[1294] in Kerberos_V5_Development
Re: Why make check fails under Ultrix....
daemon@ATHENA.MIT.EDU (Mark Eichin)
Fri Jun 7 03:46:13 1996
To: "Theodore Y. Ts'o" <tytso@MIT.EDU>
Cc: Mark Eichin <eichin@MIT.EDU>, Ezra Peisach <epeisach@MIT.EDU>,
krbdev@MIT.EDU, krb5-bugs@MIT.EDU
From: Mark Eichin <eichin@MIT.EDU>
Date: 07 Jun 1996 03:45:30 -0400
In-Reply-To: "Theodore Y. Ts'o"'s message of Thu, 6 Jun 1996 01:19:40 -0400
"Theodore Y. Ts'o" <tytso@MIT.EDU> writes:
> Just for the record, the code in appl/gss-ftp which was checked in by
> Cygnus had getdtablesize has a separate function, and AC_REPLACE_FUNCS
> was used in configure.in. So whoever split it out must have been
> someone at Cygnus.
Are we looking at the same thing? appl/gssftp/ftp/pclose.c has always
had, and still has, an #ifndef HAVE_GETDTABLESIZE, and
appl/gssftp/ftp/configure.in has the corresponding
AC_HAVE_FUNCS(getcwd getdtablesize). In fact, configure.in is
untouched since 3/19, pclose.c is untouched since 2/7. The issue in
question is why this AC_HAVE_FUNCS test might have suddenly started
failing for Tom on Ultrix -- or did that actually happen?
For AC_REPLACE_FUNCS to work, there would have to be a
getdtablesize.c, and not only isn't there one, there isn't one in the
repository, nor is there an Attic version.
As far as I can tell, neither ftp or ftpd, in the cygnus or the MIT
trees, has *ever* had the AC_REPLACE_FUNCS usage you're talking
about. Can you figure out which revision of which file, and quote me
the relevant "cvs log" output (since the ChangeLog entries don't
mention anything relevant...)
