[Libreoffice-qa] [ANN] LibreOffice 4.0.2 RC1 available

[Yifan Jiang]

Hi,

On Fri, Mar 22, 2013 at 12:17:35PM +0100, Petr Mladek wrote:

> > So it may make sense to keep the concept as simple as "operating
> > system locale".
> 
> I would use the locale that is used in "Tools/Options/Language
> Settings/Languages/Locale Setting". I think that it affects more things
> than UI Language or the other settings.

Yes, I agree with that. It makes sense to make it as a "confirm point"
what the test locale is if any conflict (for example not have l10n
pacakges installed). :)

> > Alternatively these language specific testing would be reflected in
> > *testcases* description and steps:
> > 
> >     * Ideally these language specific *testcases* should be tested in
> >       relevant environments.
> > 
> >     * Conversely those non-language specific *testcases* should be
> >       tested in only one environment.
> 
> Are these two types of test cases already somehow marked in Moztrap?
> Unfortunately, manual-test.libreoffice.org is not accessible right
> now.

No, they were not implemented. All cases are equally distributed in
each test environment.

> >     which I think are Moztrap's future roles to take care about how to
> >     map different cases to different locales.
> > 
> > Finally as far as I can concern, everything manually set, inluding
> > language settings, which is relevant to the results should go to
> > "comments" field when marking the test results.
> 
> It might be too long list :-) I think that it makes sense to mention
> extra information only there are problems or if you are not sure about
> the test results and some information looks important.

Yes, only relevant settings, which usually cause a failure/invalid
should be noted.

Actually there seems no way to comment on a Pass mark.

> > > Why is the Install test (which is first action you take) numbered
> > > 16?
> > 
> > hehe, I'll have to hide it if really confusing.
> 
> Well, the information might help to discuss particular test case. It is
> similar to bugzilla. We speak about bug XXXXX. We could speak also about
> test case XX.

wrt the last column:

    http://manual-test.libreoffice.org/runtests/run/40/env/642/

I think it was the undocumented "order" id (instead of test case id)
have got concerned. It seems a new feature and I just did a slight dig
to the code base, the feature is designed to allow test cases sorting
in *run* by editing *testcases* order in *testsuite* and editing
*testsuite* order in *testrun*.

For example, we have a test run with this layout:

testrun:
    testsuite_A:
        - case_a
        - case_b    
    testsuite_B:
        - case_c
        - case_d

where the case order is drag/drop-adjustable when editing
testsuite. The above layout will generate test cases order in this run
when people start to test:

    case_a 1
    case_b 2
    case_c 3
    case_d 4

Meanwhile, as mentioned, the *order* does not only depend on cases
position in suite, but also depend on suite position in a run:

testrun:
    testsuite_B:
        - case_d
        - case_c
    testsuite_A:
        - case_a
        - case_b

will lead =>

    case_d 1
    case_c 2
    case_a 3
    case_b 4

In conclusion, it brings Admin people more flexibility to decide which
test cases should be put on the top of Tester's run list.

Meanwhile we should be more cautious to create test suite and run
since the cases and suites are now order sensitive :)

Sadly the order cannot be updated for a run already created...so we
will do this better in next regression test, at leaset with high
priority test cases on the top of the test list.

Thanks to raise up this actually and I'll update wiki later.

Best wishes,
Yifan

-- 
  Yifan Jiang
  Libreoffice / SUSE
  Contact: yifan - irc.freenode.net/libreoffice
  =============================================  
  http://www.libreoffice.org/
  http://www.documentfoundation.org/