Hi Thorsten, all Thorsten Ziehm wrote: > Hi Sophie, > > it's not so easy to answer. I am not a friend of doing useless work or > time consuming testing which does not identify any problem over a long > time. So I am not a friend of doing all TCM tests for each release > again and again. I said that often. We need sanity checks for the > releases (perhaps the existing ones have to be expended) which tests > some critical points (e.g. does spell checker works, is it still > possible to write/save/load files in the tested language etc.)
This is exactly the sanity test we are doing using few scenarios and checking for compatibility too. For new > languages all TCM tests are very helpful to identify easily language > depended issues. But do you think all these test cases make sense > on a stable language, which is released and tested since years? I think > it take a mass of time, but the result will be that the known issues > still exists and new ones aren't came up. No I agree with you, this is why we are not doing all the tests, of course. > > The test case specifications for general testing of functionality (new > feature or older functionality) are more complex, as the test cases in > TCM. Therefore my team asked for new features in TCM, when Petr ask for > wishes for OpenTCM [1]. A tooling is needed, where you can integrate > L10N test cases, and general functionality test cases. And a tracking > of the results should be possible. In my view it isn't needed to run > all these test cases for each release/each language in the last weeks > of the release cycle. For general functionality tests it should be > enough, when one run the test case - because most feature are not > language depended. Translation is here a different part. Checking the > translation in new features, does not need test cases for functionality. > It is only needed to know, where the changes were done in UI. Yes but in several cases we have seen that issues appears differently depending on your local environement and/or the local you use in OOo. And if we want to attract more testers on developer releases, we need to have a clear process and know what tool to use. If we abandon the TCM, ok, but the tests on the QA site are unusable for us. > > I do not want to change processes in the L10N-teams. But I think, it > is a point which should be discussed in the L10N project and now we > have a QA Lead for this, Jaqueline Rahemipour. There is two different process that NLC are running : one for L10n, another one for functionalities on a lot of different platforms. For both, here are my questions : - what tests will we run for the TCM l10n testing for 2.3 ? - what tests will we run the 2.3 rcx ? For me currently no tests are available for both, because none exists for l10n, and none are translated (localized) for functionalities (sorry if this effort does exist in the reloaded site, I didn't take the time to dig in it). BTW, there is an encoding issue for me in this page : http://tinyurl.com/yrxl36 Kind regards Sophie --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
