|Deletions are marked like this.||Additions are marked like this.|
|Line 57:||Line 57:|
|As some testing is carried out on Virtiual Machines (VMs), a quick note on them.||Some testing is carried out on Virtiual Machines (VMs), for details on these head over to:|
|Line 59:||Line 59:|
| * Advantage: If it dies, you can restart / re-install it easily without affecting your 'main' machine.
* Dis-advantage: A VM does not fully check out your own computers' hardware setup.
Each cycle we hold tutorial / classroom sessions on these invaluable tools. If you'd like more information click on the link below to take you to that area.
'''[[https://wiki.ubuntu.com/Testing/Activities/Classroom | All about tutorial / classroom sessions.]]'''
|'''[[https://wiki.ubuntu.com/Testing/Virtual_Machines | Virtual Machines.]]'''|
There are some differences between how the ubuntu family carry out their testing cycles. Some only release at LTS, some use alpha testing, some use only beta testing and everyone has to use release candidate. Add to this the sequence that is cadence testing, it can get confusing. You need to check with the team(s) who you are testing for to know which of the various options they are using. The dates of when important things happen?
Testing is split into distinct, but joined areas. The Daily Builds, the QA-testing of Milestone releases and the Milestone releases themselves.
As new features are added at the beginning of the cycle, Quality assurance and the testers check for bugs and test for correct function.
The daily images are built automatically each 24 hours when not 'turned off' for mile stone testing.
The most important goal with these images is simply to find as many bugs as possible and report them with enough detail that they can be fixed. Finding the bugs ahead of the milestone crunch is helpful as it gives us more time to fix them. Of course finding and reporting bugs in the released milestones is also useful for improving the overall quality of the distribution.
As there is less time pressure on testing these images than the pre-release ones, we would encourage more in-depth testing on a range of hardware to flush out the deeper hidden bugs. If possible, please perform installs using a variety of methods and focus on the quality and depth of testing as opposed to the quantity.
The Dailies are automatically generated every 24 hours, the get suspended as the QA cycles kick in.
Scary, expect them to break! At this point in the cycle the devolopers (devs) are pushing the boundaries to the absolute limit to what they would like to be in the next release. Testing is always important during this time, so that the devs have some feed back as to how things are going. The use of Virtual Machines is handy at this point.
With the launch of the alpha's, we are into the area of 'it should not break too often'. The pre Alpha testing is to ensure that they boot. With a bootable Alpha 1 out, the devs then go to try adding things to it and updating things to the latest versions available for things. Things may well get broken. With the test version of the Alpha 2, things are little less concentrated on getting 'new' stuff in, but checking what is fixable.
The purpose of the Alpha milestones is to encourage wider public testing of the image builds. Before we can recommend the images for public testing we must at least ensure that they have all built correctly (are not in themselves corrupt) and are at least known to boot on some machines. Some form of installation should also be tested. We do not require that these images be tested for any possible minor defect or that all possible install methods work, since that is rather the goal of the wider image testing itself. Find the show stopper bugs and get them fixed.
When an important error is found, the current frozen images are rejected and new ones are built after the fix has been applied. It is therefore important to make sure that you are testing the most recent images.
If the bug is specific to one flavour, only those will be re-built and the other flavour builds remain valid. So if we find a serious bug in the Ubuntu version we will rebuild those images, which will take several hours in total. Likewise, if we find a bug that exists across all flavors, but only on i386, we will re-build i386 and all amd64 and PPC images remain valid.
Beta and Final Release Candidates
The first glimpse of what the final release may well look like. With the Betas, we mainly switch from testing on Virtual Machines to 'Real Computers'. No changes to what applications that will be shipped from this point should be altered without a very good reason. The first Beta is for people to widely check on as many computers as possible to check if certain types of computer and their associated internal parts are having problems. The second Beta is the prelude for the final version, things are now concentrated on bug-fixing.
For Beta and Final milestones we have extended testing periods where we try to find as many bugs as we can. With the Beta we want to put our best foot forward across the distribution and want to fix as many bugs as we have the manpower for, large or small. For the Final release on the other hand, we will only fix the remaining release critical bugs. This is because seemingly small changes can have unintended consequences and break other things. Minor changes that are merely a good idea to fix will be deferred to downloadable updates.
The procedure of rejecting known bad images and continuing testing of unaffected ones also applies here as with the Alphas.
And that, is how we test & do things. If you fancy a roller-coaster ride then do get on. You will have the ability when the Release is made, to smile to yourself and to others. And to say "I was a part of that".
Some testing is carried out on Virtiual Machines (VMs), for details on these head over to: