I have a module that includes a few dependencies and one main package.

A colleague submitted a build earlier this week and the dependent packages built fine but the main package FTBFS so the module build failed.

I fixed the failing package and committed it to dist-git and build it.

I went to try my own module spin and it just revived his failed module build which re-ran it using the same build id and package set and it failed (of course, the package n-v-r is still and will always be broken)

I can't cancel it because I didn't submit it.

sgallagh recommended i do an empty commit to the module file and resubmit and that did the trick.

I guess it makes some sense to lock a set of package versions to a module build id for predictability purposes but this could lead to a ton of empty commits to a module just to get another crack at building it. Is that expected?

I also get that I can't cancel someone elses build. I probably don't want others killing mine.

Resubmitting the same build, I guess I get that too. The commit id is the same and there is an existing module for it, it is just a bit surprising. Might be nice to have some feedback from the command on options, or at least a more extended description of what it is doing and why.

Modules are expensive to build. I assume these are protections put into place to prevent the same module being built by multiple people at the same time?