Comment 5 for bug 412058

Revision history for this message
Gary van der Merwe (garyvdm) wrote :

So there are 2 thing we can try first.

1. Reduce the time before we run the initial_load. It was 1000 msec. We can change that to 0.

2. 2 Steps: Change Repository.find_branches in bzrlib to a Generator. Then when we iterate through find_branches, we should call self.processEvents()

Lets try this before we start trying out more complex architecture.
Here is my branch with the qbzr changes: lp:~garyvdm/qbzr/qswitch. The changes to bzr are here: lp:~garyvdm/bzr/find_branches_generator. I will submit a merge request as soon as I have run the entire test suit.