* configure.ac, bus/Makefile.am: Generalize kqueue support so that
it works on any system providing this interface, not only FreeBSD.
For example, NetBSD. (Patch by Julio M. Merino Vidal <jmmv at NetBSD dot org>)
Patch provided by Ralf Habacker (ralf dot habacker at freenet dot de)
* dbus/dbus-sysdeps.c, dbus/dbus-threads.c, dbus/dbus-internals.h:
Add two more global locks for use on windows platforms. These are
unused on non-windows platforms but are not ifdefed out to avoid
potential bugs (i.e. the few bytes lost does not warrent the extra
maintanence and complexity that having seperate sets of locks would
cause)
* configure.in: use instead of
to determine where to store the pid since the init scripts manipulate
the pid file (patch from Marcelo Ricardo Leitner
<marcelo dot leitner at terra dot com dot br>.
* dbus/dbus-threads.c: Add static DBusList *uninitialized_mutex_list and
static DBusList *uninitialized_condvar_list to support new late
initialization threading model. In this model threads can be initialized
even after the D-Bus API has been used but still needs to be initialized
before the second thread has been started. Mutexes and condvar addresses
are stored in the two static lists and are replaced with actuall locks
when threads are initalized.
(_dbus_mutex_new_at_location): New method for creating a mutex and placing
the location into the static list
(_dbus_mutex_free_at_location): New method for removing a mutex location
from the static list and freeing the mutex
(_dbus_condvar_new_at_location): New method for creating a conditional
variable and placing the location into the static list
(_dbus_condvar_free_at_location): New method for removing a conditional
variable location from the static list and freeing the conditional variable
(init_uninitialized_locks): Atomic method which goes through the static
lists of mutex and condvar location and updates them with actuall locks
(init_global_locks): changed to init_locks
* dbus/dbus-connection.c:
(_dbus_connection_test_get_locks): New method for tests to check connections
(_dbus_connection_new_for_transport): Use the new at_location mutex and
condvar API
(dbus_connection_allocate_data_slot): Pass in the global lock address
to _dbus_data_slot_allocator_alloc
* dbus/dbus-dataslot.c:
(_dbus_data_slot_allocator_alloc): Use the address of the mutex
instead of the mutex itself
* dbus/dbus-message.c:
(dbus_message_allocate_data_slot): Pass in the global lock address
to _dbus_data_slot_allocator_alloc
* dbus/dbus-pending-call.c:
(dbus_pending_call_allocate_data_slot): Pass in the global lock address
to _dbus_data_slot_allocator_alloc
* dbus/dbus-server.c:
(_dbus_server_init_base): Use the new at_location mutex API
(dbus_server_allocate_data_slot): Pass in the global lock address
to _dbus_data_slot_allocator_alloc
* test/name-test/test-threads-init.c: New test case for late thread
initialization
* dbus/dbus-dataslot.c (_dbus_data_slot_allocator_alloc):
Change _dbus_abort to _dbus_assert_not_reached because _dbus_abort
causes compile problems when asserts are turned off
Keeping _dbus_warn for printing out the message so even if
asserts are turned off the user gets the messages that something is
wrong
* dbus/dbus-marshal-recursive-util.c (node_new):
return a NULL instead of FALSE since the return type
is not expecting a boolean
* dbus/dbus-server-debug-pipe.c (_dbus_transport_debug_pipe_new):
Send a NULL into _dbus_transport_new_for_fd instead of a FALSE
because we are expecting a pointer not a boolean
* dbus/dbus-sysdeps-util.c (_dbus_get_tmpdir):
add void as the parameter so some compilers
don't complain
* dbus/dbus-transport-unix.c (_dbus_transport_new_for_domain_socket,
_dbus_transport_new_for_tcp_socket):
Send a NULL into _dbus_transport_new_for_fd instead of a FALSE
because we are expecting a pointer not a boolean
* test/shell-test.c (test_command_line):
cast the second argument to _dbus_list_append to avoid compiler
warnings
* test/test-service.c (check_hello_from_self_reply):
Initialize echo_message and echo_reply to NULL
* test/test-shell-service.c (handle_echo):
Remove unused variable and cast the third parameter passed to
dbus_connection_get_object_path_data to avoid compiler warrnings
* test/name-test/test-names.c (clear_message_queue):
Remove unused function
* test/name-test/test-pending-call-dispatch.c:
Fix format string in printf
* dbus/dbus-bus.c:
* test/name-test/test-names.c:
Remove test method dbus_bus_connection_get_unique_name because
we already have public method dbus_bus_get_unique_name that
does the exact same thing
* dbus/dbus-address.c (_dbus_address_test): Revert leaking strcmp.
