|Portability & Compatibility
|Note Also see FlightGear Benchmark and FlightGear Headless
All suggestions and edits welcome.
I want to develop a set of tests for flightgear next using various scenarios and options. This would include testing of build options and various runtime settings in flightgear.
The purpose of testing is to find issues that need fixing, not proving that a certain combination of options is without problems.
Success will be finding and properly documenting an issue for the devs to fix.
A recent good example of what I'd like to find is described the recent post osg multithreading crashes.
To get started I think we need a list of features to test and the options that enable them in builds and in runtime. Remember that these should be areas likely to exhibit bugs:
I'll also need advice on how to gather the right relevant information so the issues can be understood and addressed
So what's on your list?
What info do you need to start testing productively?
- Do we need more basic information about setting up a test? (Yes. Some of us come from other platforms, c++ skills vary etc.)
- What basic skills should we assume? Do we want to encourage testing by someone who is good at testing, can read code, but can't write to the level needed? What support would they need to be effective testers?
- What would you need to learn how to do?
- build a debug version of flightgear
- learn how to start and operate the debugger for your platform.
- learn to read c++ (not write, read!)
- know the basic structure of the flightgear application, major components, interfaces, protocols, event loops.
Would it help to include suggested command lines for starting these scenarios?
|more than one graphics card
|any, Possibly multiple machines/monitors
|Reset & re-init
Ensure the sim can reset/re-init
|built in telnet daemon
|built in httpd daemon
|in air startups
|need list of command line parameters for this, otherwise, reposition to in-air.
|do the log levels give us the right information to pinpoint a problem?