Date: Thu, 12 Sep 1996 17:53:10 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Robert Hoehne <robert DOT hoehne AT mathematik DOT tu-chemnitz DOT de>
Cc: djgpp-workers AT delorie DOT com
Subject: Re: Make 3.74 diffs
In-Reply-To: <Pine.HPP.3.91.960912160412.8310E-100000@newton.mathematik.tu-chemnitz.de>
Message-Id: <Pine.SUN.3.91.960912174336.16328A@is>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII


On Thu, 12 Sep 1996, Robert Hoehne wrote:

> test:
> 	echo test
> 
> If the $(SHELL) variable is set to $COMSPEC, then a batch-file with
> contents like
> 
> i:/djgpp.v2/bin/make.exe test
> 
> is creatated and executed by command.com. But command.com does not know
> about the forward slashes. (The above example is not exactly tested by
> me, because I have my DOS-machine at home, but I remember this, when
> I debugged make)
> 
> Any other command with long commandlines will be also broken, because
> command.com truncates them.

Sorry, all those problems are again caused by the wrong diffs that I
mailed.  In the final version, batch files aren't used at all: Make calls
`system' instead (unless $SHELL is set to a unixy shell with a name such
as `sh.exe', in which case it calls the shell passing it the command line
via a response file).  So in the above example, if $SHELL is set to
COMMAND.COM or 4DOS.COM, `system' will find `echo.exe' in the DJGPP bin
directory and not call COMMAND.COM at all (even though `echo' is a
built-in command).  If $SHELL is set to something like c:/bin/sh.exe, 
`system' will call that shell to do everything.

I cannot begin telling you how sorry and ashamed I am that my stupid
mistake in generating the diffs cost you hours of debugging!  If you
aren't mad with me too much, please try the binary from my machine: 

	ftp://is.elta.co.il/pub/make.exe