While the first interpreters where busy filling the language databases we recognized that the performance gets pure on the “Entries”-page at a certain amount of languages and/or keys.
The oTranCe team reacted immediately and already could find a solution for this behavior.
Before, the search mechanism took effect in both, language keys and translations. Additionally, a filter was applied for languages, assigned templates and for not yet translated phrases. Thus, the generated SQL query became too complex and brought the MySQL server to its knees. Putting out the search results, an additional query was generated per output line to pick up all related information.
To avert this performance issues, oTranCe now became more intelligent. The single different queries have been capsuled so that every single area can care about its specific requirements. Moreover, the additional output data will be picked up in a single step and thus will keep the load of the MySQL server on a lower level.
At the end of the day, a much higher speed of replies has been reached. Interpreters don’t have to wait too long for the reaction of the surface and can work much faster.
You can download the changes with a new SVN checkout.