Date: Thu, 29 Aug 1996 13:17:53 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Alexander Lehmann <lehmann AT mathematik DOT th-darmstadt DOT de>
Cc: djgpp AT delorie DOT com
Subject: Re: LFN=y is NOT necessary under Win95 by the way
In-Reply-To: <4vvk3h$10ka@rs18.hrz.th-darmstadt.de>
Message-Id: <Pine.SUN.3.91.960829131126.16449A-100000@is>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII


On 27 Aug 1996, Alexander Lehmann wrote:

> Sortof, if you run a single djgpp compiled program, you will get long
> filenames, even though it is set to n in djgpp.env, but if this
> program runs another djgpp program (e.g. gcc running cc1 or make
> running anything), LFN=n will come into effect (this is due to a bug
> in the startup code). So when you are using make, you have to set it
> one way or the other, since otherwise make will find files that gcc
> doesn't find (something like that).

Actually, it's the other way around usually: the spawned programs behave
more cosistently wrt LFN settings than the parent program.  That's because
child programs usually inherit the environment of the parent, so when the
child program starts, LFN is already explicitly in the environment, and
the child program doesn't need to issue LFN calls to find out whether LFN
is supported by default.  Since there are bugs in the startup code 
exactly in the case where LFN API is supported, but LFN=y is NOT set 
explicitly, child programs behave better.

The problems with Make are specific to Make: the current DJGPP port just 
plainly doesn't support long filenames, no matter what you set LFN to.