From: Martin Stromberg <Martin DOT Stromberg AT lu DOT erisoft DOT se>
Message-Id: <200009141551.RAA28831@lws256.lu.erisoft.se>
Subject: Re: (fwd) startup-code
To: djgpp-workers AT delorie DOT com
Date: Thu, 14 Sep 2000 17:51:44 +0200 (MET DST)
In-Reply-To: <200009141507.LAA04685@envy.delorie.com> from "DJ Delorie" at Sep 14, 2000 11:07:54 AM
X-Mailer: ELM [version 2.5 PL3]
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

> The existence of both end and _end is historical and typical of older
> unix platforms.

Ok. The right fix seems to be removing _end from djgpp.djl. At the
same time I suggest we remove _etext and _edata as well. They have
similar detrimental effects if user code suddenly decides to use
variables with the names etext or edata.

libc compiles and works fine after this patch, so it should be right.


Right,

						MartinS

--- djgpp.urg/lib/djgpp.djl	Fri Jul 30 03:50:42 1999
+++ djgpp/lib/djgpp.djl	Thu Sep 14 15:43:59 2000
@@ -6,8 +6,8 @@ SECTIONS
     *(.text)
     *(.gnu.linkonce.t*)
     *(.gnu.linkonce.r*)
-    etext  =  . ; _etext = .;
-    . = ALIGN(0x200);
+     etext  =  . ;
+     . = ALIGN(0x200);
   }
   .data  ALIGN(0x200) : {
     djgpp_first_ctor = . ;
@@ -23,7 +23,7 @@ SECTIONS
     *(.eh_fram*)
     ___EH_FRAME_END__ = . ;
     LONG(0)
-     edata  =  . ; _edata = .;
+     edata  =  . ;
      . = ALIGN(0x200);
   }
   .bss  SIZEOF(.data) + ADDR(.data) :
@@ -32,7 +32,7 @@ SECTIONS
     . += 24 ;
     *(.bss)
     *(COMMON)
-     end = . ; _end = .;
+     end = . ;
      . = ALIGN(0x200);
   }
 }