We need automated tests showing that concurrent editing converges. This is the main evidence that the feature works.
The integration tests should generate random concurrent edits rather than only hand-picked cases. The interesting bugs tend to be the ones nobody thought to write down.
Depends on #18, #19 and #20.
We need automated tests showing that concurrent editing converges. This is the main evidence that the feature works.
The integration tests should generate random concurrent edits rather than only hand-picked cases. The interesting bugs tend to be the ones nobody thought to write down.
Depends on #18, #19 and #20.