From mboxrd@z Thu Jan  1 00:00:00 1970
To: Fans of the OS Plan 9 from Bell Labs <9fans@cse.psu.edu>
Subject: Re: [9fans] More 'Sam I am' 
In-Reply-To: Your message of "Fri, 10 Feb 2006 08:09:47 CST."
	<8c5fce165415fdcb81e598bfee75740d@quanstro.net> 
References: <8c5fce165415fdcb81e598bfee75740d@quanstro.net> 
Date: Fri, 10 Feb 2006 15:17:46 +0000
From: John Stalker <stalker@maths.tcd.ie>
Message-ID: <200602101517.aa21844@salmon.maths.tcd.ie>
Topicbox-Message-UUID: f989d38e-ead0-11e9-9d60-3106f5b1d025


> i use the change times on files, so writing only the changed files
> is important to me.

Important to you or to mk?  Not that I recommend this as a solution
to the problem of writing only changed files, but I tend to write
mk recipes like

%.dvi:DP false: %.tex %.aux
        touch md5/$target
        md5 $prereq >junk/$target
        while
                ! cmp -s junk/$target md5/$target
        do
                latex $stem.tex </dev/null
                mv junk/$target md5/$target
                md5 $prereq >junk/$target
        done
        rm junk/$target

so that the target is rebuilt if the prerequites have really changed
rather than if their timestamps have.  In the example of LaTeX this
is needed for a variety of reasons stemming from the innate stupidity
of the program, but I've started doing it for pretty much everything.
It's very helpful if, for example, you want to rebuild an old version
of the target.  The md5 and cmp overhead is tolerable for small
projects, but I admit that it scales badly.

ps Sorry for giving the p9p version rather the native.
--
John Stalker
University of Dublin, Trinity College
School of Mathematics