platform/upstream/dbus.git
17 years agoImprove description of allowed variant signatures (#10185)
Simon McVittie [Tue, 9 Oct 2007 13:23:28 +0000 (14:23 +0100)]
Improve description of allowed variant signatures (#10185)
Patch based on one from Kristoffer Lundén, amended as per Havoc's comment.

17 years agodoc/dbus-specification.xml: Clarify description of DBUS_COOKIE_SHA1 (#10184).
Simon McVittie [Tue, 9 Oct 2007 13:12:31 +0000 (14:12 +0100)]
doc/dbus-specification.xml: Clarify description of DBUS_COOKIE_SHA1 (#10184).
Patch from Kristoffer Lundén.

17 years agoMerge branch 'master' of ssh://git.freedesktop.org/git/dbus/dbus
Simon McVittie [Tue, 9 Oct 2007 13:06:54 +0000 (14:06 +0100)]
Merge branch 'master' of ssh://git.freedesktop.org/git/dbus/dbus

17 years agodoc/dbus-specification.xml: Specifically forbid empty structs (#7969)
Simon McVittie [Tue, 9 Oct 2007 13:06:35 +0000 (14:06 +0100)]
doc/dbus-specification.xml: Specifically forbid empty structs (#7969)

17 years agofd.o bug #11678 Don't error out if compiler does not support vararg macros
John (J5) Palmieri [Wed, 3 Oct 2007 21:54:09 +0000 (17:54 -0400)]
fd.o bug #11678 Don't error out if compiler does not support vararg macros

* _dbus_verbose is the only function that does this so make it a noop if
  vararg macros are not supported
* https://bugs.freedesktop.org/show_bug.cgi?id=11678

17 years agofd.o bug #12429 Reverse check to setpcap and only init audit if we were root
John (J5) Palmieri [Wed, 3 Oct 2007 21:29:45 +0000 (17:29 -0400)]
fd.o bug #12429 Reverse check to setpcap and only init audit if we were root

* patch by Dan Walsh <dwalsh@redhat.com>
* https://bugs.freedesktop.org/show_bug.cgi?id=12429
* Reverse we_were_root check to setpcap if we were root.  Also only init
  audit if we were root.  So error dbus message will not show up when policy
  reload happens.  dbus -session will no longer try to send audit message,
  only system will.

17 years agofd.o bug #11872 improve linker test for --gc-sections
John (J5) Palmieri [Wed, 3 Oct 2007 21:08:44 +0000 (17:08 -0400)]
fd.o bug #11872 improve linker test for --gc-sections

* patch by Tim Mooney <enchanter at users.sourceforge.net>

17 years agofd.o bug #11872 fix clearenv for systems that do not have it
John (J5) Palmieri [Wed, 3 Oct 2007 21:02:23 +0000 (17:02 -0400)]
fd.o bug #11872 fix clearenv for systems that do not have it

* patch from Brian Cameron <brian.cameron at sun.com>

17 years agofd.o bug #12547 remove superfluous if
John (J5) Palmieri [Wed, 3 Oct 2007 20:43:22 +0000 (16:43 -0400)]
fd.o bug #12547 remove superfluous if

* also convert tabs to spaces

17 years agoCorrectly implement -fPIC and -fPIE
John (J5) Palmieri [Wed, 3 Oct 2007 19:21:57 +0000 (15:21 -0400)]
Correctly implement -fPIC and -fPIE

* For security reasons we want possition independent code for libraries
  and possition independent executable for executables
* before we were just enabling -fPIC
* now we correctly enable -fPIC and -PIE for libdbus and the bus respectively
* propper LD_FLAGS are set for each also

17 years agoAdd argument path matching support. Bug #11066.
Ryan Lortie [Thu, 20 Sep 2007 17:04:38 +0000 (13:04 -0400)]
Add argument path matching support.  Bug #11066.

2007-09-20  Ryan Lortie  <desrt@desrt.ca>

        * dbus/signals.c (struct DBusMatchRule, bus_match_rule_new,
        bus_match_rule_set_arg, bus_match_rule_parse_arg_match,
        match_rule_matches): Add support for parsing and matching on
        arg0path='/some/path' type rules.

        * dbus/signals.h (bus_match_rule_set_arg): change to take const
        DBusString instead of const char * for the string to match against.

        * dbus/dbus-bus.c: add a quick note to dbus_bus_add_match
        documentation about the path matching.

        * doc/dbus-specification.xml: add a more detailed description of the
        changes here.

17 years agoAdd support for compacting DBusStrings to release wasted memory.
Ryan Lortie [Thu, 20 Sep 2007 04:13:35 +0000 (00:13 -0400)]
Add support for compacting DBusStrings to release wasted memory.

2007-09-19  Ryan Lortie  <desrt@desrt.ca>

        * dbus/dbus-string.[ch] (compact, _dbus_string_compact,
        _dbus_string_lock): new compact function to free up allocated memory
        that is no longer used.

        * dbus/dbus-message.c (load_message): call _dbus_string_compact on the
        message loader buffer.

