On Mon, 2010-04-26 at 00:16 +0100, Jeff wrote:
On Sun, 2010-04-25 at 21:31 +0200, Patrick Ohly wrote:
> On Sat, 2010-04-24 at 23:12 +0100, Jeffrey Perry wrote:
> Intel QA runs tests with Bluetooth sync. They don't test the HTTP
> server. We (= developers) could and should test the SyncEvolution HTTP
> server like we do with other SyncML servers in the nightly testing, but
> we haven't set that up yet.
When the HTTP server configuration is tested, which tests are run?
Various SyncML semantic related tests (different sync modes,
update/delete/add). Large items. Sending various items to the server and
back, then check for data loss ("testItems", test data is in
test/testdata). For a full list you would have to look at the source.
My intent is to avoid testing manually anything you may have covered
with automated tests,
If you want to maximize the bang for the buck, then focus on the data
and properties that you want to have synchronized. We might not have
that covered by our tests.
> > I see there are about 450 CPPUnit tests...very nice.
> I'm not sure how you counted those. What we have and use a lot is the
> automated testing of SyncEvolution as client with various third-party
> HTTP SyncML servers.
The readme says to run client-test with no params to get a list of
tests... I piped it to wc -l and learned that there are about 450 tests
That ends up counting tests multiple times, as they are parameterized:
each test runs once for each data source alone, then also with all
sources enabled. For historic reasons, they are also enabled in two
different orders. This used to make a difference when using the Funambol
engine, not sure whether it still has an effect with Synthesis as
> > Are all the same tests that are run for Bluetooth sync
> > also run for the HTTP server sync scenario?
> Not yet. The same tests could be used for both, but they need be
> parameterized differently (many phones are too dumb to handle the same
> data as SyncML servers). I'm also not sure how to integrate phone
> testing with the nightly test runs - place all phones next to the server
> in the machine room?! Not possible right now, with the server in a
> remote location ;-}
Sounds like you can run all the phone client tests against the HTTP
In principle, yes, although it is more the other way around: the tests
were originally written for HTTP servers, now they can be reused for
Do you do this ?
We run tests against HTTP servers regularly, but not yet against phones.
Best Regards, Patrick Ohly
The content of this message is my personal opinion only and although
I am an employee of Intel, the statements I make here in no way
represent Intel's position on the issue, nor am I authorized to speak
on behalf of Intel on this matter.