Mail Archives: djgpp/1997/10/29/08:37:47
hi!
Eli Zaretskii wrote:
>On Mon, 27 Oct 1997, Vik Heyndrickx wrote:
>
>> You are right. Relying on the default setting of -fsigned-char or
>> -funsigned-char is dangerous. But '\xFD' should be positive when
>> compiled with -funsigned-char. I tested this and it also *is*. So either
>> Roman Suzi must use an older buggy compiler or he is using a defective
>> pgcc or I am missing something.
I got latest version of DJGPP only 1 month ago! Or are there bugs
in RHIDE 1.1?
Because when I manually add -funsigned-char,
it works. (In contrast to when I use -funsigned-char
checkbox in RHIDE in Options/Compilers/C-flags)
>I think you are missing the fact that '\xFD' is NOT a character constant,
>it's an int. Therefore, -funsigned-char doesn't affect it.
NO!!! It has sizeof('\xFD') == 1
I have just checked it.
That is, it is char all right!
>You are right. Relying on the default setting of -fsigned-char or
>-funsigned-char is dangerous. But '\xFD' should be positive when
>compiled with -funsigned-char. I tested this and it also *is*. So either
>Roman Suzi must use an older buggy compiler or he is using a defective
>pgcc or I am missing something.
Possibly, the fact, that my file is probe.cc
and not probe.c ?
That is, its C++ and -funsigned-char is C-flag.
(It is only silly suggestion. I am only beginner with
DJGPP/Rhide to understand all things in details)
>If you want avoid trouble always use explicit casts in these cases.
-- Then the program will be a nightmare...
OK. I can use:
#define UC (unsigned char)
..
c = UC'\xFD'
No problem here.
Sincerely yours, Roman
--
-- * --
-- Roman A. Suzi * Petrozavodsk Karelia Russia --
-- http://www.angelfire.com/ks/sudensivu --
- Raw text -