        * dbus/dbus-transport-socket.c (do_reading, do_writing): call
        _dbus_string_compact on the incoming/outgoing "encoded" buffers.

        * dbus/dbus-string-util.c (_dbus_string_test): add a few tests for
        string compacting.

17 years agoHACKING: add more explicit git branch/tag instructions
Ryan Lortie [Thu, 13 Sep 2007 16:42:32 +0000 (12:42 -0400)]
HACKING: add more explicit git branch/tag instructions

17 years agomigrate from cvs to git (cvs2svn -> git-svnimport).
Ryan Lortie [Thu, 13 Sep 2007 15:13:40 +0000 (11:13 -0400)]
migrate from cvs to git (cvs2svn -> git-svnimport).

2007-09-13  Ryan Lortie  <desrt@desrt.ca>

        migrate from cvs to git (cvs2svn -> git-svnimport).

        * HACKING: update release/branch/tag instructions
        * */.cvsignore: rename to .gitignore

        also, clean up tags and branch names to conform to HACKING

17 years ago2007-08-17 William Jon McCann <mccann@jhu.edu>
Havoc Pennington [Fri, 17 Aug 2007 18:51:41 +0000 (18:51 +0000)]
2007-08-17  William Jon McCann  <mccann@jhu.edu>

* update-dbus-docs.sh: upload DTD to server

17 years ago2007-08-17 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Fri, 17 Aug 2007 16:43:57 +0000 (16:43 +0000)]
2007-08-17  Havoc Pennington  <hp@redhat.com>

* tools/dbus-launch-x11.c (set_address_in_x11): fix from Michael
Lorenz to use long not int with XChangeProperty format 32

* dbus/dbus-sysdeps-util-unix.c
(_dbus_write_pid_to_file_and_pipe): factor this out, and use the
same code in _dbus_become_daemon (where the parent writes the pid
file and to the pid pipe) and in bus_context_new (where the daemon
writes its own pid file and to its own pid pipe)

* bus/bus.c (bus_context_new): close the pid pipe after we print
to it. Also, don't write the pid to the pipe twice when we fork,
someone reported this bug a long time ago.

17 years ago2007-08-03 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Fri, 3 Aug 2007 19:05:29 +0000 (19:05 +0000)]
2007-08-03  Havoc Pennington  <hp@redhat.com>

* configure.in: add major/minor/micro version number AC_SUBST

* dbus/dbus-arch-deps.h.in (DBUS_MAJOR_VERSION,
DBUS_MINOR_VERSION, DBUS_MICRO_VERSION, DBUS_VERSION_STRING,
DBUS_VERSION): collection of macros to get version of library we
are compiled against.

* dbus/dbus-misc.c (dbus_get_version): new function, to get
version of library we are linked against at runtime.

17 years ago2007-07-30 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Mon, 30 Jul 2007 16:16:47 +0000 (16:16 +0000)]
2007-07-30  Havoc Pennington  <hp@redhat.com>

* bus/activation-helper.c (check_bus_name): don't use
_dbus_check_valid_bus_name() which is only around with
--enable-checks, instead use _dbus_validate_bus_name().
Bug #11766 from Diego <diego@pemas.net>

17 years ago2007-07-27 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Fri, 27 Jul 2007 21:01:22 +0000 (21:01 +0000)]
2007-07-27  Havoc Pennington  <hp@redhat.com>

* configure.in: post-release version bump

17 years ago2007-07-27 Havoc Pennington <hp@redhat.com> dbus-1.1.2
Havoc Pennington [Fri, 27 Jul 2007 20:59:42 +0000 (20:59 +0000)]
2007-07-27  Havoc Pennington  <hp@redhat.com>

* release 1.1.2

17 years ago2007-07-26 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Thu, 26 Jul 2007 23:01:14 +0000 (23:01 +0000)]
2007-07-26  Havoc Pennington  <hp@redhat.com>

* bus/config-parser-trivial.c (check_return_values): disable a
test that hardcoded the bus user's name

* bus/dispatch.c (bus_dispatch_test_conf): remove the "if
(!use_launcher)" around the tests, they were only failing because
we didn't pass through all the expected errors from the helper.

* bus/activation-exit-codes.h
(BUS_SPAWN_EXIT_CODE_CHILD_SIGNALED): add a code for child segfaulting
(BUS_SPAWN_EXIT_CODE_GENERIC_FAILURE): make "1" be a generic
failure code, so if a third party launch helper were written it
could just always return 1 on failure.

