Hi Vito,
what should I say, thanks for your hard work and work with the
Automation team together to find an easy way to bring localization
into the scripts. There is more than 500.000 lines of code in the
scripts. It is in some parts is simply complex as the Office code.
So it will take a while to create a L10N framework for these tests.
Aa I know you were in contact with Helge Delfs, he is responsible
for the Automation team on OOo. Your knowledge can help to create a
simple framework. Perhaps you will join the OOo Automation team.
On the other side TCM and Automation testing are for different parts
of testing. TCM includes localization test cases, where a native speaker
should check special functionality - eg. is the date format the common
one for their country/language? The Automation testing with TestTool
check general functionality in Office and some parts are only reachable,
if localized strings are included in the test cases. So TCM is for a
part, which cannot be realized by a tooling.
Why TCM doesn't include new features is easy to say. Most of the new
features do not have such functionality, which have to be checked by an
native speaker. Yes, the translation can be checked for the new dialogs,
but this isn't part of TCM testing in general. So for the new features
test case specifications are created. These files are stored on the QA
pages in the TCS area (eg.
http://qa.openoffice.org/ooQAReloaded/TestcaseSpecifications/OpenOffice.org_2.3/Word_Processor/).
Perhaps these test cases can be handled in TCM too. Then general
testing and localization testing is mixed in this tooling. I think,
we should find a place were all test cases can be handled, but this
is another question, which have to be cleared.
Thorsten
Vito Smolej schrieb:
Hi Thorsten:
I may be a snail, but momentarily I am at 294 completed cases out of
309 available in qatesttool (2.1.0 slovenian win). I started 21st of
June, so one can safely assume it takes at least a month of work to get
through all of them (I did some of the three four times, because I
learnt L10n by doing it).. I found 228 errors on the way (not to talk
about warnings). Probably half of the errors are irrelevant or to be
ignored because of incomplete/incorrect localisation and/or context (I
am not sloppy, but its my night/weekend time). Lets say 100 errors are
reportable - and so far I have succeeded to report less than 10 issues.
One month and < 10 issues !? - well, the developers would fall asleep on
their watch, if they had to wait for me. It is all against the basic
idea of (mutatis mutandis) "test early and report often"
What I mean to say is the same thing I meant in my previous post: surely
there must some a priori knowledge or expectation on the part of the
developers, as regards the possible weak spots (like: "run x.bas and
y.bas in framework first and report on the double if you see anything
fishy"). If I could get a list of say 30 most probable *.bas suspects,
I'll be happy (for feeling useful) and the developers will have a much
faster feedback (i.e. within a week) , if anything suspicious surfaces.
Please note that while I am plodding through the tests, I am learning
more and more about the product, so theres definitely some sort of a
payout for me. However, this does not help the product.
Regarding TCM ... correct me, if Im wrong, but, since two years that I
have been playing this game, I have yet to see some NEW use case,
intended to test NEW features, . Old features? Oh well, correct me
(again) if I am wrong, but there's enough old testtool routines to run
regression tests. Maybe TCM gets more errors, but then it just means
theres more and more regressions in the system - and it does not take a
human brain to get them.
Regards to all
Vito
Thorsten Ziehm wrote:
Hi Vito,
you will find test cases for nearly all features and for some
enhancements on the wiki page. Some of them are automated test
cases. Perhaps it is possible to add a list of new test cases,
but why is it needed? We have to find the regressions in the
rest of the Office and not only in the new features. So the
test cases are only on part of the whole project, which have
to be run before the release?
For me it is important, that manual test cases exists for the new
feature. Then it is possible for everybody to work with this feature
and can check in each language (when the translation is integrated and
the builds are available).
Regards,
Thorsten
Vito Smolej wrote:
One practical point for the six:
TCM:
* add test cases for the new functionality * add regression tests
testtool:
* provide the list of testtool scripts, where the new features are
adressed
* provide as list of scripts to test for regressions
and of course, let us know!
Regards
Vito
Maho NAKATA wrote:
Hello, Chiris, Helge, Oliver, Stefan, Thorsten, Wolfram and Chirstoph,
my name is Maho, (lazy) QA project lead.
For better 2.3 release, I'd like to ask you which part is new
and important between 2.2 and 2.3. Well, checking cwses are of course
enough.
I find your names on
http://wiki.services.openoffice.org/wiki/Team_Leads,
and I'd like to ask you about:
Chris: you are responsible for Base. In 2.3, which part should be
extensively tested on TCM?
Helge, you are responsible for automation. Is there any progress in
automate VCLtesttool? or should I ask jogi or I know quite well?)
Oliver, you are responsible for Calc and chart. I know brilliant
new chart2 module has been introduced. Could you please suggest us
which parts are extensively tested by hand on TCM?
Stefan, You are responsible for writer and math,
which part should be extensively tested by hand on TCM?
Thorsten, you are responsible for installation and framework,
What progress has been made for installation? or which part
should be tested as framework?
Wolfram Gaten, you are responsible for Impress and Drawing.
Which part should be extensively tested by hand on TCM?
BTW: I love your
http://qa.openoffice.org/ooQAReloaded/Docs/QA-Reloaded-HowToStart.html
this one and sometimes I point this URI to the new contributors as
it is a good start:)
Christoph, you are respobsilbe for API. What are the new API
and how it is tested?
Any suggestion is really appreciated.
-- Nakata Maho ([EMAIL PROTECTED])
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
