On Wed, Feb 13, 2019 at 12:38:28AM +0100, Phil Sutter wrote: > diff --git a/configure.ac b/configure.ac > index 048d261ac1088..5384d3c80962c 100644 > --- a/configure.ac > +++ b/configure.ac > @@ -73,6 +73,12 @@ AM_CONDITIONAL([HAVE_SYSTEMD], [test "x$enable_systemd" = "xyes"]) > > AC_CHECK_HEADERS([linux/capability.h],, [AC_MSG_ERROR([Cannot find linux/capabibility.h])]) > > +# check for rpc_msg.h existence > +AC_CHECK_HEADER([rpc/rpc_msg.h], [have_rpc_msg_h=yes], [have_rpc_msg_h=no]) > +AS_IF([test "x$have_rpc_msg_h" = "xno"], > + [PKG_CHECK_MODULES([LIBTIRPC], [libtirpc], [], > + [AC_MSG_ERROR([No provider for rpc/rpc_msg.h found])])]) Better skip this is --disable-cthelper is passed to ./configure? BTW, I have no libtirpc-dev / libtirpc here and this compiles. I can find the headers that I need in libc6-dev. Probably different approach there? Asking to make sure if this dependency is superfluous.