17 years agofix whitespace
Havoc Pennington [Thu, 26 Jul 2007 22:32:09 +0000 (22:32 +0000)]
fix whitespace

17 years agoSwitch over to using getaddrinfo for TCP clients & servers to enable IPv6
Daniel P. Berrange [Wed, 25 Jul 2007 02:46:52 +0000 (02:46 +0000)]
Switch over to using getaddrinfo for TCP clients & servers to enable IPv6

17 years ago2007-07-24 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Tue, 24 Jul 2007 22:11:00 +0000 (22:11 +0000)]
2007-07-24  Havoc Pennington  <hp@redhat.com>

* configure.in: add AM_PROG_CC_C_O to allow per-target CPPFLAGS

* bus/dispatch.c (bus_dispatch_test_conf): Fix up setting
TEST_LAUNCH_HELPER_CONFIG to include the full path, and enable
test shell_fail_service_auto_start when use_launcher==TRUE

* bus/activation-helper-bin.c (convert_error_to_exit_code): pass
through the INVALID_ARGS error so the test suite works

* bus/activation.c (handle_activation_exit_error): return
DBUS_ERROR_NO_MEMORY if we get BUS_SPAWN_EXIT_CODE_NO_MEMORY

* dbus/dbus-spawn.c (_dbus_babysitter_get_child_exit_status):
return only the exit code of the child, not the entire thingy from
waitpid(), and make the return value indicate whether the child
exited normally (with a status code)

* bus/bus.c (process_config_first_time_only): _dbus_strdup works
on NULL so no need to check
(process_config_every_time): move servicehelper init here, so we
reload it on HUP or config file change

* bus/Makefile.am (install-data-hook): remove comment because
Emacs make mode seems to be grumpy about it

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 13:03:31 +0000 (13:03 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/Makefile.am:
* bus/test-system.c: (die), (check_memleaks), (test_pre_hook),
(test_post_hook), (main):
Add back the test-system.c file - not sure now this got ignored in the
diff. I blame git.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 12:55:32 +0000 (12:55 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* configure.in:
Use ustar to generate the tarball; this fixes the make distcheck
problem when the data files do not fit in the archive:
tar: dbus-1.1.2/test/data/valid-service-files/org.freedesktop.DBus.
TestSuiteShellEchoServiceFail.service.in: file name is too
long (max 99); not dumped