In any case it was wrong since this is a test checking to see if
address parsing is correct. There was no need to get the true
tmp directory.
* dbus/dbus-signature.c:
Fix the unit tests so that they work if assertions are not enabled.
(patch from Julio M. Merino Vidal <jmmv at NetBSD dot org>)
* Change an assert to a detailed warning and abort.
When allocator->lock != mutex it indicates that the user has failed
to initalize threads before using the D-Bus library. This warning
helps the user identify the issue and fix their app.
* Doxyfile.in: fix the INPUT line so we can generate docs
when sourcedir != builddir (patch from Cygwin Ports maintainer
<yselkowitz at users dot sourceforge dot net>
* dbus/dbus-sysdeps.h:
* dbus/dbus-sysdeps.c:
* dbus/dbus-string.c:
s/_dbus_printf_length/_dbus_printf_string_upper_bound to comform with
GLib's function which does the same thing
* configure.in:
* bus/Makefile.am:
* bus/dir-watch-default.c:
* bus/dir-watch-dnotify.c:
* bus/dir-watch-kqueue.c:
Add kqueue directory watching for freebsd and split the directory
watching code into seperate files per method/arch
(patches from Timothy Redaelli <drizzt at gufi dotorg>)
* configure.in:
* tools/Makefile.am:
* tools/dbus-launch.c:
* bus/Makefile.am:
allow --with-dbus-daemondir switch to be used to make the
daemon install to a seperate bindir like /usr/libexec
(patch from Brian Cameron <brian dot cameron at sun dot com)
* bus/Makefile.am (install-data-hook): removed the slash after
$(DESTDIR) so we don't get the double slash which does not work
in windows (patch from Andras Porjesz
<andras dot porjesz at siemens dot com>
* dbus/dbus-internals.c: unistd.h is not used
(_dbus_verbose_real): only enable verbose printing
if DBUS_VERBOSE environment var is set to '1'
(patch from Peter Kümmel <syntheticpp at gmx dot net>
* confifure.in: add a GCC flag check for -Wdeclaration-after-statement
so we still compile on older gcc (patch from Frederic Peters
<fpeters at 0d dot be>
* dbus/dbus-connection.c:
(_dbus_connection_attach_pending_call_unlocked):
(connection_timeout_and_complete_all_pending_calls_unlocked):
Make sure we set timeout_added on pending calls to FALSE when
we remove the timeout from the connection
* Removed some extra bindings stuff lingering around (thanks timo)
* dbus-pendingcall.c (_dbus_pending_call_new):
s/dbus_connection_ref/_dbus_connection_ref_unlocked fixes assertion
when we tried to take a lock on an already locked connection
* dbus-connection.c (dbus_connection_send_with_reply): return TRUE
and set pending_reply out arg to NULL is connection is disconnected
(connection_timeout_and_complete_all_pending_calls_unlocked): New
static method for cleaning up pending calls on disconnect
(_dbus_connection_get_dispatch_status_unlocked): If we have pending
calls queued timeouts on disconnect
* dbus/dbus-connection.c (dbus_connection_close): removed deprecated
function
(dbus_connection_dispatch): On disconnect unref any shared connections
* dbus/dbus-bus.c (_dbus_bus_check_connection_and_unref): new function
for cleaning up shared connections on disconnect
(internal_bus_get): get a hard refrence to shared connections when
they are created
* doc/TODO: Remove items which are no longer relevent or have been fixed
Split 1.0 todo items with a 0.90 freeze todo list
Thiago Macieira [Mon, 12 Jun 2006 09:18:45 +0000 (09:18 +0000)]
* qt/src/qdbusintegrator.cpp: Fix bug in parsing async methods
that took a QDBusMessage parameter.
* qt/src/qdbusbus.h: Add a default flag for RequestName.
* qt/tools/dbus.cpp: Don't use automatic call because we might
be calling an async method: request a reply.
Thiago Macieira [Sun, 11 Jun 2006 12:18:23 +0000 (12:18 +0000)]
* test/qt/*: Update the testcases, including testing the new
functionality of sending null QByteArray and QString over the
bus. Add new headertest test and restore the old
qdbusxmlparser test.
Thiago Macieira [Sun, 11 Jun 2006 12:16:30 +0000 (12:16 +0000)]
* qt/tools/dbuscpp2xml.cpp: Compile on Windows.
* qt/tools/dbusidl2cpp.cpp: Add missing newline.
* qt/examples/Makefile.am:
* qt/examples/chat.h: Use UI-generated files with the ui_*.h
form.
* qt/src/qdbusmarshall.cpp: Allow sending of QString() and
QByteArray() (nulls) over the bus.
