Date: Sun, 28 Feb 1999 15:05:27 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: Paul Derbyshire <pderbysh AT usa DOT net>
cc: djgpp AT delorie DOT com, bug-gnu-utils AT gnu DOT org
Subject: Re: Make bug/wart/oddity, Make v 3.76.1
In-Reply-To: <3.0.6.32.19990228024437.0089e310@pop.globalserve.net>
Message-ID: <Pine.SUN.3.91.990228145932.5950p-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp AT delorie DOT com


On Sun, 28 Feb 1999, Paul Derbyshire wrote:

> I made a makefile one of whose targets creates one of its dependencies
> (perfectly normal) and in which another target shares the same dependency.
> When run, it will (sometimes, but not always) emit a message about the
> dependency in question having a modification time in the future, and
> complaining about clock skew.

Short answer: Upgrade to Make 3.77 (v2/mak377b.zip from the usual DJGPP 
places); it should solve this problem.

Longer answer: this is (not surprisingly) a Windows problem.  The DJGPP 
FAQ list the underlying bug in some detail in section 22.17.  The DJGPP 
port of Make 3.77 works around the Windows bug.  If, after upgrading,
you still see these messages, please post the details to the DJGPP news 
group.

> Since the file is written and the clock is not subsequently changed except
> to allow it to follow its normal strict increasing trend, it is logically
> impossible for the file's timestamp to be greater than the clock time on
> the subsequent read

Ha!  Since when is logic part of MS-Windows (or any MS software, for that 
matter)???