We have to have the 'long' names as the service helper matches by
filename rather than by the name in the service file.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 12:48:45 +0000 (12:48 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* configure.in:
* test/Makefile.am:
* test/data/invalid-service-files-system/org.freedesktop.DBus.TestS
uiteNoExec.service.in:
* test/data/invalid-service-files-system/org.freedesktop.DBus.TestS
uiteNoService.service.in:
* test/data/invalid-service-files-system/org.freedesktop.DBus.TestS
uiteNoUser.service.in:
* test/data/valid-config-files-system/debug-allow-all-fail.conf.in:
* test/data/valid-config-files-system/debug-allow-all-pass.conf.in:
* test/data/valid-config-files/debug-allow-all-sha1.conf.in:
* test/data/valid-config-files/debug-allow-all.conf.in:
* test/data/valid-service-files-system/org.freedesktop.DBus.TestSui
teEchoService.service.in:
* test/data/valid-service-files-system/org.freedesktop.DBus.TestSui
teSegfaultService.service.in:
* test/data/valid-service-files-system/org.freedesktop.DBus.TestSui
teShellEchoServiceFail.service.in:
* test/data/valid-service-files-system/org.freedesktop.DBus.TestSui
teShellEchoServiceSuccess.service.in:
* test/data/valid-service-files/debug-echo.service.in:
* test/data/valid-service-files/debug-segfault.service.in:
* test/data/valid-service-files/debug-shell-echo-fail.service.in:
* test/data/valid-service-files/debug-shell-echo-success.service.in:
* test/data/valid-service-files/org.freedesktop.DBus.TestSuiteEchoS
ervice.service.in:
* test/data/valid-service-files/org.freedesktop.DBus.TestSuiteSegfa
ultService.service.in:
* test/data/valid-service-files/org.freedesktop.DBus.TestSuiteShell
EchoServiceFail.service.in:
* test/data/valid-service-files/org.freedesktop.DBus.TestSuiteShell
EchoServiceSuccess.service.in:
Add the data files needed by the system activation unit checks.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 12:39:30 +0000 (12:39 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/dispatch.c: (check_segfault_service_no_auto_start),
(check_launch_service_file_missing),
(check_launch_service_user_missing),
(check_launch_service_exec_missing),
(check_launch_service_service_missing), (bus_dispatch_test_conf),
(bus_dispatch_test_conf_fail), (bus_dispatch_test):
Add unit tests for system activation. Most are copied from the
session activation tests, but some didn't apply when using a laucher.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 12:36:41 +0000 (12:36 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/activation.c: (bus_activation_activate_service):
If the bus uses a service-laucher, then use the setuid laucher.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 12:33:35 +0000 (12:33 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* configure.in:
Add the needed library exports for the new laucher.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 12:31:01 +0000 (12:31 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* configure.in:
Check for -Wl,--gc-sections so we can really reduce the size of the
setuid binary.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 12:22:43 +0000 (12:22 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/activation.c: (handle_activation_exit_error),
(babysitter_watch_callback):
Map the child exit status integer to a proper dbus error.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 12:19:59 +0000 (12:19 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/bus.c: (process_config_first_time_only),
(process_config_every_time), (bus_context_unref),
(bus_context_get_servicehelper):
* bus/bus.h:
Add the concept of a service-helper and allow it's value to be read.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 12:17:39 +0000 (12:17 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/activation.c: (bus_activation_entry_unref),
(update_desktop_file_entry):
Add the concept of, and read the value of user from the desktop file.
The user string is not required unless we are using system activation.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 12:14:51 +0000 (12:14 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/activation.c:
* bus/desktop-file.h:
Move the defines into the header file, as we use these in the lauch
helper as well as the desktop file parsing.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 12:12:10 +0000 (12:12 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/.cvsignore:
Add the autogenerated binary files.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 12:10:58 +0000 (12:10 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/Makefile.am:
* bus/test.h:
Add the build glue for the lauch helper, and also add the launch-helper
OOM checks into make check. I've probably broken the build, give me 2.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 12:03:04 +0000 (12:03 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/test-launch-helper.c: (die), (check_memleaks),
(test_post_hook), (bus_activation_helper_oom_test), (main):
Add a test wrapper to allow OOM checks on the launch helper.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 12:01:32 +0000 (12:01 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/activation-helper-bin.c: (convert_error_to_exit_code),
(main):
* bus/activation-helper.c: (desktop_file_for_name),
(clear_environment), (check_permissions), (check_service_name),
(get_parameters_for_service), (switch_user),
(exec_for_correct_user), (check_bus_name), (get_correct_parser),
(launch_bus_name), (check_dbus_user), (run_launch_helper):
* bus/activation-helper.h:
Add the initial launch-helper. This is split into a main section and a
binary loader that allows us to lauch the main section in another test
harness to do stuff like OOM testing. No build glue yet.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 11:58:08 +0000 (11:58 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/Makefile.am:
* bus/config-parser.c: (bus_config_parser_unref),
(start_busconfig_child), (bus_config_parser_end_element),
(servicehelper_path), (bus_config_parser_content),
(bus_config_parser_finished),
(bus_config_parser_get_servicehelper),
(test_default_session_servicedirs),
(test_default_system_servicedirs), (bus_config_parser_test):
* bus/config-parser.h:
Make the config-parser code use the common config code.
Also add the session and systemdirs stuff, and make the config parser
aware of the servicehelper field.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 11:53:35 +0000 (11:53 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/system.conf.in:
Add new servicehelper fields to the default system.conf file.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 11:50:59 +0000 (11:50 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/config-parser-trivial.c: (service_dirs_find_dir),
(service_dirs_append_link_unique_or_free), (bus_config_parser_new),
(bus_config_parser_unref), (bus_config_parser_start_element),
(bus_config_parser_end_element), (bus_config_parser_content),
(bus_config_parser_finished), (bus_config_parser_get_user),
(bus_config_parser_get_type), (bus_config_parser_get_service_dirs),
(check_return_values), (do_load), (check_loader_oom_func),
(process_test_valid_subdir), (make_full_path), (check_file_valid),
(bus_config_parser_trivial_test):
* bus/config-parser-trivial.h:
Add a security sensitive stripped down config parser for the setuid
launcher. This file only reads what it needs, and doesn't try to do
anything remotely clever like including external files.
It is not intended to validate the config file; it is expected that
config-parser will do that before the setuid program tries to read it.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 11:47:31 +0000 (11:47 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/config-parser-common.c:
(bus_config_parser_element_name_to_type),
(bus_config_parser_element_type_to_name):
* bus/config-parser-common.h:
We don't want to run the whole config parser with all it's deps in the
setuid program. We need to implement a stripped down config parser just
for the launcher, and to do so I need some common functions and
defines; add them here.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 11:44:35 +0000 (11:44 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* dbus/dbus-sysdeps-unix.c:
(_dbus_get_standard_system_servicedirs):
* dbus/dbus-sysdeps-win.c:
Provide a way to get the standard system servicedirs, just like we do
for the session service dirs. These should be seporate, as there may
be a security issue starting up some session stuff as root.
The use-case for the same binary starting up per-system _and_
per-session is also not valid.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 11:41:19 +0000 (11:41 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/dbus-daemon.1.in:
Add standard_system_servicedirs and servicehelper into the man file
and explain what each does.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 11:38:47 +0000 (11:38 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* doc/busconfig.dtd:
Add servicehelper into the dtd, it will soon be a valid part of the
config file.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 11:37:20 +0000 (11:37 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* dbus/dbus-spawn.c: (read_data),
(_dbus_babysitter_get_child_exit_status):
* dbus/dbus-spawn.h:
Add a function so we can get access to the exit status of the launch
helper.
By providing the return code and not the error we can leave the
'what does this mean?' to the bus launch code and not include it in the
dbus directory.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 11:34:40 +0000 (11:34 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* bus/activation-exit-codes.h:
Add defines which specify the output codes of the launch helper.
We have to use exit codes as this is the only way we can return failure
type without going grotty things like redirecting possibly-nonsecure
stderr into the error.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 11:27:09 +0000 (11:27 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* dbus/dbus-protocol.h:
Add new error names needed for the launch helper.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 11:24:52 +0000 (11:24 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* dbus/dbus-sysdeps.c: (_dbus_clearenv):
* dbus/dbus-sysdeps.h:
Add a wrapper for clearenv.

