[net.micro.pc] Let's C Compiler

jld@ulysses.UUCP (Jeff David) (09/17/86)

I apologize if this has been discussed before, but I only started playing
with IBM clones 2 weeks ago.

I was in a bookstore at lunch today and saw a C compiler called "Let's C"
by Mark Williams Company.  It caught my eye because of the price - it was
only $67.50 (retails for $75)!  It had quotes from two recent reviews in
Byte and PC Tech Journal (favorable, of course).  My question:  Can it do
everything it claims to do - i.e. "Full Kernigan & Ritchie C and extensions;
Full UNIX compatibility and complete libraries"?  I mean, it seems like a
steal next to the MEGABUCKS Microsoft C compiler.

Has anybody used Let's C?  Anybody like it?  Thanks.

Jeff David

jnl@inuxh.UUCP (John Le) (09/18/86)

> I apologize if this has been discussed before, but I only started playing
> with IBM clones 2 weeks ago.
> 
> I was in a bookstore at lunch today and saw a C compiler called "Let's C"
> by Mark Williams Company.  It caught my eye because of the price - it was
> only $67.50 (retails for $75)!  It had quotes from two recent reviews in
> Byte and PC Tech Journal (favorable, of course).  My question:  Can it do
> everything it claims to do - i.e. "Full Kernigan & Ritchie C and extensions;
> Full UNIX compatibility and complete libraries"?  I mean, it seems like a
> steal next to the MEGABUCKS Microsoft C compiler.
> 
> Has anybody used Let's C?  Anybody like it?  Thanks.
> 
> Jeff David

No, I haven't used "Let's C" compiler yet.  But I think it cannot do
bit manipulation and it can only create small memory model (<=64K size).
I bought MIX-C compiler earlier this year for $39 (it still costs the same),
which includes the compiler on the disk and the manual book.  I have used it
extensively and love it.  At $39 it's a steal.  It's an one-pass compiler
which compiles and links code quickly (faster than my Microsoft C compiler).
I would recommend this MIX-C compiler if you're looking for a low cost, full-
feature C compiler.

	John N. Le (AT&T Consumer Products Laboratories, Indpls, IN)
	..!inuxh!jnl

mjb224@uiucuxf.CSO.UIUC.EDU (09/19/86)

I also have the MIX-C Compiler and have found it very complete, very usefull
and VERY VERY SLOW!!!  A recent article in PC-Magazine or PC Tech Journal 
did a comparison of C compilers and found that the generated code was up to
ten times slower than the Microsoft product.  If you can deal with that aspect
(as I was able to) MIX-C is the way to go.

Mike Bruno
Arpanet: mjb224%uiucuxf@a.cs.uiuc.edu
Csnet:   mjb224%uiucuxf@uiuc.csnet
Bitnet:  mjb224@uiucuxf
UUCP:    {ihnp4,pur-ee}!uiucdcs!uiucuxc!uiucuxa!uiucuxf!mjb224
********* Only through December of 1986 **********************

dtg@houem.UUCP (F.GRIZMALA) (09/19/86)

	I have used most of the C Compilers EXCEPT Let's C. Of all the
compilers I have tried, the two I use most often are The Mark Williams C
Compiler for my serious work and MIX-C for my "fun" work.

	MIX-C has its good points and its bad points. The biggest plus 
it has going for it is the "extra" routines it gives you - ie. the
screen functions: clear screen, position cursor, etc. - I wrote these
routines using the Mark Williams Compiler but it was not very much fun to
do. 

	The biggest drawback to the MIX-C Compiler is that it does not
create standard object code and it creates a .COM file rather then a .EXE
file. This limits your compiled code to 64K. They do give you a way to "link"
these .COM files but it is not very elegant. With me, this limitation
provided me with a "fun" challange: to create a full blown data base (I/O
screens, add, modify, delete, search on any field(s), etc) in the 64K
permitted. While it wasn't easy, it WAS do-able.

