Hi, Andy Thanks for the comment.
On 4 February 2013 02:48, Andy Doan <[email protected]> wrote: > On 02/03/2013 04:37 AM, YongQin Liu wrote: > >> Hi, Andy Doan & Paul Sokolovsky & Vishal Bhoj & All >> >> Now lava-test-shell is supported by LAVA, >> and android team is considering adding the support on android-build. >> >> I want to add following 3 interfaces on the android-build, >> >> *1. >> lava-test-shell-url(url=http:/**/cbuild.validation.linaro.org:** >> 80/helpers/scheduler/lava/**testdef/lava-pandaes-usbdrive/** >> gcc-4.8~bzr121031~christophe-**lyon-vect-cost-model<http://cbuild.validation.linaro.org:80/helpers/scheduler/lava/testdef/lava-pandaes-usbdrive/gcc-4.8~bzr121031~christophe-lyon-vect-cost-model>timeout=59400)* >> > > I'd say this is optional. Do you guys really see a usage for this one? > > Yes, in some case this may be an option. > >> *2. >> lava-test-shell-git(repo=git:/**/git.linaro.org/qa/test-**definitions.git<http://git.linaro.org/qa/test-definitions.git> >> > > This seems like the main one to me. I suspect being android, you'll be > keeping all your definitions in git. Git also happens to be the primary > place we've been doing our work: > > > http://git.linaro.org/gitweb?**p=ubuntu/test-definitions.git<http://git.linaro.org/gitweb?p=ubuntu/test-definitions.git> > > That repo might be a good place for you to work off of as well. > Yes, but seems we'd better to store the android and ubuntu test definition separately. Then we can specify one git url and run all the test there. otherwise seems we need to specify lava-test-shell-git(repo, testdef) for each test. > *3. >> lava-test-shell-bzr(repo=lp:~**terceiro/+junk/lava-dev revision=112 >> timeout=-1)* >> > > Again, this is optional. I'm not sure you'll have a use case for it in > Android. > > yes, I guess this will be rarely used in the future on android too. > > By these 3 types interface, we can do most of the things lava-test-shell >> supported. >> Although we can't specify the array form with one interface one time, >> we can do the same things by specifying multiple times the interface of >> the same type. >> > > Some people might *prefer* that :) Joking aside, there has been debate in > the past whether or not we should reboot in between each test. This > "limitation" would ensure reboots happen and might make Zach and Alexander > happier. > I understand what you said here now. For each lava-test-shell action, lava will reboot to master image to install and then reboot to android image to run the test. This may make Zach happier:) > > like *lava-test-shell-url(url=url1)**,**lava-test-shell-url(url=**url2),* >> >> so I think it's enough for us to use on android-build. >> >> About this, how do you think about that? >> Please feel free to give your comments. Thanks in advance! >> > > I think you are on the right track. I'd suggest just focusing on approach > #2 since that's probably the only way you guys are ever going to use this. > > -- Thanks, Yongqin Liu --------------------------------------------------------------- #mailing list [email protected] <[email protected]> http://lists.linaro.org/mailman/listinfo/linaro-android [email protected] <[email protected]> http://lists.linaro.org/pipermail/linaro-validation
_______________________________________________ linaro-validation mailing list [email protected] http://lists.linaro.org/mailman/listinfo/linaro-validation
