Authentifications with DBUS_COOKIE_SHA1 auth prints a message if server and client sid/uid are different. In recent implementations (I looked into git dbus-1.2 and master branch) the log print contains the server id as client id, which is wrong. An example is displayed below: 272: Client authorized as SID 'S-1-5-18' but our SID is 'S-1-5-18', disconnecting Index: dbus/dbus-transport.c =================================================================== --- dbus/dbus-transport.c (Revision 812) +++ dbus/dbus-transport.c (Arbeitskopie) @@ -651,12 +651,12 @@ if (_dbus_credentials_include(our_identity,DBUS_CREDENTIAL_WINDOWS_SID)) _dbus_verbose ("Client authorized as SID '%s'" " but our SID is '%s', disconnecting\n", - _dbus_credentials_get_windows_sid(our_identity), + _dbus_credentials_get_windows_sid(auth_identity), _dbus_credentials_get_windows_sid(our_identity)); else _dbus_verbose ("Client authorized as UID "DBUS_UID_FORMAT " but our UID is "DBUS_UID_FORMAT", disconnecting\n", - _dbus_credentials_get_unix_uid(our_identity), + _dbus_credentials_get_unix_uid(auth_identity), _dbus_credentials_get_unix_uid(our_identity)); _dbus_transport_disconnect (transport); allow = FALSE; ---- Changelog Entry -------------------------------------------------- 2009-11-20 Ralf Habacker <ralf.habacker@freenet.de> * dbus/dbus-transport.c(auth_via_default_rules): fixed incorrect displayed client sid/uid value
Would you mind attaching a Git commit as a patch?
Created attachment 31364 [details] [review] added patch in git format provided patch for git
Btw, this has already been handled in April in the dbus4win fork (and originally in my patch from last year), and is part of the suggested merge back (cherry-pick extravaganze) of the dbus4win code into dbus master.
Tor, then you can take it out of the patch.
Tor, this patch shows the way to merge the big one patch you refered into upstream. The big patch should be splitted into peaces with a specific context (for example automake build system, cmake build system, windows only fixes and main code fixes).
Of course, that is exactly what I did. I quote my mail: I spent some hours cherry-picking most of the commits from the dbus4win-noncetcp branch of the dbus4win git repository to the master branch of my local clone of the upstream dbus repository. Did you check the http://tml.pp.fi/dbus-dbus4win-merge.diff I pointed to in my mai? It is the output of git format-patch, so there are indeed separate easily digestable pieces. (114 of them...) Do I understand correctly that you now then want the small pieces to be combined together again into a small number of larger patches? Is that really worth it? The amount of patch lines to review won't get any smaller by that.
I created a dedicated bug https://bugs.freedesktop.org/show_bug.cgi?id=25213 for further discussion of the merge.
*** This bug has been marked as a duplicate of bug 25213 ***
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.