Vito: My two cents: Testtool saves time. To replicate the actions that this program performs would take weeks and is subject to error.
Testtool for 2.1 is broken on the Mac Intel platform. I am troubleshooting it and discovered this occurred between milestones 172 and 173. A look at the code reveals many problems with the code. I am still investigating but suspect that, in the effort to rid the code of warnings during building, the code was too refined and never tested properly on the Mac Intel platform. Serious investigation will take several weeks and many code changes. However, removal of this program from OpenOffice.org would seriously slow the release process or require many more people to become involved in the QA process. Both will slow the progress of development and resolution of issues brought to the attention of the OpenOffice.org project to a slow crawl. James McKenzie -----Original Message----- >From: Vito Smolej <[EMAIL PROTECTED]> >Sent: Dec 28, 2006 8:52 AM >To: [email protected] >Subject: Re: [qa-dev] IRC Meeting - l10n QA - re testtool > >Hi Throsten: > >I can just offer my services regarding i18n and l10n of testtool. >Regarding testtool status ... from a pragmatical point of view, as long >as it helps to get more bugs out of the product, who cares if it is >broken (g). It just does not make nice impression if it is in there >among test cases and everybody makes a wide turn around it. Next time we >may just as well drop the "does a win98 doc file load OK?" test. > >"wide is the road to untested products, and stony and steep is the path >to QA'd stuff" - to paraphrase. > >Re working with jsi - Id say he tried once to show me how to ride the >bicycle (gg). Anyhow, as yet, I am still far away from any serious >work on the subject - at the moment I try to start thinking the cvs way- >I can feel years of working with SourceSafe on my back.... > >Regards to all > >smo > >Thorsten Ziehm wrote: >> Hi Vito, >> >> you pointed out a major problem I see. Everyone want to use the TestTool >> for their sanity checks. But a whole test cycle isn't needed for >> languages and platforms which are provided by Sun. These languages and >> platforms are tested by the Sun QA-team and we announce the results of >> the testing in this list. Actually only a small sanity check is needed >> to approve a release. But this sanity check isn't defined until now. So >> in my opinion the L10N team do too much QA for a release. >> >> The other situation are languages or platforms which are not supported >> by Sun. The language dependencies are not integrated in the test scripts >> and they throw many errors and warning. So the teams do not know, if >> the release can be get an approval or not. For this we need patches or >> fixes by the L10N teams. As I heard you work together with Jogi (JSI) to >> identify all these parts in the scripts. If this could be finished and >> the dependencies for a new language are integrated each other L10N team >> can benefit from these results. >> >> Yes, these are points we have to discuss in the meeting or have to find >> solutions in near future. >> >> Please stay tuned with Jogi, even if he is on vacation now. >> >> Thorsten >> >> >> Vito Smolej wrote: >>> Id like you to address the question of the role testtool has within >>> TCM and the question of an update to the (l10n) Testtool >>> environment. Given the fact that testtool is part of Release sanity >>> test and also a number of other test scenarios, one can not simply >>> think it away. Otoh I have yet to find somebody among 136 testers >>> who has passed (or failed) any entries involving testtool. My >>> opinion: striking it out test cases would mean lowering qa >>> standards. Otoh keeping it in would involve first some more effort, >>> before the tool gets out to the community - right now I have my >>> doubts about it. I am spending more time on adjusting testtool to my >>> local conditions than on testing 2.1. jsi is away on holidays etc... >>> >>> In any case a policy statement would be welcome. >>> >>> TiA >>> >>> smo >>> >> [...] >> >> --------------------------------------------------------------------- >> 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]
