Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!linus!philabs!cmcl2!harvard!husc6!panda!genrad!decvax!cwruecmp!bammi
From: bammi@cwruecmp.UUCP (Jwahar R. Bammi)
Newsgroups: net.micro.atari16
Subject: Re: Mini-review of Megamax C
Message-ID: <1497@cwruecmp.UUCP>
Date: Tue, 10-Jun-86 12:39:31 EDT
Article-I.D.: cwruecmp.1497
Posted: Tue Jun 10 12:39:31 1986
Date-Received: Sat, 14-Jun-86 03:56:28 EDT
References: <564@ssc-bee.UUCP> <526@druhi.UUCP>
Distribution: net
Organization: CWRU Dept. Computer Eng., Cleveland, OH
Lines: 37

I have been reading the Megamax Mini reviews with interest. I agree
100% with the comments about its speed etc. I was just a little
surprised that no one mentioned some of its flaws and better known
bugs.

Flaws:
- I am shocked that no one mentioned the fixed sized 32K text/data/bss
segment code model. Ok you can get around the text segement by using
the overlay (non standard C), but how about data (you cannot malloc
and statically initialize it, so you end up writing a malloc followed
by an init routine). Doing a malloc for each data structure larger can
32k in the bss segement can get painful too, especially when porting
large programs. Anyone who has written a program more than a couple
of line long knows what a headache fixed sized segments can give you.

-Shorts are 1 byte long. One of the entusiasts here just got burned on
that one.

-No forward pointer reference of typedef'ed structures (because of its
one pass nature, and short sightedness of the parser implementor).
This broke a couple of my programs.

Well known errors that come to mind:
- Comparison of floats.
- Sizeof struct arrays bombs compiler.
- I have not been able to try the release version to verify if the
preprocessor errors have been fixed.

	I continue to stick to Alcyon with a hard/ram disk. It is very
livable. 

-- 
					Jwahar R. Bammi
			       Usenet:  .....!decvax!cwruecmp!bammi
			        CSnet:  bammi@case
				 Arpa:  bammi%case@csnet-relay
			   CompuServe:  71515,155