Re: [bug #29074] -include target fails to issue Error in 3.81

[Lee]

This is a simplified version to get the point across. Hope you noticed that. :-)

all:; @exit 0
done:; @echo $(MAKE_VERSION); exit 1
-include done

I hope you have tested with 3.80 (up to 3.81beta1) and 3.81(from 3.81beta3 and up)
and seen the difference.

We adopt heavy parallel make, so -include target is very useful to build this target
before anything else. In 3.80 era, this worked fine meaning issuing the expected error if there's an error, but in 3.81 it is silent, so we thought the build was ok, but not true.
That concerns us.

ChangeLog may have mentioned this difference, but it is a surprise that the behavior has changed drastically, and we've been using 3.81 for over 2 years now.

hslee

--- On Tue, 3/16/10, Edward Welbourne <address@hidden> wrote:

From: Edward Welbourne <address@hidden>
Subject: Re: [bug #29074] -include target fails to issue Error in 3.81
To: address@hidden
Cc: address@hidden, address@hidden, address@hidden, address@hidden
Date: Tuesday, March 16, 2010, 1:58 AM

> So there was a bug before 3.81, since they issued an Error in this case. And
> the bug was fixed in 3.81. Is that what you are saying?

I think this misses the objection.

> include done
> => is looking for a file named done and fails since no file named done.
>
> -include done
> => is building a target called done (if no file is found).

My understanding of the earlier report was that, in the -include case,
the rule that attempts to create done was not being exercised - the
user expects that, if there's a rule to make done, it will be run if
done does not exist; this happened in older versions (and the rule in
question happened to consist only of producing an error message) but,
apparently, no longer happens.

But I may have mis-understood the reporter ...

    Eddy.