From: pavenis AT lanet DOT lv
Message-ID: <B0000098360@stargate.astr.lu.lv>
To: "Mark E." <snowball3 AT bigfoot DOT com>, djgpp-workers AT delorie DOT com
Date: Tue, 17 Aug 1999 19:35:25 +0300
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
Subject: Re: crtbegin.o and crtend.o
In-reply-to: <199908171505.PAA97176@out4.ibm.net>
X-mailer: Pegasus Mail for Win32 (v3.12)
Reply-To: djgpp-workers AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

I built GCC-2.95.1 release using old approach. Perhaps we should 
leave new one for DJGPP port of gcc-3.0.

Andris

On 17 Aug 99, at 11:05, Mark E. wrote:

> Greetings,
> 
> Several folks have said recently and in the past that using gcc's crtbegin.o and 
> crtend.o would be "a good thing". I agree, and for the past several days I've been 
> working on getting crtbegin.o and crtend.o to work with DJGPP and today I succeeded. 
> Making crt*.o work with DJGPP will require some changes in libc. However, these 
> changes allow the current method of initializeing EH and of calling 'constructors' 
> and 'destructors' to continue working for as long as neccessary. I'll send the GCC 
> folks the patch to fix up crt*.o for DJGPP as soon as have time to generate the 
> patch and write the ChangeLog. 
> 
> I'm not up on my x86 assembly, so my crt0.S may not be as efficient as it could be. 
> BTW, the current crt0.S used spaces instead of tabs where it was patched initialize 
> C++ EH.
>