On Mon, Aug 14, 2006 at 07:26:33PM +0200, Kurt Roeckx wrote: > On Mon, Aug 14, 2006 at 11:28:38AM -0500, Zinx Verituse wrote: > > No, it's not the OSs, it's Debian's libtool package. Anything that uses > > file_magic is *still broken* because EGREP still isn't properly passed to > > the libtool script. The OSs I listed are ones that the *NEW* macros still > > use file_magic for. > > Recent version of libtool.m4 have this: > # AC_PROG_EGREP > # ------------- > # This is predefined starting with Autoconf 2.54, so this > # conditional > # definition can be removed once we require Autoconf 2.54 or > # later. > m4_ifndef([AC_PROG_EGREP], [AC_DEFUN([AC_PROG_EGREP], > [AC_CACHE_CHECK([for egrep], [ac_cv_prog_egrep], > [if echo a | (grep -E '(a|b)') >/dev/null 2>&1 > then ac_cv_prog_egrep='grep -E' > else ac_cv_prog_egrep='egrep' > fi]) > EGREP=$ac_cv_prog_egrep > AC_SUBST([EGREP]) > ])]) > > It should create your EGREP. If it doesn't, you're still not > using the latest libtool.m4.
It does create EGREP, in the Makefile, successfully. As I said, this is not passed in any way to the libtool script. I have tested this behavior by making my architecture use file_magic *in the new libtool.m4*, and verified that the new macros were being used, and EGREP was not set *as an environment variable*, or passed in any other way *to the libtool script*. It was set in the Makefile. Please tell me if my testing method is flawed. > > Note that the source has 2 copies of acinclude.m4. You will > probably need to remove everything that comes from an old libtool > version from both of them. > I was aware, and have already supplied this information in the xmms bug. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
