Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!nic.MR.NET!umn-cs!bungia!orbit!pnet51!steve
From: steve@pnet51.cts.com (Steve Yelvington)
Newsgroups: comp.sys.atari.st
Subject: Re: Sozobon C: I need graphics ability!  How to?
Message-ID: <610@orbit.UUCP>
Date: 5 Feb 89 01:45:10 GMT
Sender: root@orbit.UUCP
Organization: People-Net [pnet51], Minneapolis, MN.
Lines: 38

Fiacha of Glencar@hjuxa.UUCP (Fiacha of Glencar) writes:
>Even easier, use the GEMFAST libraries posted in the binaries topic.
>
>I tested them, to a limited extent, last night and they worked fine for me.
 
I've run into a couple of minor bugs in GEMFAST, but I think it'll be a nice
library once Ian gets them fixed. His versions of vq_extnd(), which apparently
is mistagged, also crashes the programs in which I've tested it. I have not
found GEMFAST to produce smaller code; in fact, it's been consistently larger.
However, I have not #defined out the VDI control arrays that GEMFAST does not
need. (It uses stack space).
 
>
>My objection to the GEMQ libraries is that AESQ needs a magic table in the
>startup module that defines the number of arguments to each of the AES calls.
>Without a copy, I don't see how you are going to get the calls to work, and
>I didn't see the table in the DASTART.S routine posted here a few days ago.
>
 
If you're referring to the _cntrl_cnts: array, it can be found in the
crystal.o module inside dLibs, thus it is not included in the startup code.
However, the crystal() function itself is not in the current dLibs object
library. I believe it is in the sources, though. I can send a copy if you need
it.
>One good thing to take from the GEMQ routines is the rectangle stuff, i.e.
>most of the routines in UTILQ.S
>The next step is to throw together a linea package.-- 
 
Yes! A linea.a library would be great.
 

UUCP: {uunet!rosevax,amdahl!bungia,chinet,killer}!orbit!pnet51!steve
ARPA: crash!orbit!pnet51!steve@nosc.mil
INET: steve@pnet51.cts.com
  -----------
  -or-
  stag!thelake!steve@pwcs.StPaul.GOV
  "A member of STdNET -- the ST Developers' Network"