> On Sept. 19, 2016, 4:55 p.m., haosdent huang wrote: > > src/common/resources.cpp, lines 616-648 > > <https://reviews.apache.org/r/52001/diff/1/?file=1501134#file1501134line616> > > > > This is conflict with range resources `ports:[21000-24000,30000-34000]`. > > Anindya Sinha wrote: > I do not think so. `ports:[21000-24000,30000-34000]` will have > `key=ports` and `value=[21000-24000,30000-34000]`. So this additional loop > would not be entered since the `key` does not include the square brackets. > Note that key is set based on number of tokens (2 in this specific case, and > 3 when we include MOUNT/PATH disks, ie. `disk[MOUNT:/tmp]:2048`). The new > logic is not applied on value, but only on key (which is whole resource > excluding the value).
Got it, thanks for your explanation! Another question is if we have more attributes in the future, would it use this form? ``` disk[attr_1:value_1,attr_2:value2] ``` - haosdent ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/52001/#review149490 ----------------------------------------------------------- On Sept. 19, 2016, 10:43 p.m., Anindya Sinha wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/52001/ > ----------------------------------------------------------- > > (Updated Sept. 19, 2016, 10:43 p.m.) > > > Review request for mesos and Jiang Yan Xu. > > > Bugs: MESOS-6062 > https://issues.apache.org/jira/browse/MESOS-6062 > > > Repository: mesos > > > Description > ------- > > Text based representation of disk resources can indicate source type > (ie. PATH or MOUNT) and its root. This makes it closer to the JSON > representation. > > > Diffs > ----- > > src/common/resources.cpp 4602bff22ec7ff77f069056085ad92bfa04595f3 > src/v1/resources.cpp 7c4db44d4c31a8295cc16b0ce1ef5adf314e6ac6 > > Diff: https://reviews.apache.org/r/52001/diff/ > > > Testing > ------- > > > Thanks, > > Anindya Sinha > >
