On (09/10/14 20:12), Jakub Hrozek wrote:
On Thu, Oct 09, 2014 at 05:07:27PM +0200, Lukas Slebodnik wrote:
> On (08/10/14 17:46), Lukas Slebodnik wrote:
> >On (06/10/14 15:46), Jakub Hrozek wrote:
> >>Hi,
> >>
> >>attached are more patches that are related to running sssd as a
> >>non-root. I split the functions to become a different user into a
> >>separate module and provided unit tests using nss_wrapper and
> >>uid_wrapper.
> >>
> >>There is a separate Makefile.am for the cwrap-based tests, because it's
> >>not easy (read: possible without a wrapper script) to run tests with a
> >>custom environment. Some build systems, like cmake, allow that, but
> >>autotools does not.
> >
> >>From 565aca30c666cb696ccb5c6a9426144276c7dce9 Mon Sep 17 00:00:00 2001
> >>From: Jakub Hrozek <jhrozek(a)redhat.com>
> >>Date: Sat, 26 Jul 2014 12:46:26 +0200
> >>Subject: [PATCH 1/3] UTIL: Move become_user outside krb5 tree
> >>
> >>In order for several other SSSD processes to run as a non-root user, we
> >>need to move the functions to become another user to a shared space in
> >>our source tree.
> >>---
> >> Makefile.am | 20
++++++++++++--------
> >> src/providers/krb5/krb5_utils.h | 8 --------
> >> .../krb5/krb5_become_user.c => util/become_user.c} | 1 -
> >> src/util/util.h | 9 +++++++++
> >> 4 files changed, 21 insertions(+), 17 deletions(-)
> >> rename src/{providers/krb5/krb5_become_user.c => util/become_user.c}
(99%)
> >>
> >>diff --git a/Makefile.am b/Makefile.am
> >>index
eb0e649437f465d3354a0c063a29c65203824506..ac3f26cad6e1d0a5ac8979ce13286bc896c62b62 100644
> >>--- a/Makefile.am
> >>+++ b/Makefile.am
> >>@@ -1286,13 +1286,14 @@ strtonum_tests_LDADD = \
> >> krb5_utils_tests_SOURCES = \
> >> src/tests/krb5_utils-tests.c \
> >> src/providers/krb5/krb5_utils.c \
> >>- src/providers/krb5/krb5_become_user.c \
> >> src/providers/krb5/krb5_common.c \
> >> src/util/sss_krb5.c \
> >> src/providers/data_provider_fo.c \
> >> src/providers/data_provider_opts.c \
> >> src/providers/data_provider_callbacks.c \
> >>- $(SSSD_FAILOVER_OBJ)
> >>+ src/util/become_user.c \
> >>+ $(SSSD_FAILOVER_OBJ) \
> >>+ $(NULL)
> >Does it make sense to add NULL at the end?
> >You put new file before SSSD_FAILOVER_OBJ.
> >If you think that it make sense I will not insist on change.
> >
I'm not sure what you propose as a better way. I'm not opposed to a
change, but I'm not sure what shall I change.
I added the conditional into specfile. I also squashed in your changes
(thanks!) and new patches are attached.
I thought that idea of NULL is to avoid changing two lines if files is added to
the end. The file "src/util/become_user.c" was added before variable
SSSD_FAILOVER_OBJ, but it isn't important. We can be consistent and add NULL
also after variable SSSD_FAILOVER_OBJ.
Patches works well. make distcheck is fixed. nitpicks were addressed as well.
ACK
LS