Re: RepositoryGrowthTest Benchmark

2017-03-03 Thread Andrei Dulceanu
Let's remove it for the time being and add it only if it really proves to be valuable. I created OAK-5884 [2] to track this. [2] https://issues.apache.org/jira/browse/OAK-5884 2017-03-03 11:30 GMT+02:00 Davide Giannella : > On 02/03/2017 16:59, Francesco Mari wrote: > > Since there is obviously

Re: RepositoryGrowthTest Benchmark

2017-03-03 Thread Davide Giannella
On 02/03/2017 16:59, Francesco Mari wrote: > Since there is obviously some work to do to adapt > RepositoryGrowthTest, it is worth asking how many times somebody > checked the results of this benchmark since it was introduced in 2014. > If nobody finds RepositoryGrowthTest useful, we might as well

Re: RepositoryGrowthTest Benchmark

2017-03-02 Thread Francesco Mari
Assuming that we decide to adapt RepositoryGrowthTest to use SegmentTarFixture, the new FileStore implementation doesn't expose a public size() method. This means that RepositoryGrowthTest will not be able to access the repository size. >From an implementation point of view, RepositoryGrowthTest v

RepositoryGrowthTest Benchmark

2017-03-02 Thread Andrei Dulceanu
Hi, Working on OAK-5834 [1], which is about removing the deprecated oak-segment module, I came across RepositoryGrowthTest. This benchmark uses *only* the deprecated SegmentFixture. Does anyone have any hints regarding what to do with this benchmark? Is it safe to remove it entirely, or does it ne