Message-ID: <19980725005503.D222@cdata.tvnet.hu>
Date: Sat, 25 Jul 1998 00:55:03 +0200
From: Laszlo Molnar <molnarl AT cdata DOT tvnet DOT hu>
To: DJGPP Workers List <djgpp-workers AT delorie DOT com>
Subject: stub.exe problem
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: Molnar Laszlo <ml AT cdata DOT tvnet DOT hu>
Precedence: bulk

Hi Workers!

I've a problem with stub.exe. If you look at the binary image
of any djgpp executable you'll see that there is a string

        \x0d\x0astub.h ...
        
beginning at the file position 0x18. However these bytes (0x18-0x1b)
are part of the exe-header according to the interrupt list. And
the word at position 0x18 should show us the beginning of the 
relocation records. In our case this word(0x0a0d) is greater than
0x40, which means that the file has a new style executable header.
Unfortunately this header is 0x40 bytes long, and should contain
some information which the djgpp stub doesn't (for example it should
contain the offset of the "real" executable at positions 0x3c-0x3f).

So, I think that these exe headers are a little bit broken. Here
is a patch that corrects this problem.

bye, Laszlo

--- src/stub/djasm.y~	Sun Jul 12 16:10:54 1998
+++ src/stub/djasm.y	Thu Jul 23 12:44:34 1998
@@ -1361,10 +1361,10 @@
   exe[23] = 0;
 
   time(&now);
-  sprintf(exe+24, "\r\n%s generated from %s by djasm, on %.24s\r\n", argv[2], argv[1], ctime(&now));
+  sprintf(exe+28, "\r\n%s generated from %s by djasm, on %.24s\r\n", argv[2], argv[1], ctime(&now));
   if (copyright)
-    strncat(exe+32, copyright, 480-strlen(exe+32));
-  strcat(exe+32, "\r\n\032");
+    strncat(exe+36, copyright, 480-strlen(exe+36));
+  strcat(exe+36, "\r\n\032");
 
   if (argv[2] == 0)
   {

EndOfPatch