This is the mail archive of the mailing list for the glibc project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Patches to improve Hurd parts: include/sys/socket.h


On Wed, Dec 17, 2008 at 03:42:10PM -0800, Paul Eggert wrote:
> Thomas Schwinge <> writes:
> > Some questions remain.  Note that all this is not Hurd-specific.
> >
> > Are we allowed to include <fcntl.h> from include/sys/socket.h?
> No, as that would infringe on the user's name space.

What Roland already said.  Thanks nevertheless!

> > Where to put the COMPILE_TIME_ASSERT definition?  For sure into some
> > other file, but into which one?
> I suggest using the macro name that glibc already uses in time/mktime.c:
> 'verify'.

Aha!  I was browsing through the glibc sources, but couldn't easily find
a place where such a compile-time check was done already.

> If you'd like to share this, you can add a file called
> <verify.h> and call the macro 'verify'.  Gnulib has a file verify.h that
> does this and is widely tested with many compilers.  For more, please
> see:
> The header and macro needn't be visible to gnulib's users, of course.
> Though I would advocate adding verify.h (or something like it) to glibc,
> as part of a separate change, at some point: it's quite useful to have a
> compile-time verification macro.

I'll be sure to have a look at this, thanks!  Would the maintainers be
fine with adding this file
to a generic place in the glibc sources or shall I #define the macro
internally where I use it?  As stated above, this already used in generic
code, so this could be unified then.


Attachment: signature.asc
Description: Digital signature

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