Firefox:1.5.0.x:Major Update 1.5.0.x to 2.0.0.x: Difference between revisions
No edit summary |
No edit summary |
||
| Line 6: | Line 6: | ||
*'''Make sure updates don't break basic functionality for user (ability to bookmark, add feeds, and such)''' | *'''Make sure updates don't break basic functionality for user (ability to bookmark, add feeds, and such)''' | ||
Currently we have a testing environment setup that allows us to test major updates end to end. Just install 1509 and edit the channel-prefs.js file to point to "releasetest" and check manually for updates. It should present you with the option to jump to 2.0.0.1. As a reference we should take a look at the test cases already in Litmus. Follow the link bellow. | Currently we have a testing environment setup that allows us to test major updates end to end. Just install 1509 and edit the channel-prefs.js file to point to "releasetest" and check manually for updates. It should present you with the option to jump to 2.0.0.1. As a reference we should take a look at the test cases already in Litmus. Follow the link bellow, and enter your [http://wiki.mozilla.org/Firefox:1.5.0.x:Major_Update_1.5.0.x_to_2.0.0.x:Test_Results '''test results here'''] | ||
This is an unorder list of tests spun off the Revised Software Update test cases in Litmus, under the [http://litmus.mozilla.org/show_test.cgi?searchType=by_category&product_id=1&testgroup_id=38&subgroup_id=623 Major/Minor Eula UI Functionality] | This is an unorder list of tests spun off the Revised Software Update test cases in Litmus, under the [http://litmus.mozilla.org/show_test.cgi?searchType=by_category&product_id=1&testgroup_id=38&subgroup_id=623 Major/Minor Eula UI Functionality] | ||
Revision as of 01:55, 23 January 2007
This page cover the strategy for testing major updates going from 150x to 200x. Users of the latest 150x should be presented with the option to do a major update. If they choose this option and agree to the licence agreement, they should be updated to a version of 200x.
At a basic level we should:
- Make sure updates are presented
- Make sure the user profile is preserved after the major update
- Make sure updates don't break basic functionality for user (ability to bookmark, add feeds, and such)
Currently we have a testing environment setup that allows us to test major updates end to end. Just install 1509 and edit the channel-prefs.js file to point to "releasetest" and check manually for updates. It should present you with the option to jump to 2.0.0.1. As a reference we should take a look at the test cases already in Litmus. Follow the link bellow, and enter your test results here
This is an unorder list of tests spun off the Revised Software Update test cases in Litmus, under the Major/Minor Eula UI Functionality
check locales including de stuff check all description content is there check eula is there check you can only install if you accept license check eulas and descriptions are localized check major update works on different locales check interface elements work, buttons, links, text selection, scrolling, closing check "fallback" major updates (should there be a failure in the update download for example) check uninstalling and going back to previous build check Details and license content after URLs have been included check only one type of update is presented at a time minor/major check major updates and how it affects when you have multiple profiles check extension compatibility check check theme compatibility check check what tabs are pre-opened (you have updated) and start page. check these start pages are localized as before. check Windows bits are signed after major update post: check minors are presented to the upgraded 2.0.0.1 users post: check major updates are presented only to those using the latest 150x post: check minors are presented for those users who said never post: check 200x users are not presented a major post: check majors a presented when user had said never, then manually checked for updates again (assuming major are once again available) check to see what happens when one type of update has been downloaded and not applied, and another type of update is made available. check to see if it is possible to download and say later and then regret and then install something better, like a major update. check what is left over after a "diff" (currently some crumbs are left for Win and Mac and it has been captured in a bug) check the state of the registry after a major update (and uninstall)