On Thu, Jul 18, 2013 at 8:40 AM, Ali Lown <a...@lown.me.uk> wrote:
>> On July 18, 2013, 3:13 p.m., Vicente J. Ruiz Jurado wrote:
>> > From my point of view the File Persistence still have bugs that we have 
>> > not addressed and aren't present in Memory Store, so I usually compare the 
>> > implementations to try to see the differences and fix it.
>> >
>> > In general there are more corruptions in waves and more indexing problems.
>> >
>> > So to remove this code before fix these issues scare me.
>
> I agree that anecdotally I see corruptions and indexing problems using the 
> file persistence systems.
> But anecdotally, they don't disappear completely using the memory systems 
> either...

Sounds like we need more tests. Although if the problems don't
disappear completely using the other file systems its possible they're
concurrency issues.

-J

> It is impossible to be running an actual server using the memory persistence 
> systems, so it serves little (no) purpose remaining in the codebase.
>
> [If it is simply that you use the code for reference, it will always be there 
> in the source control system, just not taking up space (and time) during 
> compilation+release].
>
>
> - Ali
>
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/12678/#review23383
> -----------------------------------------------------------
>
>
> On July 17, 2013, 3:20 p.m., Ali Lown wrote:
>>
>> -----------------------------------------------------------
>> This is an automatically generated e-mail. To reply, visit:
>> https://reviews.apache.org/r/12678/
>> -----------------------------------------------------------
>>
>> (Updated July 17, 2013, 3:20 p.m.)
>>
>>
>> Review request for wave, Bruno Gonzalez, Vicente J. Ruiz Jurado, and Yuri 
>> Zelikov.
>>
>>
>> Repository: wave-git
>>
>>
>> Description
>> -------
>>
>> Following on from the MongoDB removal, I then pondered why these 
>> Memory-backed stores still exist.
>> The only reason I could come up with was the test-suite, so this patch also 
>> fixes the test-suite to use File-backed stores.
>>
>> This removes a second implementation of everything (which we have to 'test' 
>> before every release) making that easier, removes hundreds of LOC to 
>> maintain and adjust if we want to add a new field/etc.
>>
>> I had also started to notice discrepancies in behaviour between 
>> implementation backings (e.g. exceptional case behaviour differences for 
>> lucene/memory search backing).
>>
>> Nobody is running an actual server using memory-backing, so we no longer 
>> have a need to keep this code. (It was primarily developed during 
>> bootstrapping until some actual persistence exists).
>>
>> I also considered removing the config switches etc, but decided to not 
>> delete too much in one review :P
>>
>>
>> Diffs
>> -----
>>
>>   server.config.example 19ba8b2
>>   src/org/waveprotocol/box/server/CoreSettings.java 5beaf65
>>   src/org/waveprotocol/box/server/SearchModule.java 2de0ef9
>>   
>> src/org/waveprotocol/box/server/persistence/FakePermissiveAccountStore.java 
>> fcefc18
>>   src/org/waveprotocol/box/server/persistence/PersistenceModule.java a430570
>>   
>> src/org/waveprotocol/box/server/persistence/memory/MemoryDeltaCollection.java
>>  da6a85e
>>   src/org/waveprotocol/box/server/persistence/memory/MemoryDeltaStore.java 
>> db1606a
>>   src/org/waveprotocol/box/server/persistence/memory/MemoryStore.java de0cbda
>>   src/org/waveprotocol/box/server/waveserver/ImportServlet.java 67c4ee4
>>   
>> src/org/waveprotocol/box/server/waveserver/MemoryPerUserWaveViewHandlerImpl.java
>>  d995d5e
>>   src/org/waveprotocol/box/server/waveserver/MemoryWaveIndexerImpl.java 
>> c5c6fe4
>>   
>> test/org/waveprotocol/box/server/authentication/AccountStoreLoginModuleTest.java
>>  48f905f
>>   test/org/waveprotocol/box/server/authentication/SessionManagerTest.java 
>> 461cd03
>>   test/org/waveprotocol/box/server/persistence/memory/AccountStoreTest.java 
>> f89f08a
>>   test/org/waveprotocol/box/server/persistence/memory/CertPathStoreTest.java 
>> 19fabbd
>>   test/org/waveprotocol/box/server/persistence/memory/DeltaStoreTest.java 
>> 2e85646
>>   test/org/waveprotocol/box/server/rpc/AuthenticationServletTest.java 2e39d2d
>>   test/org/waveprotocol/box/server/rpc/FetchServletTest.java 2ae4dbc
>>   test/org/waveprotocol/box/server/rpc/UserRegistrationServletTest.java 
>> bd83db8
>>   
>> test/org/waveprotocol/box/server/waveserver/CertificateManagerImplTest.java 
>> 75ac795
>>   
>> test/org/waveprotocol/box/server/waveserver/DeltaStoreBasedWaveletStateTest.java
>>  6b09778
>>   
>> test/org/waveprotocol/box/server/waveserver/LocalWaveletContainerImplTest.java
>>  5355e2b
>>   
>> test/org/waveprotocol/box/server/waveserver/MemoryPerUserWaveViewProviderTest.java
>>  89006a4
>>   
>> test/org/waveprotocol/box/server/waveserver/SimpleSearchProviderImplTest.java
>>  30e0c2d
>>   test/org/waveprotocol/box/server/waveserver/WaveMapTest.java e161490
>>   test/org/waveprotocol/box/server/waveserver/WaveServerTest.java 1da4f7b
>>   test/org/waveprotocol/box/server/waveserver/WaveletContainerTest.java 
>> 92d5baa
>>   test/org/waveprotocol/box/server/waveserver/WaveletStateTestBase.java 
>> db3e86e
>>
>> Diff: https://reviews.apache.org/r/12678/diff/
>>
>>
>> Testing
>> -------
>>
>> Builds and passes test suite (Junit maintains temporary directories for us).
>> The composition of all 7 of these 'related' (but independent) patches is 
>> verified to still work as a wave server.
>>
>>
>> Thanks,
>>
>> Ali Lown
>>
>>
>

Reply via email to