> On 2011-10-31 00:30:58, Michael MacFadd wrote: > > My only small comment is why the creation of an object to be the mutex for > > the synchronized blocks. Why not just use the object itself?
That would work, too, but this logic only needs mutual exclusion to two variables, nothing else in the class. I think it's clearer this way. - soren ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/2632/#review2932 ----------------------------------------------------------- On 2011-10-30 20:18:42, Yuri Zelikov wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/2632/ > ----------------------------------------------------------- > > (Updated 2011-10-30 20:18:42) > > > Review request for wave, Michael MacFadd, vjrj, and soren. > > > Summary > ------- > > Improves persistence logic and addresses some old TODOs. Most of the patch > was contributed by ([email protected]). > -Removes the requirement for persistExecutor being single threaded. > -Refines the persist() logic to make batch successive writes to storage. > -Injects shared instance of persistExecutor instead of creating one instance > per wavelet. > > > Diffs > ----- > > > /src/org/waveprotocol/box/server/waveserver/DeltaStoreBasedWaveletState.java > 1195229 > /src/org/waveprotocol/box/server/waveserver/WaveServerModule.java 1195229 > /test/org/waveprotocol/box/server/waveserver/WaveServerTest.java 1195229 > > Diff: https://reviews.apache.org/r/2632/diff > > > Testing > ------- > > yes. > > > Thanks, > > Yuri > >