	One little problem I encountered with the MIX Compiler (and there
is still much debate between MIX, myself and other users of the Compiler
if this is a "problem") is when you have a newline (\n) is a file you are
trying to read. If you create a file and do a dir on it, look at the size
and then try to read that many characters, you end up with garbage at the
end of the file. Reason: the \n counts as two characters to the DOS dir
command but only 1 character to the read command. No problem? 

	Just subtract 1 for each newline in the file, adjust your size
accordingly (if you have 3 newlines, instead of char inp[33] you use
char inp[30] right?). NOW: when you do a lseek you can't use sizeof(inp)
because the lseek counts the \n as two characters again.

	Once you get accustomed to this little problem (they do have a way
around this also but again...not very elegant) the MIX-C Compiler is a
SUPER Compiler and for the money...everyone should have one.

jsm@vax1.ccs.cornell.edu (Jon Meltzer) (09/20/86)

In article <476@inuxh.UUCP> jnl@inuxh.UUCP (John Le) writes:
>> Has anybody used Let's C?  Anybody like it?  Thanks.
>> 
>> Jeff David
>
>No, I haven't used "Let's C" compiler yet.  But I think it cannot do
>bit manipulation and it can only create small memory model (<=64K size).

I have it, and John is right. It's limited to the small model. The thing
also produces object files in its own , non Microsoft, format, and thus can
link only to its own assembler.

Dr Dobbs says that the best cheap C is the Datalight Developer's version -
$99, Lattice compatible, large memory models.

droye@tdms2.UUCP (Doug Royer) (09/23/86)

>I was in a bookstore at lunch today and saw a C compiler called "Let's C"
>by Mark Williams Company.  It caught my eye because of the price - it was
>only $67.50 (retails for $75)!  ........

It is available from 'egghead software' for $54.

>My question:
> Can it do everything it claims to do - i.e. "Full Kernigan> & Ritchie C
> and extensions; Full UNIX compatibility and complete libraries"? ......
> I mean, it seems like a steal next to the MEGABUCKS Microsoft C compiler.

Full K&R: Yes + enum and void.

Unix Libraries: YES, lots of them. It has something called ececall which is 
like execev + system, and the stdio library, + more.

>Has anybody used Let's C?  Anybody like it?  Thanks.

I have used the Lets C compiler. A friend of mine bought it at a bookstore
for $54. This version only supports < 64k memory models. This included
examples, egrep, MicroEmacs (with source), assembler, linker (with user made
library support).

I understand that a development package for around $350 includes the ability
to expand past the 64k limit. It also includes a symbolic debugger, and more
unix like programs. The same manual is included with both versions, the shaded
pages only apply to the development system.

--------------------------------------------------
Doug Royer, Hughes Aircraft GSG, Fullerton, CA

...!scgvaxd!droye	usenet
droyer@ecla		arpanet
--------------------------------------------------

guest@csustan.UUCP (Chris Rhodes) (09/23/86)

(eat me.)

People,
	This discussion of the various C compilers has been helpful,
but I have a question:  What are the system requirements for all
of these various compilers in terms of disk space?  The reason why
I ask is because I have only one floppy (!) and a 300k RAMdrive.
So, those of you who have worked with a particular compiler, please
tell me:  can it work with one floppy?  how about two?  if it's
a two-floppy design, can one of the disks be xfered to RAMdisk and
still run OK (i.e. would it let you have your compiler in A: and
your other stuff in C: or does it need them in a: and b:)?

	Just for the record, the reason why I only have one floppy
is because I just finished putting this thing together myself.
Total cost w/8mhz board, case, multifunction card, color card,
640k RAM, keyboard and one floppy was about $550.


/*  Chris Rhodes / Shooting Shark
 *  Currently cowering behind lll-crg!csustan!guest
 *  real uucp : lll-crg!csuh!shark -or- lll-crg!ptsfa!harlie!shark
 *  My opinions *are* those of csustan!  Yeah, they made me spokesman, ah,
 */ president of the university!  Yeah, *that's* the ticket!