Re: GNU make 3.81beta4 released

[Eli Zaretskii]

> From: "Markus Mauhart" <address@hidden>
> Date: Mon, 23 Jan 2006 00:38:29 +0100
> 
> Eli, you did not include my addition of the IMHO missing "goto error"
> in function start_job_command(..) once process_easy(..) failes.
> In my case, whithout this patch, make entered an infinite loop.

I didn't have enough time to handle everything.  I just verified that
with my changes, process_easy never fails even when there are 64
processes running.

> I havnt tried your current patches, but you probably have verified
> that with them my experiment's call into process_easy() will either
> no more happen or no more fail.
> 
> Additionally, are you sure that no more other failing process_easy()
> inside start_job_command(..) will ever happen, or that only the failure
> in my tests (as other failures inside start_job_command()) would need
> the "goto error" ?

Your failures were partially induced by your change that forced a
failure inside process_easy when it attempted to launch a 65th
process.  I added changes in job.c to prevent this from happening.  So
the scenario where process_easy fails should be limited to real
failures.

(You will see that I also added to process_easy a test for when it is
called to launch processes when there are already 64 processes
waiting, but in my testing I never saw the error message from that
test, and the list printed by "--debug=j" never showed more than 64
live processes.)

> IOW, if you dont have good reasons against my added "goto error",
> then we should make sure that it goes into the make381 release.

I will look into this when I have more time.