Re: Mac OS X 10.2.8 HEAD test failures.

[Ralf Wildenhues]

Hi Peter,

* Peter O'Gorman wrote on Mon, Dec 10, 2007 at 08:42:25PM CET:
> So, I dug my old g3 tower out of the closet, started it up and ran the
> libtool testsuite (I wanted to see what failed currently before trying a
> patch to see what failed after it).
> 
> There are a number of failures related to the older autotools in the
> default install of 10.2.8. (autoconf-2.52, automake-1.6.1)
> 
> Test 28 fails because autoconf-2.52 does not grok --force.
> "autoconf: invalid option --force"

Please name the tests; I don't think my test 28 (DESTDIR with in-package
deplibs) is your test 28.  Is it `Support for older m4 interface.'
Then please try dropping --force, similar for your test 29 which should
then fail, too.

> Tests 46, 47 and 48 fail because autoconf-2.52 does not have
> AC_CONFIG_MACRO_DIR:
> "./configure: line 894: syntax error near unexpected token
> `LT_CONFIG_LTDL_DIR(ltdl,'"
> 
> Tests 05,29,30,31,39,40,41,43,44,45,46,47,48,52 and 53 fail because
> "aclocal: macro `_LT_CHECK_BUILDDIR' required but not defined"
> Looks like _LT_CHECK_BUILDDIR is m4_defun'ed but AC_REQUIRED.

That's a bug, thanks for catching.  Does it work if _LT_CHECK_BUILDDIR
is only m4_require'd?

I assume if that's fixed, there will still be more issues.

> Should we skip 28,46,47 and 48 if autoconf is too  old, and m4_require
> _LT_CHECK_BUILDDIR?

The "older m4 interface" tests should not be skipped.

I would not be surprised if the nonrecursive libltdl tests required a
newer Automake, and am not sure about the recursive libltdl ones, but
again, I'm not even sure which ones you see failing.

Cheers,
Ralf