Fix spurious remaining issues in the master testsuite

[Francesco Casella]

After @AnHeuermann merged in ​PR 6978, one cause for random improvements and regressions in each run of the entire testsuite was fixed for good. Unfortunately, we still have some issues that are causing this behaviour in a few models involving Python and DeviceDrivers (see below). This is no big deal, but it would be nice to get rid of them, and at long last getting clean regression reports.

The UDP models are reported as regressions even though the ​simulation log seems to indicate success. I guess there is something fishy in the way the simulation process is terminated that triggers a non-zero exit code or something. The Python examples just ​segfault every now and then.

Could someone please try to look into these two issues and fix them for good?

[Francesco Casella]

Here is ​another issue with TCP sockets.

[Francesco Casella]

@sjoelund.se reports

The UDP issues are because they all use the same port. We'd need to run each worker in a separate docker instance and it's not that easy to make it all work the way I want. It would also double the time it takes to run the entire thing unless I spent additional time to make attaching to running containers the default :)

Maybe we should just find a way to selectively disable the tests we cannot run. Or at least make them fail consistently...

[Christoph Buchner <buchner@…>]

From looking at the log: Are you still supporting applications using Python 2, which has been unsupported since Jan 2020?

[Francesco Casella]

Replying to Christoph Buchner <buchner@…>:

From looking at the log: Are you still supporting applications using Python 2, which has been unsupported since Jan 2020?

Good question, we should discuss this next year, see #6305.

[Francesco Casella]

Retargeted to 1.18.0 because of 1.17.0 timed release.

[Francesco Casella]

Ticket retargeted after milestone closed