17 years ago2007-07-24 Richard Hughes <richard@hughsie.com>
Richard Hughes [Tue, 24 Jul 2007 11:23:46 +0000 (11:23 +0000)]
2007-07-24  Richard Hughes  <richard@hughsie.com>

* doc/system-activation.txt:
Add design document for the system activation parts. I'll shortly be
committing many patches that add system activation using a setuid
launcher into CVS, so expect things to be broken for a few hours.

17 years ago* cmake/modules/FindKDEWIN.cmake: fixed comment
Ralf Habacker [Thu, 19 Jul 2007 20:16:09 +0000 (20:16 +0000)]
* cmake/modules/FindKDEWIN.cmake: fixed comment
* cmake/modules/FindKDEWIN32.cmake: removed obsolate cmake module

17 years ago2007-07-18 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Wed, 18 Jul 2007 21:43:23 +0000 (21:43 +0000)]
2007-07-18  Havoc Pennington  <hp@redhat.com>

* dbus/dbus-message.c (dbus_message_get_cached)
(dbus_message_cache_or_finalize): don't mess with message from
message cache outside of the cache lock. Bug #9164 from Jonathan
Matthew.

17 years ago2007-07-13 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Sat, 14 Jul 2007 02:44:01 +0000 (02:44 +0000)]
2007-07-13  Havoc Pennington  <hp@redhat.com>

* Add indent-tabs-mode: nil to all file headers.

17 years ago2007-07-12 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Thu, 12 Jul 2007 22:18:05 +0000 (22:18 +0000)]
2007-07-12  Havoc Pennington  <hp@redhat.com>

* dbus/dbus-sysdeps-util.c (_dbus_sysdeps_test): invert the test
for parsing hex as double to be sure it fails to work

* dbus/dbus-sysdeps.c (_dbus_string_parse_double): don't allow hex numbers.

17 years ago2007-07-10 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Tue, 10 Jul 2007 16:00:44 +0000 (16:00 +0000)]
2007-07-10  Havoc Pennington  <hp@redhat.com>

* dbus/dbus-connection.c (struct DBusConnection): Fix from Olivier
Hochreutiner to avoid trying to protect individual bits in a word
with different locks (make dispatch_acquired and io_path_acquired
dbus_bool_t rather than bitfields)

17 years ago* dbus/dbus-sysdeps-win.c,dbus-sysdeps-win.h,dbus-sysdeps-win-util.c,dbus-sysdeps...
Ralf Habacker [Mon, 9 Jul 2007 06:24:50 +0000 (06:24 +0000)]
* dbus/dbus-sysdeps-win.c,dbus-sysdeps-win.h,dbus-sysdeps-win-util.c,dbus-sysdeps-spawn-win.c: synced with windbus sources

17 years ago* dbus/dbus-sysdeps-win.c
Ralf Habacker [Sat, 7 Jul 2007 11:14:44 +0000 (11:14 +0000)]
* dbus/dbus-sysdeps-win.c
(_dbus_getsid): remove unused jump label
(_dbus_read_credentials_socket): _dbus_string_init could fail, check initialisation
(_dbus_get_working_dir, _dbus_init_working_dir): remove unused functions

17 years ago* cmake/modules/FindKDEWIN.cmake: search in <program files dir>\win32libs for windows...
Ralf Habacker [Wed, 4 Jul 2007 06:41:56 +0000 (06:41 +0000)]
* cmake/modules/FindKDEWIN.cmake: search in <program files dir>\win32libs for windows supplementary packages too

17 years ago* cmake/dbus/CMakeLists.txt: handle userdb as unix file
Ralf Habacker [Sat, 30 Jun 2007 11:46:48 +0000 (11:46 +0000)]
* cmake/dbus/CMakeLists.txt: handle userdb as unix file

17 years ago* dbus/dbus-sysdeps-win.c,dbus/dbus-sysdeps-win.h: removed obsolate DBusUserInfo...
Ralf Habacker [Sat, 30 Jun 2007 11:38:05 +0000 (11:38 +0000)]
* dbus/dbus-sysdeps-win.c,dbus/dbus-sysdeps-win.h: removed obsolate DBusUserInfo code

17 years ago* dbus/dbus-sysdeps-win.c (_dbus_daemon_init): reduced compiler warnings
Ralf Habacker [Sat, 30 Jun 2007 11:29:17 +0000 (11:29 +0000)]
* dbus/dbus-sysdeps-win.c (_dbus_daemon_init): reduced compiler warnings

17 years ago* dbus/dbus-auth-script.c (_dbus_auth_script_run): added UNIX_ONLY and WIN_ONLY comma...
Ralf Habacker [Sat, 23 Jun 2007 07:45:17 +0000 (07:45 +0000)]
* dbus/dbus-auth-script.c (_dbus_auth_script_run): added UNIX_ONLY and WIN_ONLY commands for auth scripts
* test/data/auth/external-root.auth-script: limit execution to unix

17 years ago. readded dbus_watch_get_fd() fix, requires for recent qt releases
Ralf Habacker [Fri, 22 Jun 2007 09:56:53 +0000 (09:56 +0000)]
. readded dbus_watch_get_fd() fix, requires for recent qt releases

