Isn't that the old zsh qouting bug? Well, people still refuse to give us an 1.4.3 anysoon, so may be you want to expand your configure scripts with: http://ac-archive.sf.net/Installed_Packages/patch_libtool_on_darwin_zsh_overquoting.html
Christoph Egger wrote: > Usually I don't reply myself, but I have some news related to my problem: > > >>>Hi! >>> >>>I am running Darwin 6.1. libtool 1.4.2, autoconf 2.52 and automake 1.6.1 >> >>are >> >>>shipped with it. >>> >>>The application I write loads dynamic libs at runtime or at least it >> >>should. >> >>>But Darwin says, the dynamic lib are not of the right type of object >> >>file. >> >>>Having a closer look how these libs are build, I noticed that gcc >> >>_always_ >> >>>uses the -dynamiclib option. Though libtool should use the -bundle >> >>option, >> >>>when the -module option is passed to libtool. >>> >>>Having a look into the libtool.m4 file, I found this line: >>> >>>archive_cmds='$nonopt $(test "x$module" = xyes && echo -bundle || echo >>>-dynamiclib) $allow_undefined_flag -o $lib >>> $libobjs $deplibs$linker_flags -install_name $rpath/$soname $verstring' >>> >>>As gcc _always_ uses the -dynamiclib option, the $module variable is not >>>set to "yes", when it should. >> > > $module _is_ set to yes. In the above mentioned line is a quoting bug. > Replacing the above line by this one works: > > archive_cmds='$nonopt $(test .$module = .yes && echo -bundle || echo > -dynamiclib) $allow_undefined_flag -o $lib $ > libobjs $deplibs$linker_flags -install_name $rpath/$soname $verstring' > > Now, I have another problem: > gcc: -install_name only allowed with -dynamiclib > _______________________________________________ Libtool mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/libtool
