Path: utzoo!attcan!uunet!bnrgate!bnr-fos!bmers58!atreus!keithh
From: keithh@atreus.uucp (Keith Hanlan)
Newsgroups: comp.sys.amiga
Subject: Re: Latest Lattice C
Message-ID: <1164@bmers58.UUCP>
Date: 11 Jan 90 15:47:09 GMT
References: <5536@fy.sei.cmu.edu> <2917@d75.UUCP>
Sender: news@bmers58.UUCP
Reply-To: keithh@atreus.UUCP (Keith Hanlan)
Organization: Bell-Northern Research, Ottawa, Canada
Lines: 53

In article <2917@d75.UUCP> robin@reed.UUCP (Robin D. Wilson/1000000) writes:
>I just got version 5.04.  Perhaps that's what you mean by 5.4?
>
>BTW, Has anybody tried to compile the examples?  When I do the only one I can
>get to work is speechtoy (I also have not read the manual (manual"s") yet, 
>but I figured it would be easy enough to "wing-it" on the examples).  All of
>them compiled with warnings.  I don't know about anybody else, but I would 
>think/hope that the compiler maker could come up with some code that doesn't 
>have any warnings on the compile.  Or maybe it didn't install properly.

You probably are experiencing the same bug I found in 5.04. It generated
incorrect code for global statically initialized arrays. That is:

char *fred = "abcdef";
main() {
	printf( "'%s'\n", fred );
}

prints out '' (the empty string). Using CPR (which I thought was great!)
showed that it was generating code using offsets from the LOCAL
frame pointer (A5) instead of from the static base (A4). (I might have
flipped the registers around - I'm new to the Amiga and its format)

By contrast:

char *fred;
main() {
	fred = "abcdef";
	printf( "'%s'\n", fred );
}

works as does:

main() {
	char *fred = "abcdef";
	printf( "'%s'\n", fred );
}

I have not been too impressed. Don't they run a regression test on all
the software they release? Even their demos didn't work with 5.04! 
Pretty slipshod.

My next biggest beef is the fact that 50% of my lookups in the Master
Index fail. The documentation is not adaquate for the experienced
programmer/novice amiga programmer - doubtless a frequent customer.

Despising the C programming language and null terminated strings
even more-so doesn't help my disposition.

Keith Hanlan
Keith Hanlan
Bell-Northern Research, Ottawa, Canada 613-765-4645
uunet!utgpu!bnr-vpa!bnr-fos!bmers58!atreus!keithh or keithh@bnr.ca