Re: BUGS when compiling using SC3.0.1 on Solaris 2.4

From: Erick Gallesio <eg_at_kaolin.unice.fr>
Date: Wed, 01 Mar 1995 22:18:30 +0000

>>>> On Mon, 27 Feb 1995 15:06:14 +0100, Giorgio Cesana writes

> A. No errors occurr when compiling using gcc 2.6.3, the produced code seems
> to run correctly.
>
> B. When I use cc (SC3.0.1) I get some warnings, the code is produced but it
> runs uncorrectly (for example, calling stk -f amib.stklos I get "*** Error:
> map: malformed list"). I recieved everal types of warnings
> 1. warning: initializer does not fit: xxx (on bitmaps and on tkCanvPs)
Yes but the array is explicitely defined as an unsigned char one.
It seems to be a bug of the compiler
> 2. warning: loop not entered at top (on mpn_add.c, mpn_sub.c, mpn_div.c)
That's true. GMP has a very weird flow of control...

> 3. warning: semantics of "<=" change in ANSI C; use explicit cast
> (on gc.c, print.c, eval.c, proc.c)
> 4. warning: trigraph sequence replaced (on slib.c, list.c, read.c, print.c,
> macros.c, dump.c)
I haven't seen this message in your post. But Shouldn't a trigraph contain 3
characters? It seems to be of length 2 here.

> 5. warning: semantics of ">=" change in ANSI C; use explicit cast
> (on port.c)
I don't understand this message
> 6. warning: assignment type mismatch:
> pointer to function() returning void "=" pointer to void
> (on dynload.c)
OK
>
> I think error 2 is not important, error 4 could be avoided by changing the
> string "??" with an other (ex. "xx") in all the comments.
> Errors 3, 5, and 6 could be the ones that causes STk to fail. Is it possible
> trying to fix them?
>

Hard for me to do it since I haven't this compiler. If you have someone send
me patch I could integrate them in the standard release.



                -- Erick
Received on Wed Mar 01 1995 - 22:18:34 CET