Must have something to do with my re-ordering of dlopen checks. I'll bet it is this part that did it:
> > @@ -1265,11 +1267,11 @@ > > [ > > AC_CHECK_FUNCS(NSLinkModule, [ tempdso="dyld" ], [ > tempdso="no" ]) > > if test "$tempdso" = "no"; then > > - AC_CHECK_LIB(dl, dlopen, [ tempdso="dlfcn" > APR_ADDTO(LIBS,-ldl) ], > > - tempdso="no") > > + AC_CHECK_FUNCS(dlopen, [ tempdso="dlfcn" ], [ > tempdso="no" ]) > > fi > > if test "$tempdso" = "no"; then > > - AC_CHECK_FUNCS(dlopen, [ tempdso="dlfcn" ], [ > tempdso="no" ]) > > + AC_CHECK_LIB(dl, dlopen, [ tempdso="dlfcn" > APR_ADDTO(LIBS,-ldl) ], > > + tempdso="no") > > fi > > if test "$tempdso" = "no"; then > > AC_CHECK_LIB(root, load_image, tempdso="yes", tempdso="no") > > What I DON'T get is why simply checking libc first should break anything, unless HP-UX has bad DSO code in their libc. Is that what may be going on? Kean
