(In reply to bhearsum@mozilla.com (:bhearsum) from comment #18) > Do we even need a real test harness for this? I admit I'm looking at this _very_ simply, but to me, all we need at a high level is: > 1) Download the fully signed browser > 2) Run it, and make sure it stays running for N seconds > 3) End task successfully > > We obviously need some sort of wrapper to do the launching and checking to see if it's running, but unless I'm missing something, a very simple python or even bash script should be enough. Perfect. It does sound like the original bug wanted us to download the *previous* build, apply an update, and verify it starts up. I'm guessing the new signature+notarization changes will apply during the update, so that will essentially test the current build? or we could have 2 flavors, one bare current build, the other previous build + update to current build.
Bug 587344 Comment 21 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
(In reply to bhearsum@mozilla.com (:bhearsum) from comment #18) > Do we even need a real test harness for this? I admit I'm looking at this _very_ simply, but to me, all we need at a high level is: > 1) Download the fully signed browser > 2) Run it, and make sure it stays running for N seconds > 3) End task successfully > > We obviously need some sort of wrapper to do the launching and checking to see if it's running, but unless I'm missing something, a very simple python or even bash script should be enough. Perfect. It does sound like the original bug wanted us to download the *previous* build, apply an update, and verify it starts up. I'm guessing the new signature+notarization changes will apply during the update, so that will essentially test the current build? or we could have 2 flavors, one bare current build, the other previous build + update to current build. [edit]: we do need to make sure that the way we launch the build may matter. (If using the commandline doesn't hit the same quarantine issues as double clicking the app, for example.)
(In reply to bhearsum@mozilla.com (:bhearsum) from comment #18) > Do we even need a real test harness for this? I admit I'm looking at this _very_ simply, but to me, all we need at a high level is: > 1) Download the fully signed browser > 2) Run it, and make sure it stays running for N seconds > 3) End task successfully > > We obviously need some sort of wrapper to do the launching and checking to see if it's running, but unless I'm missing something, a very simple python or even bash script should be enough. Perfect. It does sound like the original bug wanted us to download the *previous* build, apply an update, and verify it starts up. I'm guessing the new signature+notarization changes will apply during the update, so that will essentially test the current build? or we could have 2 flavors, one bare current build, the other previous build + update to current build. [edit]: we do need to double-check if the way we launch the build matters. (If using the commandline doesn't hit the same quarantine issues as double-clicking the app, for example.)