raiden00pl commented on issue #17721: URL: https://github.com/apache/nuttx/issues/17721#issuecomment-3703951985
What's more complicated, building and flashing 10 different configurations, or flashing one large configuration and running the applications one by one? I think the first option is much more complicated, time-consuming, and error-prone. You can run the same tests with a single configuration that enables multiple test apps, which saves time on compiling, cleaning, and flashing. We can run apps from NSH, so why not use this feature? This is exactly how the current `citest` works. Serious tests should also check the system state after the application is launched, which can be done with a simple 'ps' or 'free' commands. Just because a test application doesn't return an error doesn't mean it's working properly. Regarding testing on small devices, I mentioned this earlier on mailing list – why use devices with low FLASH for testing at all. In most cases, you'll find chips in the same family but with more FLASH, and that should be the one used for testing. For testing purposes, we can just choose the most powerful chip from a given chip family and it should provide test coverage also for smaller chips, without worrying about available resources. > What solution can you see here @raiden00pl? It would be simpler to have specific configurations list for the CI build rather than making runtime testing a lot more complex? :-) Adding complexity to NuttX to make work easier when implementing a test tool doesn't seem like a good approach to me. The goal should be reducing complexity in NuttX, not in NXDART :) Besides, I don't agree that many separate applications simplify anything - it seems to me that it does quite the opposite. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
