Re: autoconf 2.5, 2.13, and Vim

[Akim Demaille]

>>>>> "Paul" == Paul Eggert <address@hidden> writes:

Akim> Well, too late, here is a proto monster patch.

Paul> It seems strange for the code to wrap sys/stat.h but not
Paul> fcntl.h, particularly since ancient hosts have the former but
Paul> not the latter.  (These hosts have chown, so the test is
Paul> relevant to them.)  

I agree, this is just one step forward (?), but this patch is already
big enough as is, others are to follow (?).

Paul> One could address this issue by wrapping _every_ include, but
Paul> this will make the code harder to read, and it won't help
Paul> portability on any plausible host.

I'm not sure.  Up to now, I see no reason not to use the default
includes in all these tests, so it there will only remain the most
exotic headers, which should all be protected anyway.

Paul> So I wouldn't bother wrapping headers inside #if except for
Paul> tests where wrapping plausibly adds to portability.

Paul> In other words, I would omit all the changes to acfunctions.m4
Paul> and acspecific.m4, as well as the changes to
Paul> _AC_CHECK_HEADER_DIRENT, AC_HEADER_DIRENT, AC_HEADER_STAT,
Paul> AC_HEADER_SYS_WAIT, AC_HEADER_TIME,
Paul> _AC_HEADER_TIOCGWINSZ_IN_TERMIOS_H,
Paul> _AC_HEADER_TIOCGWINSZ_IN_SYS_IOCTL, AC_TYPE_GETGROUPS, and
Paul> AC_C_BIGENDIAN.  

Including if we just use the $ac_includes_default?  If you say so,
then I'll remove them.  But I confess I'm more than happy when I see
things converging altogether.  So unless there is a problem with using
the default includes everywhere, I'd rather go that way.

Paul> Admittedly this is a judgment call.  It's not incorrect to wrap
Paul> every include; it's more of a style/maintenance issue.

My goal, precisely, is to simplify the maintenance, augment the
homogeneity, and have the tests be more faithful to the actual uses
(where headers etc. are combined altogether).

Paul> PS.  Also, it seems to me that _AC_CHECK_TYPE_OLD's reference to
Paul> sys/types.h (in the generated comment) needs to be fixed.

Hm, I'll have a look at it, thanks!