You may be familiar[1] with the c++ test utilities qpid-send,
qpid-receive and the qpid-cpp-benchmark script that harnesses them for
simple performance testing.
Pavel Moravec has been good enough to contribute a python port of these
utilities[2] and a modification (and rename) for qpid-cpp-benchmark[3]
to allow the script to drive either version of the utilities (in any
combination). (Thanks, Pavel!)
There are a couple of questions I wanted to ask before getting these
contributions into the tree.
(i) Where should the python utilities live? I was thinking about
creating a 'perf' directory under examples (alongside 'api) and putting
them in there. Though they aren't restricted to perf testing, that would
see to at least be obvious location).
(ii) Is everyone ok with renaming qpid-cpp-benchmark? (Just taking the
'cpp' out of it, since it can now run against either python or c++ - the
script itself is python). The script will otherwise be entirely backward
compatible, so for those who do rely on the current name and can't
change a symbolic link should be fine.
(iii) Should we move the qpid-[cpp]-benchmark script out of the c++
tree? This would be logically more consistent I think, e.g. it could go
under extras? (It could go under tools, but we may want to distinguish
between management tools and test utilities).
Any thoughts or opinions?
--Gordon.
[1] For those who are not, they are essentially richer versions of the
spout and drain examples. They are not restricted to performance
testing, but do provide a simple way to explore the performance impact
of different settings/choices.
[2] https://issues.apache.org/jira/browse/QPID-4115
[3] https://issues.apache.org/jira/browse/QPID-4116
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
