Would you recommend using the incubating feature as opposed to the released plugin? I would gladly provide feedback about the task concurrency feature. My only concerns are the length of the incubation period and whether or not there could be drastic changes based on community feedback. On the other hand, the plugin I put together suffers from the same issues you mentioned (doing things concurrently could cause issues if you aren’t careful). And I could easily say that the Gradle team has given much more thought that I.
BTW, thanks for your response Mark.
Continuing the discussion from Continuing the discussion from [What is the status and future of parallel task execution?]:
The current support for parallel task execution is more generic than what is implemented in this plugin. Your use case is more unique, in that you want to explicitly define what tasks get executed in parallel. More specifically, you have a number of long living processes that you want to spawn, which is actually different than a typical task (consumes input, produces output, done). We are actually working on a way of supporting that exact use case. Right now it is internal, and being used for the Play deployment support. The plan is to make it more generic, so it can be used for non-Play projects. For now however there isn’t really a good way to run multiple processes that don’t immediately return. Your option is to use the plugin you have now, or do something like use an Exec task in conjunction with a shell script that spawns the process.