Pardus Testing Team

From PardusWiki
Revision as of 21:12, 19 January 2009 by Erkan Yilmaz (Talk | contribs)
Jump to: navigation, search

Contents

Pardus Testing Team

For every product or service that is produced by human beings, it is inevitable to contain bugs or errors. That's why quality assurance policies are needed for every product to decrease the rate of malfunctions.

The aim of the Pardus Testing Team is to minimize the rate of malfunctions before and after the releases. Another aim is to support the fixing of bugs as soon as possible.

Contribution

Free Softwares are not developed and maintained by only the developers but also by the people who volunteer to make translations, help bug-fixing, test, contribute in terms of graphics and package the software. Testing is a good starting point to begin to contribute to free software as it does not require much technical information and besides, helps in learning the softwares that are being tested.

How Can I Be a Member of Pardus Testing Team?

We are expecting you to isolate the testing environment from your daily usage. You can ensure this via virtual machines, for testings inside Releases(packages, interim releases..etc.). However for testing the actual releases, testings before the releases, you definitely need a machine dedicated to this job.

Besides; you are expected to have basic skills of knowledge about Linux and Pardus. Although you are not expected to have a developer level of knowledge, questions like "What is your wireless card's name and its' vendor id?", "Can you add the repo at the address xxxxxx, and then upgrade the system?" should not sound unfamiliar to you.

You must register every computer you are going to use by filling the form at this address and then send it to psts@ @.pngpardus.org.tr. A completed form can be seen here.

How are the things going on for the Testing Team before a Pardus Release?

Before every main release, a form is sent to the Pardus Testing Team Members, and they provide hardware information of the machines which they plan to use as a testing machine. For every system a number is assigned.

With every testing release, Pardus Testing Team Members make a fresh installation and fill in the form they are provided according to the test results in the form. After that, they send these forms to the tester list when the forms are wanted.

If any bugs appear, they are entered to bugzilla (the bug tracking system for Pardus) and can be tracked by the reporter and the Testing Team Responsible till the bug is closed.

Pardus Testing Team Members answer to the feedbacks from the Testing Team Responsible in a certain period of time and keep on doing these things till the stable release is announced.

Reporting Before Release Tests

After the release announcement, the Pardus Testing Team starts testing with the help of the guide they are provided with.
The succesful steps are signed as:

1.1 Yes
1.2 Yes
1.3 Yes

but the steps that are failed should be written in detail as much as possible and be added to the final report. The final report is send to the tester list via e-mail as a "plain-text" . The header of the post should be in the form of "Test No:NN Sistem-XXX" .

How about the Tests during a Release?

Tests during a Release are different from Before-The-Release Tests as this kind of tests start at the beginning of the new stable release and go on till the release is maintained officially. That kind of Tests are seperated into two groups:

Upgrade Tests

For that kind of testing mechanism, you should have a freshly installed version of both the stable release ( e.x. Pardus-2007 ) and interim versions of this release ( e.x. 2007.1 , 2007.2 , 2007.3 ). After every Test process, as we are going to need a fresh version of these installations, installing these systems as a virtual image (e.x. using VirtualBox, see HOWTO:VirtualBox) is useful for your own health :) We are going to use these images in update tests.

Besides, we need another virtual image that is periodically upgraded from the stable repository after every repo updates.

The Process generally works in that way: the Testing Team Responsible starts an ACK process for the packages waiting in the testing repository. The packages that receive ACK from the developers are merged to the current stable repository and to a temporary repository consisting of newly ACK'ed packages. Freshly installed releases are given the address of this temporary repository and being upgraded from this repository.

Every upgraded release is being tested after restarting the system, in order to learn whether the system is capable of doing basic things or not. After this, it is checked with revdep-rebuild command to control if any broken distributed library files in reverse dependencies are missing.

Function Tests

For the Function Tests, the actual version of the latest stable release is upgraded from the testing repository and every upgraded program is being tested one by one.


Surum ici1.jpg


After you make an installation once on a virtual machine (such as VirtualBox - more at: Virtualization:Contents), you can take a screenshot of your machine every time you upgrade it to the latest version of the stable repository, so that you can use this image(screenshot) for your upcoming tests. That prevents installing the same packages every time you use a machine upgraded to the latest version of the repository.

The upgraded release is being tested after restarting the system, in order to learn whether the system is capable of doing basic things or not. After this, it is checked with revdep-rebuild command to control if any broken distributed library files in reverse dependencies are missing.

As testing all the characteristics of the programs and the libraries by the Members of Pardus Testing Team isn't possible in terms of knowledge, experience and time; the packages are classified in 3 categories:

Testing Repository Adresses

Before Reporting Any Bugs

See also

Personal tools
Namespaces
Variants
Actions
Navigation
Print/export
Toolbox
In other languages