Hi André, Am 24.03.09 10:34, Andre Schnabel schrieb: >> Btw. please do me a favour and contact me or the responsible engineer of >> an autotest[1] directly if you encounter any problems with the tooling >> in the future.
> Sorry - no! > If we have problems with the tooling, all qa-members need to be aware of > this. Or each member will run into the same problems. Furthermore - if we do > not report the problems we will go on in the "all is fine" mood. I did not ask you not to report errors but report them to the responsible engineer first. Problems in tooling (QUASTe, VCLTestTool, Testscripts etc.) should be handled like issues in OpenOffice.org and tracked via Issuetracker at first. Of course severe issues can be reported on mailinglist to avoid other testers running into the same problem and to give a hint of this issue. >> This problem was solved within minutes and I think it's >> not worth to challenge the whole tooling as it works in majority of >> cases ( I would say 99%) . > > And 1% failure might be critical if this hits stopper bugs. In this case the stopper you mentioned was already known (#100235) and fixed in a CWS. The error was found in autotest but not seen in QUASTe. So I would continue in saying tooling worked 100% barring of some small issues. Never told tooling is error free... As I mentioned QUASTe is still in beta-phase and currently it is always a good choice to crosscheck resultfiles created by VCLTestTool. >> This does not help and unsettles the >> community members working with this tooling. Furthermore those problems >> had already been identified (#100235) and meanwhile fixed in CWS sw31fb08. > > This is not the point - or it even proves that we find stoppers much to late > due to several shortcomings in our processes. > In theory all was fine: automation team corerctly prepared the testscripts, > run the tests and VCLTesttool was able to identify the problem. if we had localized builds much earlier and if the results had ben analyzed correctly we could have prevented one of the very late stopper reports. (I'd guess the issue has been introduced months ago). This issue you are talking about was introduced by CWS 'oooimprovement3' and integrated into OOO310m5. It was immediately found by testers on this milestone (OOO310m5) fixed in CWS 'sw31bf08' ...currently waiting for integration. You are talking about months here.....from my point of view it were some hours at most and in this case there was no reason at any time to ask if tooling can be trusted. -- Regards Helge -- =============================================================== Sun Microsystems GmbH Helge Delfs Nagelsweg 55 Quality Assurance Engineer 20097 Hamburg OOo Team Lead Automation http://qa.openoffice.org mailto:[email protected] http://wiki.services.openoffice.org/wiki/User:Hde --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
