X-Authentication-Warning: delorie.com: mailnull set sender to djgpp-bounces using -f
Message-ID: <3C489F4F.5BB1D77@netstep.net>
From: Rodeo Red <rodeored AT netstep DOT net>
X-Mailer: Mozilla 4.7 [en] (Win98; I)
X-Accept-Language: en
MIME-Version: 1.0
Newsgroups: comp.os.msdos.djgpp
Subject: Re: My program produces nonsense with 3.0.3
References: <3C483E73 DOT 54F9B484 AT netstep DOT net> <a29gvq$1ta$1 AT nets3 DOT rz DOT RWTH-Aachen DOT DE> <3C485969 DOT A862A251 AT netstep DOT net> <a29n0s$6b9$1 AT nets3 DOT rz DOT RWTH-Aachen DOT DE>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Lines: 40
Date: Fri, 18 Jan 2002 22:13:14 GMT
NNTP-Posting-Host: 205.232.135.150
X-Complaints-To: Abuse Role <abuse AT netstep DOT net>, We Care <abuse AT newsread DOT com>
X-Trace: newshog.newsread.com 1011391994 205.232.135.150 (Fri, 18 Jan 2002 17:13:14 EST)
NNTP-Posting-Date: Fri, 18 Jan 2002 17:13:14 EST
Organization: Crossroads Systems (netstep.net)
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com



Hans-Bernhard Broeker wrote:
> 
> Rodeo Red <rodeored AT netstep DOT net> wrote:
> 
> > The compiler took twice as long, the the program it produced was twice
> > as big, and when I ran my program, it took 50% longer and turned the
> > begining of the web pages into gibberish. (A repeating pattern which,
> > strangly had the word java in it.)
> 
> This could all be caused by changes in GCC 3 to get it closer to ANSI
> C++ compliance.  

So does that mean its not really a djgpp problem? (I'm still trying to
understand the difference)

>Hard, if not impossible to tell, without seeing the
> source code and the exact kind of misbehaviour.
> 
Short web pages are not effected. In long pages it turns the first six
hundred lines 
into one line of a repeating pattern: 
tes ÿÿÿÿ	   
@  
@  	   tes ÿÿÿÿ	   
@  
@  	   tes ÿÿÿÿ	   
@
It also changes the file so my editor thinks its a hex file. 

The good version produced by GCC-2.95.2 was 514 kb
This bad version produced by 3.0.3 has 926 kb and took twice as long to
compile

If anyones interested, this page links to the code and a download with
the code and the offending executable:
http://reenie.org/data/badindenter/index.htm

I'm not urgently asking for help or anything like that, I just thought
it might help someone out there track down the bugs in 3.0.3.  
GCC-2.95.2 works fine for my purposes, but I'm glad someones trying to
make it more compliant. 

-red