17 years ago* dbus/dbus-tranport.c (auth_via_default_rules): made the verbose spam cross-platform
Ralf Habacker [Thu, 21 Jun 2007 21:58:36 +0000 (21:58 +0000)]
* dbus/dbus-tranport.c (auth_via_default_rules): made the verbose spam cross-platform

17 years ago2007-06-21 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Thu, 21 Jun 2007 21:28:59 +0000 (21:28 +0000)]
2007-06-21  Havoc Pennington  <hp@redhat.com>

* dbus/dbus-watch.c (dbus_watch_get_fd): 1) its behavior should
not be the same as before, the ABI has never been declared stable
on Windows and 2) do not commit to cross-platform files without
posting the exact patch to the mailing list

* dbus/dbus-sysdeps-util.c (_dbus_sysdeps_test): as I have pointed
out before, either 0xff needs to work on both platforms, or none
of the dbus code can rely on it working. That means the options
are 1) audit the code for anywhere that relies on 0xff working, if
none found ideally add a test that it *doesn't* work and make unix
reject it explicitly, but in any case this test would go away or
2) make it work on Windows also, then we don't have to figure out
whether we rely on it.

And in either case, post the exact patch to the mailing list and
don't just commit.

17 years ago* cmake/CMakeLists.txt: added VERSION_PATCH to be able to distinguish win32 binary...
Ralf Habacker [Thu, 21 Jun 2007 15:08:50 +0000 (15:08 +0000)]
* cmake/CMakeLists.txt: added VERSION_PATCH to be able to distinguish win32 binary releases from dbus versioning

17 years ago* dbus/dbus-sysdeps-win.c: deleted local DBusCredentials structure
Ralf Habacker [Thu, 21 Jun 2007 13:11:10 +0000 (13:11 +0000)]
* dbus/dbus-sysdeps-win.c: deleted local DBusCredentials structure
    (_dbus_getsid): new function
    (_dbus_read_credentials_socket): fixed
    (_dbus_append_user_from_current_process, _dbus_credentials_add_from_current_process): added real sid reading
    (_dbus_credentials_parse_and_add_desired,_dbus_parse_uid): deleted

17 years ago- recommit with unix eol, cvs does not support native eol
Ralf Habacker [Thu, 21 Jun 2007 12:58:52 +0000 (12:58 +0000)]
- recommit with unix eol, cvs does not support native eol

17 years ago* dbus/dbus-sysdeps-win.c (_dbus_sysdeps_test): don't check 0xff as floating point...
Ralf Habacker [Thu, 21 Jun 2007 12:56:56 +0000 (12:56 +0000)]
* dbus/dbus-sysdeps-win.c (_dbus_sysdeps_test): don't check 0xff as floating point, this isn't supported on win32 math implementation

17 years ago* dbus/dbus-sysdeps-win.c (_dbus_homedir_from_username, _dbus_homedir_from_current_pr...
Ralf Habacker [Thu, 21 Jun 2007 12:34:06 +0000 (12:34 +0000)]
* dbus/dbus-sysdeps-win.c (_dbus_homedir_from_username, _dbus_homedir_from_current_process, _dbus_append_desired_identity): removed obsolate functions

17 years ago* dbus/dbus-sysdeps-win.c, dbus/dbus-sysdeps-spawn-win.c, dbus/dbus-sysdeps-win.h...
Ralf Habacker [Thu, 21 Jun 2007 12:23:30 +0000 (12:23 +0000)]
* dbus/dbus-sysdeps-win.c, dbus/dbus-sysdeps-spawn-win.c, dbus/dbus-sysdeps-win.h: disabled DBusSocket implementation by Peter Kuemmel

17 years ago* dbus/dbus-watch.c (dbus_watch_get_fd): this function is deprecated and its behavior...
Ralf Habacker [Thu, 21 Jun 2007 08:29:24 +0000 (08:29 +0000)]
* dbus/dbus-watch.c (dbus_watch_get_fd): this function is deprecated and its behavior should be as before until all client code is migrated.

17 years ago* dbus/dbus-sysdeps-util-win.c, tools/dbus-launch-win.c: msvc7.1 fixes by Jaroslaw...
Ralf Habacker [Tue, 19 Jun 2007 20:12:11 +0000 (20:12 +0000)]
* dbus/dbus-sysdeps-util-win.c, tools/dbus-launch-win.c: msvc7.1 fixes by Jaroslaw Staniek tested with mingw

17 years ago* dbus/dbus-sysdeps-win.c, dbus/dbus-sysdeps-spawn-win.c, dbus/dbus-sysdeps-win.h...
Ralf Habacker [Tue, 19 Jun 2007 20:04:34 +0000 (20:04 +0000)]
* dbus/dbus-sysdeps-win.c, dbus/dbus-sysdeps-spawn-win.c, dbus/dbus-sysdeps-win.h: disabled DBusSocket implementation Patch by Peter Kuemmel

