From: sandmann AT clio DOT rice DOT edu (Charles Sandmann)
Message-Id: <9706151959.AA15968@clio.rice.edu>
Subject: Re: Latest stub
To: eliz AT is DOT elta DOT co DOT il (Eli Zaretskii)
Date: Sun, 15 Jun 1997 14:59:06 -0600 (CDT)
Cc: robert DOT hoehne AT mathematik DOT tu-chemnitz DOT de, djgpp-workers AT delorie DOT com
In-Reply-To: <Pine.SUN.3.91.970615114528.14561E-100000@is> from "Eli Zaretskii" at Jun 15, 97 11:53:02 am
Content-Type: text
Precedence: bulk

> On the related issue, I would like to suggest a possibility to add code 
> to the stub that will allow it to open the .exe file and read the COFF 
> info even if the first 20 handles are all taken (because the parent 
> program has used them and passed them to the child).  Probably the 
> easiest way would be to make the stub call the DOS function that enlarges 
> the handle table (Int 21h/AH=67h) right from the start.

Right from the start would be too early, since this function allocates DOS
memory for the handle table (around 1020 bytes) which would cause problems
in making the transfer buffer the right size, etc.  After the initial 
resize request, this would be OK if we want to live with each DJGPP program
always using an additional 1K DOS memory.  Some code in libc which handles
the 20 handle transition could be removed.

The other alternative would be to add a feature to let the libc allocate
handles without the inherit bit, or to have it only set the inherit bit
for handles in the first 16 or so.