POSIX compliance issue with mmap(2)

[Garrett Cooper]

On Thu, Jul 29, 2010 at 10:01 AM, Garrett Wollman <wollman at csail.mit.edu> wrote:
> <<On Thu, 29 Jul 2010 09:38:59 -0700, Garrett Cooper <yanegomi at gmail.com> said:
>
>> There are a number of opengroup manpages I've seen use the `shall
>> fail' tort in the ERRORs sections -- some being connect(2), open(2),
>> etc. I'll see if I can get clarification on whether or not there is
>> any wiggle room if it states "shall fail if".
>
> "Shall" is a mandatory requirement; if it were optional, it would say
> "may" instead.  (A conformance test has to include at least one test
> for every instance of the word "shall" in the standard.)

    According to the Austin Group folks, shall is a very clear term [1]

For an implementation that conforms to POSIX.1-2008, describes a
feature or behavior that is mandatory. An application can rely on the
existence of the feature or behavior.

For an application or user, describes a behavior that is mandatory.

    should [1] is a different item entirely.
    So mmap(2) needs to be fixed. Shall I create a patch for this?
Thanks,
-Garrett

[1] http://www.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap01.html