If a task fails, and it's manualy modified/corrected and started again, it will always show the state of the first task job.
That is, if the first task finished state is FAILED, although the last one (more recent) was successful by looking at the logs, the state will always remain FAILED.

I've noticed a strange behaviour with this, that I don't know if it's intended or not.

It seems that, if we change the task configuration, when trying to run the task again (without running a new job build), it seems the task will run with the configuration it had when the "main build" run.
Is it supposed to happen like this?

Wouldn't tasks be independent of the job's main configuration, and be independent mini-jobs of their own?

I mean, if we change the task's configuration and run it again, and it will keep running like the first time it was run triggered by the job at the end of the main build, it's ultimate purpose becomes a bit awkard to have, since if the task failes and we have to correct/change it, we would have to run a new job's build again, only to test the task.

Thanks in advance.