17 years agoadd a comment about the match rules limit
Havoc Pennington [Tue, 19 Jun 2007 17:23:55 +0000 (17:23 +0000)]
add a comment about the match rules limit

17 years ago* removed obsolate patches
Ralf Habacker [Tue, 19 Jun 2007 16:35:40 +0000 (16:35 +0000)]
* removed obsolate patches

17 years ago* eol fix
Ralf Habacker [Tue, 19 Jun 2007 16:08:09 +0000 (16:08 +0000)]
* eol fix

17 years agofix location of releases in HACKING
Havoc Pennington [Mon, 18 Jun 2007 19:55:50 +0000 (19:55 +0000)]
fix location of releases in HACKING

17 years ago2007-06-18 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Mon, 18 Jun 2007 19:51:18 +0000 (19:51 +0000)]
2007-06-18  Havoc Pennington  <hp@redhat.com>

* configure.in: bump version to 1.1.2 so CVS is higher than last
release (this is not the 1.1.2 release)

17 years ago2007-06-18 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Mon, 18 Jun 2007 19:49:40 +0000 (19:49 +0000)]
2007-06-18  Havoc Pennington  <hp@redhat.com>

* Release 1.1.1

17 years ago2007-06-18 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Mon, 18 Jun 2007 19:32:51 +0000 (19:32 +0000)]
2007-06-18  Havoc Pennington  <hp@redhat.com>

* doc/dbus-specification.xml: document org.freedesktop.DBus.GetId()

* bus/driver.c (bus_driver_handle_get_id): implement org.freedesktop.DBus.GetId()

* bus/bus.c (bus_context_new): generate a unique ID for each bus context

* dbus/dbus-connection.c (dbus_connection_get_server_id): new function

* dbus/dbus-bus.c (dbus_bus_get_id): new function

* dbus/dbus-server.c (dbus_server_get_id): new function

17 years ago2007-06-18 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Mon, 18 Jun 2007 18:05:21 +0000 (18:05 +0000)]
2007-06-18  Havoc Pennington  <hp@redhat.com>

* dbus/dbus-sysdeps-unix.c (_dbus_read_credentials_socket): clean
this up a little bit, to try and understand why telnet'ing to a
server and sending a non-nul byte didn't disconnect immediately;
now it seems that it does disconnect immediately as it should,
though I don't understand what has changed.

17 years ago2007-06-18 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Mon, 18 Jun 2007 16:24:03 +0000 (16:24 +0000)]
2007-06-18  Havoc Pennington  <hp@redhat.com>

* dbus/dbus-watch.c (dbus_watch_get_socket)
(dbus_watch_get_unix_fd): new API to match DBusConnection
(dbus_watch_get_fd): deprecate this

Throughout: just s/dbus_watch_get_fd/dbus_watch_get_socket/g for
now since all the transports use sockets anyway

17 years ago* dbus/dbus-macros.h, dbus/dbus-message.c, dbus/dbus-message.h: renamed DBUS_GNUC_DEP...
Ralf Habacker [Sun, 17 Jun 2007 16:43:39 +0000 (16:43 +0000)]
* dbus/dbus-macros.h, dbus/dbus-message.c, dbus/dbus-message.h: renamed DBUS_GNUC_DEPRECATED  to DBUS_DEPRECATED and extended to msvc compiler

17 years ago- debug options should not be in cvs
Ralf Habacker [Sat, 16 Jun 2007 08:14:47 +0000 (08:14 +0000)]
- debug options should not be in cvs

17 years ago* cmake/CMakeLists.txt: use local include header first
Ralf Habacker [Sat, 16 Jun 2007 08:09:41 +0000 (08:09 +0000)]
* cmake/CMakeLists.txt: use local include header first
* dbus/dbus-sysdeps-win.c: mingw compile fix of DBusCredential struct

17 years ago* cmake/ConfigureChecks.cmake,cmake/config.h.cmake: added check for HAVE_ERRNO_H
Ralf Habacker [Fri, 15 Jun 2007 21:04:43 +0000 (21:04 +0000)]
* cmake/ConfigureChecks.cmake,cmake/config.h.cmake: added check for HAVE_ERRNO_H
* cmake/dbus/CMakeLists.txt: added missing files
* dbus/dbus-transport-win.c/.h: new files
* dbus/dbus-sysdeps-win.c,.h: added required _unix functions to make dbus compilable on win32
* dbus/dbus-sysdeps-win-utils.c,.h: moved some functions to dbus-sysdeps-win.c
* dbus-win.patch: removed applied or obsolate patches
Note: dbus-win32 is now compilable, no guarantee that it runs without any problems

17 years ago2007-06-15 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Fri, 15 Jun 2007 20:10:09 +0000 (20:10 +0000)]
2007-06-15  Havoc Pennington  <hp@redhat.com>

* dbus/dbus-sysdeps-unix.c (_dbus_append_session_config_file)
(_dbus_append_system_config_file): new functions

* bus/main.c (main): use _dbus_append_system_config_file() and
_dbus_append_session_config_file()

* dbus/Makefile.am (INCLUDES): move DBUS_SYSTEM_CONFIG_FILE and
DBUS_SESSION_CONFIG_FILE into this makefile

