This is the mail archive of the cygwin@sourceware.cygnus.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: cygwin-1999-12-01: Prototypes in system headers not standard-compliant


>>>>> "Chris" == Chris Faylor <cgf@cygnus.com> writes:

Chris> On Fri, Dec 03, 1999 at 08:13:43PM -0800, Martin Buchholz wrote:
>> The prototype for read() and write() in Unix98 don't match those in
>> the Cygwin header files.
>> 
>> http://www.opengroup.org/onlinepubs/007908799/xsh/read.html
>> http://www.opengroup.org/onlinepubs/007908799/xsh/write.html
>> 
>> The standard return type for these functions is `ssize_t'.  
>> The traditional return type for these functions is `int'.
>> The return type in Cygwin is `int'.
>> 
>> On 32-bit platforms, like Cygwin, the best choice is probably to use
>> `ssize_t' in the prototypes (they are standard, after all, right?),
>> but to then have
>> 
>> typedef int ssize_t;
>> 
>> This is effectively what Linux does.

Chris> We're discussing this now.  It's not a simple issue since the "newlib"
Chris> code used by cygwin is also used by other Cygnus products so we have
Chris> to make sure that we don't break anything just to benefit Cygwin's
Chris> portability.

Chris> I do agree with this change, however, and hope to things fixed soon.

I understand that there may be engineering issues associated with the
transition, but the change has to be made sometime soon.  You just
don't have a choice when it comes to standards.  Besides, Cygwin's
primary purpose is to make standard Unix code run on Windows.

Like this (untested) line of code:

ssize_t (*my_read) (int, void *, size_t) = read;

--
Want to unsubscribe from this list?
Send a message to cygwin-unsubscribe@sourceware.cygnus.com


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]