Problem

The repetition editor may runs into timeouts with larger tasks / multiple (> ~20 or 50) repetitions.

Also saving a single repetition may need 1-2 seconds, which is to long for a fast working translator.

First more details about the problem are needed:

• how is repetitions implemented in the backend ?
• can we speed it up by send only the first / master segment to termtagger / spellchecker and apply the tagged content to the repetitions. Is that already done that way or do we send each segment repetition to the mentioned services? Can we benefit from TRANSLATE-4353?

Solution

An async operation can be very likely implemented too. What mus be kept in mind: The user can configure if the repetition editor is popping up, or all repetitions are saved automatically.

• The saving could also be done in a worker (see TRANSLATE-4353) so that the pure request to start it is not running into a time out.
• How do we communicate the problems back to the user when now asynchronous?
• All segments, that are loaded already in the UI and which are affected repetitions should be locked → should already be the case
• All segments in the back-end where something will be saved should be locked
• When search and replace is done, UI is notified via messagebus, that segments are unlocked. → already implemented? / Implement along with TRANSLATE-4353?