[PATCH libXaw] Add definition of getpagesize(), missing on Windows.

[James Larrowe]

Please disregard this patch. I sent a better one.

On Sat, Mar 9, 2019 at 9:33 AM James Larrowe <larrowe.semaj11 at gmail.com>
wrote:

> Sorry about that, accidentally responded to the digest:
>
> So add another check for cygwin or disable getpagesize use on windows?
>
> On Fri, Mar 8, 2019 at 8:19 PM Jon Turney <jon.turney at dronecode.org.uk>
> wrote:
>
>> On 05/03/2019 17:43, Alan Coopersmith wrote:
>> > What's wrong with it is mainly that the set of people who know about X
>> > libraries
>> > on Windows is small, and the set of people who deal with libXaw is
>> > small, and
>> > the intersection of those two sets is tiny.   (It also doesn't help
>> that
>> > we've
>> > sucked for years at handling patches by email for the unmaintained
>> > modules like
>> > libXaw - we're trying to get better with gitlab merge requests now, but
>> > we're
>> > still not great - unmaintained modules are still not well maintained.)
>> >
>> > I was hoping one of our Windows folks would comment on whether or not
>> > that's the
>> > right way to handle this, or if there's some Windows header to include
>> > instead.
>> > But if no one pipes up soon, I'll likely just go ahead and push it and
>> > see if
>> > anyone complains afterwards that it broke something.
>> >
>> >      -Alan Coopersmith-               alan.coopersmith at oracle.com
>> >       Oracle Solaris Engineering - https://blogs.oracle.com/alanc
>> >
>> > On 03/ 4/19 06:54 AM, James Larrowe wrote:
>> >> Ping. Anything wrong with this? It's only defined in the plugin header
>> >> files for gcc.
>>
>> I guess this means that you're using a MinGW-type Windows-targetted gcc
>> and this function is present in libgcc.a (for who knows what historical
>> reason...), which is turning on HAVE_GETPAGESIZE?
>>
>> In that environment, I think there is a unistd.h, which could be
>> included to provide a prototype.
>>
>> Alternatively, since getpagesize() is a POSIX function, and not in the
>> Windows API, it's possibly more correct for this code not to be used at
>> all on Windows?
>>
>> James: Going forward, if you'd like to Cc: me on your patches, I will do
>> my best to give them review.
>>
>> >> On Thu, Feb 28, 2019 at 12:56 PM James R Larrowe
>> >> <larrowe.semaj11 at gmail.com <mailto:larrowe.semaj11 at gmail.com>> wrote:
>> >>
>> >>     Fix build with --enable-selective-werror.
>> >>
>> >>     Signed-off-by: James R Larrowe <larrowe.semaj11 at gmail.com
>> >>     <mailto:larrowe.semaj11 at gmail.com>>
>> >>     ---
>> >>       src/OS.c | 3 +++
>> >>       1 file changed, 3 insertions(+)
>> >>
>> >>     diff --git a/src/OS.c b/src/OS.c
>> >>     index 4c69005..cd0c517 100644
>> >>     --- a/src/OS.c
>> >>     +++ b/src/OS.c
>> >>     @@ -37,6 +37,9 @@ _XawGetPageSize(void)
>> >>       #endif
>> >>
>> >>       #ifdef HAVE_GETPAGESIZE
>> >>     +# if defined(_WIN32)
>>
>> Also note that, unfortunately, since _WIN32 is overloaded as meaning
>> both "the target is Win32" and "the Win32 API is available", this
>> construction can cause difficulties when compiling for the Cygwin target
>> if "windows.h" has got included somewhere.
>>
>> >>     +extern int getpagesize(void);
>> >>     +# endif
>> >>           if (pagesize == -1)
>> >>              pagesize = getpagesize();
>> >>       #endif
>> >>     --     2.11.0
>> >>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.x.org/archives/xorg-devel/attachments/20190309/7d630d7e/attachment.html>