17 years ago2007-06-15 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Fri, 15 Jun 2007 19:28:36 +0000 (19:28 +0000)]
2007-06-15  Havoc Pennington  <hp@redhat.com>

* dbus/dbus-sysdeps.c (_dbus_set_errno_to_zero)
(_dbus_get_is_errno_nonzero, _dbus_get_is_errno_eintr)
(_dbus_strerror_from_errno): family of functions to abstract
errno, though these are somewhat bogus (really we should make our
socket wrappers not use errno probably - the issue is that any
usage of errno that isn't socket-related probably is not
cross-platform, so should either be in a unix-only file that can
use errno directly, or is a bug - these general errno wrappers
hide issues of this nature in non-socket code, while
socket-specific API changes would not since sockets are allowed
cross-platform)

17 years ago- minor spelling fix
Ralf Habacker [Fri, 15 Jun 2007 18:15:41 +0000 (18:15 +0000)]
- minor spelling fix

17 years ago2007-06-14 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Thu, 14 Jun 2007 22:02:10 +0000 (22:02 +0000)]
2007-06-14  Havoc Pennington  <hp@redhat.com>

* bus/dispatch.c (check_get_connection_unix_process_id): mop up
getpid() (noticed by Peter KKümmel) and adapt the test to
expect a "pid unknown" error when running on Windows.

17 years ago2007-06-14 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Thu, 14 Jun 2007 21:15:33 +0000 (21:15 +0000)]
2007-06-14  Havoc Pennington  <hp@redhat.com>

* dbus/dbus-sysdeps-unix.c (_dbus_credentials_parse_and_add_user):
delete this function since it was effectively the same as
_dbus_credentials_add_from_username()

17 years ago2007-06-14 Havoc Pennington <hp@redhat.com>
Havoc Pennington [Thu, 14 Jun 2007 20:59:16 +0000 (20:59 +0000)]
2007-06-14  Havoc Pennington  <hp@redhat.com>

* dbus/dbus-auth.c: adapt to keyring changes

* dbus/dbus-keyring.c: change to avoid using user ID and home
directory directly; instead use a
keyring-location-from-credentials function in dbus-sysdeps

* fix to use _dbus_append_user_from_current_process() instead of
_dbus_username_from_current_process() or _dbus_append_desired_identity().

17 years ago* reverted global rename of function _dbus_username_from_current_process.
Ralf Habacker [Thu, 14 Jun 2007 20:08:48 +0000 (20:08 +0000)]
* reverted global rename of function _dbus_username_from_current_process.
It needs to much tests to verify that the change does not break anything. I had overseen that the signatures are different (** to *) which requires non trivial changes. This is one *major* disadvantage of elumating oop functionality with c. You are responsible for cleaning every object on every function return point, which could be a nightmare if you are not working with dbus all the days.

17 years ago* dbus/dbus-auth.c (handle_client_initial_response_cookie_sha1_mech): fixed usage...
Ralf Habacker [Thu, 14 Jun 2007 19:28:07 +0000 (19:28 +0000)]
* dbus/dbus-auth.c (handle_client_initial_response_cookie_sha1_mech): fixed usage of _dbus_append_desired_identity()
* dbus/dbus-sysdeps.h (_dbus_username_from_current_process): removed prototype

17 years ago* dbus/dbus-sysdeps.c: moved global lock system_users from dbus-userdb.c
Ralf Habacker [Thu, 14 Jun 2007 18:34:28 +0000 (18:34 +0000)]
* dbus/dbus-sysdeps.c: moved global lock system_users from dbus-userdb.c

17 years ago* global rename of function dbus_username_from_current_process to _dbus_append_desire...
Ralf Habacker [Thu, 14 Jun 2007 18:25:22 +0000 (18:25 +0000)]
* global rename of function dbus_username_from_current_process to _dbus_append_desired_identity

17 years ago* dbus/dbus-sysdeps-win.c: disabled DBusUserInfo related code
Ralf Habacker [Thu, 14 Jun 2007 17:37:27 +0000 (17:37 +0000)]
* dbus/dbus-sysdeps-win.c: disabled DBusUserInfo related code
(_dbus_append_desired_identity,_dbus_windows_user_is_process_owner): new win32 functions as counterpart of unix related
        (_dbus_send_credentials_socket,_dbus_read_credentials_socket): renamed from ..._unix_socket
(_dbus_send_credentials_unix_socket): removed obsolate function
* dbus/dbus-sysdeps-win-util.c: disabled DBusGroupInfo related code
(_dbus_verify_daemon_user,_dbus_change_to_daemon_user): new win32 functions as counterpart of unix related

17 years agodoc/dbus-specification.xml: say the protocol version is 1 instead of 0
Simon McVittie [Thu, 14 Jun 2007 09:50:17 +0000 (09:50 +0000)]
doc/dbus-specification.xml: say the protocol version is 1 instead of 0
(patch from Kristoffer Lundén, fd.o#10033)
Also remove the FIXME about removing protocol version from messages (as per
Havoc's comment on that bug)