Date: Sun, 16 Sep 2001 21:46:33 +0300 From: "Eli Zaretskii" Sender: halo1 AT zahav DOT net DOT il To: "Mark & Candice White" Message-Id: <2593-Sun16Sep2001214633+0300-eliz@is.elta.co.il> X-Mailer: Emacs 20.6 (via feedmail 8.3.emacs20_6 I) and Blat ver 1.8.9 CC: djgpp AT delorie DOT com In-reply-to: (mhewii@home.com) Subject: Re: addressing question References: Reply-To: djgpp AT delorie DOT com Errors-To: nobody AT delorie DOT com X-Mailing-List: djgpp AT delorie DOT com X-Unsubscribes-To: listserv AT delorie DOT com Precedence: bulk > From: "Mark & Candice White" > Newsgroups: comp.os.msdos.djgpp > Date: Sun, 16 Sep 2001 13:40:42 GMT > > The code is not running in the normal DJGPP run-time environment. > That is made with the help of DOS and DPMI. Neither are there. > Just a simple flat mem model with 3 GDT entries: code, data, and video. > > You are quite right about how djgpp is normaly used but dima made > it clear that DJGPP is being used more as a pure port of gcc hosted > in dos. My responce was in regards to the exact use it is being used for, > and as such is correct. If so, your wording was inaccurate and misled me into thinking you are talking about the normal DJGPP runtime: > > > GCC and its dos port DJGPP are made for flat mem models > > > where ds,es,fs,gs,ss all point to the same mem, and cs is all so > > > the same phisical mem. This explicitly talks about DJGPP, not just GCC. It also specifically says that FS and GS are set up to the same segment as DS, SS, and CS, which is not true in the DJGPP setup. If Dima sets up the environment differently, and doesn't use the DJGPP library and startup code, then he is not using DJGPP at run time. > As an aside I do still belive that DJGPP is made for flat mem model. I don't understand how can you believe that, since facts contradict this. For starters, you can't even access the entire address space without setting the segment limit to -1 and using the address wrap. This is not what the flat model is about. > DJGPP uses a flat mem model, it is just its hosted OS that does > not. Not true: the DJGPP library and the startup code themselves make extensive use of segmentation, and so does the debug support. > Do not worry about debating it, you will not change my mind > and I will not likely change yours, so just belive your right. This is not a question of belief. There are facts out there, so the issue can (and should) be unequivocally decided to be one or the other.