> On July 15, 2013, 7:30 p.m., Ben Mahler wrote:
> > Can you provide some context for the use of '=' vs ':' in the resource
> > strings?
> >
> > "cpus=45.55 (*); ports=[10000-20000, 30000-50000] (*); disks={sda1} (*)"
> >
> > vs.
> >
> > "cpus:45.55 (*); ports:[10000-20000, 30000-50000] (*); disks:{sda1} (*)"
We were merging the parse and stringify formats - parse used ':' and stringify
used '='. The choice of ':' was arbitrary, and I'm certainly open to arguments
about why one would be better than the other.
> On July 15, 2013, 7:30 p.m., Ben Mahler wrote:
> > src/tests/resources_tests.cpp, line 89
> > <https://reviews.apache.org/r/12341/diff/2/?file=321794#file321794line89>
> >
> > EXPECT?
Sorry about that.
> On July 15, 2013, 7:30 p.m., Ben Mahler wrote:
> > src/tests/resources_tests.cpp, lines 254-255
> > <https://reviews.apache.org/r/12341/diff/2/?file=321794#file321794line254>
> >
> > EXPECT_LE
I don't think that works - since we're expecting it to be false, we obviously
can't use EXPECT_LE, and we also can't just do the reverse and use EXPECT_GT,
since cpus1 > cpus2 should also be false. (admittedly, I'm not crazy about
these semantics, since if a <= b is false, a > b should be true, but given the
difficulty of comparing Resources, I'm not sure how else to make it work).
> On July 15, 2013, 7:30 p.m., Ben Mahler wrote:
> > src/tests/resources_tests.cpp, lines 262-265
> > <https://reviews.apache.org/r/12341/diff/2/?file=321794#file321794line262>
> >
> > Can you use EXPECT_LE?
Same as above, except for the EXPECT_TRUE.
- Thomas
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/12341/#review23169
-----------------------------------------------------------
On July 15, 2013, 6:36 p.m., Thomas Marshall wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/12341/
> -----------------------------------------------------------
>
> (Updated July 15, 2013, 6:36 p.m.)
>
>
> Review request for mesos and Benjamin Hindman.
>
>
> Bugs: MESOS-505
> https://issues.apache.org/jira/browse/MESOS-505
>
>
> Repository: mesos
>
>
> Description
> -------
>
> Adds parsing of roles to the Resources class. Also updates parse to return a
> Try, in the case that parsing fails, and adds tests for role parsing and for
> Resources operators in the presence of roles.
>
>
> Diffs
> -----
>
> src/common/resources.hpp 42dfb6a
> src/common/resources.cpp 8bc1057
> src/master/drf_sorter.cpp 510acc3
> src/slave/slave.cpp 9a2c671
> src/tests/gc_tests.cpp 1745800
> src/tests/master_tests.cpp 5ac4d5f
> src/tests/resource_offers_tests.cpp beab2d8
> src/tests/resources_tests.cpp 67a11b3
> src/tests/slave_recovery_tests.cpp b0843d7
> src/tests/sorter_tests.cpp 619e786
>
> Diff: https://reviews.apache.org/r/12341/diff/
>
>
> Testing
> -------
>
> make check
>
>
> Thanks,
>
> Thomas Marshall
>
>
