Add chainable type to `core` that holds multiple Filesystem mounts

[sipsma]

Right now, there are common use cases that call for running a chain of execs that have multiple mounts in static locations. For example, in yarn we want to run a series of yarn commands with the base image plus source code mounted at /src.

Right now, because you can only obtain one mount at a time, you can't really use chaining very well. This could be fixed by adding a resolver to Exec that obtains a type like Filesystems (note the plural), which has the same interface as Filesystem but will run any execs using multiple mounts (specifically, the output of each mount from the previous step).

[shykes]

Maybe a simpler way would be for exec to accept an array of commands to execute, in addition to executing just one? Then all the commands would benefit from the mount config from the exec input

[sipsma]

Maybe a simpler way would be for exec to accept an array of commands to execute, in addition to executing just one? Then all the commands would benefit from the mount config from the exec input

That would work for chaining exec specifically, but it would also be nice if this had the same extendability you get today with extend type Filesystem. Like what if I want to use the yarnScript Filesystem extension but chained to do multiple builds? If the builds need multiple mounts, then I couldn't today.

Not saying that this means we shouldn't just update exec to accept an array of commands, that's quite beneficial and simpler than the idea here, just that there may be additional benefits to having a type like this too.

[sipsma]

This is supported in the new API 🎉