Date: Tue, 21 May 1996 14:13:12 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Martynas Kunigelis <martynas DOT kunigelis AT vm DOT ktu DOT lt>
Cc: j DOT aldrich6 AT genie DOT com, DJGPP mailing list <djgpp AT delorie DOT com>
Subject: Re: 'make' problem
In-Reply-To: <199605210914.FAA18536@delorie.com>
Message-Id: <Pine.SUN.3.91.960521140818.14153A-100000@is>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII


On Tue, 21 May 1996, Martynas Kunigelis wrote:

> >%.o : %.cpp
> >
>   Yup, I knew bout that, but I thought this was not standard. Do other makes
>   support this? I thought that was a GNU-Make-only feature. Tell me if it's
>   not.

It is quite standard.  I even think POSIX specifies it, but I don't have 
a POSIX reference handy to check.  Anyway, many Make programs support it, 
although if you want to also support Make programs from DOS compilers, 
you might consider staying away from static pattern rules.

>   BTW, I found a problem with djgpp 2.0 and Borland make. gcc dies with a
>   floating point exceptions when invoked from Borland make. I'm not quite
>   sure now, but I think it WORKED with Beta 3. Has this something to do with
>   the fact that Borland make is a 16 bit DPMI program?

Borland Make (at least the revisions of BC I use) is a 16-bit DPMI 
program, and it can never work with GCC.  16-bit and 32-bit DPMI clients 
cannot be mixed at all.  Beta 3 couldn't have supported this because this 
is a fundamental limitation of the DPMI spec.