Re: Bug in update_goal_chain() logic in make 3.79.1

[Paul D. Smith]

OK, I'll look at it.

However, your suggestion of moving the increment to start_waiting_job()
isn't sufficient, because we only want to increment it if we actually
ran something, and we don't know for sure until we get into
start_job_command() whether the command is empty or not (variable
expansion to nothing).

I agree that this is quite messy: for example, it seems there are a
number of places where we check for empty commands, and drop out of the
invocation at different times.

It would be nice to clean _all_ that stuff up; for example, right now we
always obtain a jobserver token for every target, even if it turns out
that we don't have to run any commands!  That's because the place where
we wait to get a token is too far removed from where we eventually
decide whether to fork/exec a command.

This whole area needs investigation, but so far it's not been high on my
agenda.

I'll see if there's something to be done about the specific case you're
talking about, though.

Thanks.

-- 
-------------------------------------------------------------------------------
 Paul D. Smith <address@hidden>          Find some GNU make tips at:
 http://www.gnu.org                      http://www.paulandlesley.org/gmake/
 "Please remain calm...I may be mad, but I am a professional." --Mad Scientist