Re: lcc and c99 conformance

En news:87myul9uok.fsf@[EMAIL PROTECTED]
 Ben Bacarisse va escriure:
> jacob navia writes:
>
>> Ben Bacarisse wrote:
>>>
>>> ---------------- 1
>>> int f(int *i) { return *i; }
>>>
>>> void t(void)
>>> {
>>>      long l = 0;
>>>      f(&l);
>>> }


lcc (4.1) chokes on this one:
    6: type error in argument 1 to `f'; found `pointer to long int'
     expected `pointer to int'


In view of the following discussion, I'd append it does it without the -A
option.


>> This provokes a warning if the -A option is present.
>> Otherwise it doesn't.
>
> I thought -ansic was the flag to request standards mode.  What is
> "-A"?

This is standard lcc option to be given more diagnostics.
The manual says:

-A   Warns about declarations and casts of function types without
     prototypes, assignments between pointers to ints and pointers to
     enums, and conversions from pointers to smaller integral types.
     A second -A warns about unrecognized control lines, nonANSI
     language extensions and source characters in literals,
     unreferenced variables and static functions, declaring arrays of
     incomplete types, and exceeding /some/ ANSI environmental limits,
     like more than 257 cases in switches. It also arranges for
     duplicate global definitions in separately compiled files to cause
     loader errors.


> MSVC doesn't warn either by default.

This is one that makes me nervious. Are you considering lcc-win32 should
follow MSVC's behaviour rather that sticks with what was/is in standard
lcc?


Furthermore, I got (-W1 selects minimal warnings)

C> path\to\VS2005\cl.exe -Za -W1 -c 1.c
Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.40904 for
80x86
Copyright (C) Microsoft Cor****ation.  All rights reserved.

1.c
1.c(6) : warning C4133: 'function' : incompatible types - from 'long *' to
'int *'


And this hardly new: I also have
C> path\to\VC++5\cl.exe -Za -W1 -c 1.c
Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 11.00.7022 for
80x86
Copyright (C) Microsoft Corp 1984-1997. All rights reserved.

1.c
1.c(6) : warning C4133: 'function' : incompatible types - from 'long *' to
'int *'


You are correct that the Microsoft compiler does not choke in
"compatibility" mode (-Ze), but then I shall argue this mode was made to
allow the quietest compilations of earlier softwares, back around 1989.

Or have you changed the -A option to really means the same as -Za?
(This is exactly what did Borland, by the way.)


Antoine