Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!dali.cs.montana.edu!uakari.primate.wisc.edu!zaphod.mps.ohio-state.edu!think.com!snorkelwacker.mit.edu!bloom-beacon!eru!hagbard!sunic!mcsun!ukc!warwick!csuwr
From: csuwr@warwick.ac.uk (Derek Hunter)
Newsgroups: comp.sys.acorn
Subject: Re: 32bit immediate load in ARM code
Summary: Plug for my forthcoming C compiler
Keywords: snafu
Message-ID: <7~R_RA`@warwick.ac.uk>
Date: 18 May 91 00:42:40 GMT
References: <+|Q_L||@warwick.ac.uk> and others
Sender: csuwr@cu.warwick.ac.uk (Derek Hunter)
Organization: Computing Services, Warwick University, UK
Lines: 44
Nntp-Posting-Host: lily

I seem to be posting quite a lot at the moment. This is probably my last
 thing on this topic:

It was nice to have had an original idea, shame it wasn't an optimal
 solution and wo'n't work with proposed machines.

As far as knowing which Ldr to produce, all my globals come before any
 program code, so they're all -ve offset.

 Is there as yet any notes on a correction to the aof deficiency?
 Is there any information further to the PRMs in this area I could
  grab hold of?

For the moment, I'm going to leave this system in my compiler because it
 works with the current machines, and because the BASIC C compiler is going
 to be retired once I've rewritten it in C, (without all the design features
 (read: `design bugs') that I `forget' to mention in my Final Year Project
 Report), and I can snaffle John's suggestions into this latter.

------

If anyone's interested in the final compiler, it should code to m/c or
 Basic assembly as RQ'd (I haven't got the ARM assembler) with a linker
 which will work out what needs recompiling a la `make', using the #includes
 (and #pragmas for the extra info it'll probably need when someone out
 there points out why .h and .o timestamps are not sufficient).

 It will also have a nicer interface to RMA moduling than Norcroft's (unless
 I've been mislead about how nasty that is), and will make those interrupt
 surrounds for the routines that you're all itching to write much easier.
 I also want to WIMP it.
 (It currently has a prettier swi interface too, but that might have to
  change when I do register colouring. Either that, or a file with the
  in/output register information.)

 I doubt it will ever be as efficient/fast as Norcroft's thing, but that can
 only improve, and it /will/ only be about 1/3..1/4 as much (Shareware).

		- I hope sh/w plugs don't transgress NETiquette.

		- Derek Hunter		csuwr@cu.warwick.ac.uk


 I thought I might and I tried, but I stopped when I found out I couldn't.