu pourras alors t'authenfier comme il se doit, avec l'accent...
4 rates 
1- Bernard
on Aug. 9 2007 -
If YACS detects that a database server has some support for UTF-8, data will be saved in this format instead of the ASCII Unicode entities we had previously.
This mechanism enables the coexistence of UTF-8 and of Unicode pages in the database. When a Unicode page is edited, it is saved in UTF-8 automatically. Therefore, no conversion script is required to start the full support of UTF-8, up to the database back-end.
One consequence of this new behavior is the better support of search requests in non-ASCII languages...