* qt/src/qdbusabstractinterface.cpp: Use the correct variable,
the one that has the signature suffix stripped.
* qt/src/qdbusreply.h: Make some methods const.
Thiago Macieira [Fri, 9 Jun 2006 21:43:14 +0000 (21:43 +0000)]
Patch from Timo Hoenig <thoenig@suse.de>.
* qt/dbus/Makefile.am: New file. Fix "make dist", add all
headers required during build to EXTRA_DIST.
* qt/src/Makefile.am: Fix "make dist", add 'qdbus.h' to
EXTRA_DIST.
* qt/Makefile.am: Fix "make dist", add 'dbus' to DIST_SUBDIRS.
* configure.in: Fix "make dist", take care that the Makefile
for qt/dbus is being generated.
Robert McQueen [Wed, 7 Jun 2006 00:03:57 +0000 (00:03 +0000)]
2005-05-06 Robert McQueen <robot101@debian.org>
* glib/dbus-gvalue-utils.c: Fix the failing test where static string
pointers were put into a GPtrArray-based specialised collection, and
then freed along with the array. GValues which you add into
collections or maps which have the NOCOPY flag set are assumed to not
belong to the caller, so rather than the existing pointer-stealing
semantics, they are copied instead. Given that the main consumers of
this abstraction are the bindings themselves, I don't think this is
too bad, but others should watch their choice of take vs set_static.
Robert McQueen [Tue, 6 Jun 2006 23:07:04 +0000 (23:07 +0000)]
2005-05-06 Robert McQueen <robot101@debian.org>
* glib/dbus-gvalue-utils.c: Spotted a warning about the return value
of g_slist_prepend not being used. Fixed copying of slist-based
specialised collections, then wrote a test case and found that it was
all broken. Went on to fix iterating and appending too. Good thing
nobody uses this code yet.
Robert McQueen [Tue, 6 Jun 2006 19:45:39 +0000 (19:45 +0000)]
2005-05-06 Robert McQueen <robot101@debian.org>
* glib/dbus-gvalue-utils.c: Remove duplicated code by having all of
the iterators use gvalue_take_ptrarray_value (the GValues themselves
are discarded without unsetting, so it makes no difference whether
we take or set_static). Remove cases for G_TYPE_POINTER because
there really is nothing useful we can do with them in our
specialised types - we *need* boxed copy/free functions at the very
least.
Thiago Macieira [Mon, 5 Jun 2006 18:13:07 +0000 (18:13 +0000)]
* qt/dbus: Add directory. I had forgotten to add this
yesterday after the move...
* qt/examples/Makefile.am:
* qt/examples/dbus.cpp: Moved to qt/tools/dbus.cpp.
* qt/tools/Makefile.am:
* qt/tools/dbus.cpp: Moved from qt/examples/dbus.cpp.
Added feature to get and set properties.
Added validation of service, object path and interface names.
* qt/tools/dbusidl2cpp.cpp: Two new features:
1) Allow specifying both the header and the source file names,
by separating them with a colon.
2) Don't write an interface output if the -p switch wasn't
given, but the -a was.
* qt/src/*: Fix usage of Iterators and ConstIterators.
Fix shadowing of variables by other variables (-Wshadow).
Fix keyword-cleanliness in headers.
Fix ASCII-cast (QLatin1String, QLatin1Char).
Fix validation of member names.
Add extra checking of introspection data during XML parsing.
Various bug fixes.
Thiago Macieira [Sun, 4 Jun 2006 15:52:05 +0000 (15:52 +0000)]
* qt/: Update to Subversion r548032.
This includes a big reorganisation of the files inside the
subdir.
We really need a version control system that supports moving of
files. I'm not bothering with history anyways anymore, since the
bindings will be moved out to git. The history should be restored from
Subversion when that happens.
Sjoerd Simons [Tue, 30 May 2006 15:34:10 +0000 (15:34 +0000)]
* dbus/dbus-sysdeps.c: Make tcp socket connection error somewhat more
clear:
"Failed to connect to socket <host>:<port> <error>" instead of
"Failed to connect to socket <host>: <error>:<port>"
* dbus/dbus-transport-unix.c: Fix crash when no host option is given
for a tcp transport.
Thiago Macieira [Mon, 29 May 2006 18:17:09 +0000 (18:17 +0000)]
* qt/*: Update the QtDBus bindings up to revision 546310 in
Subversion.
This adds the dbuscpp2xml tool, that parses a C++ header and
outputs a D-BUS Introspection XML.
Havoc Pennington [Sun, 21 May 2006 05:33:52 +0000 (05:33 +0000)]
2006-05-21 Havoc Pennington <hp@redhat.com>
* glib/dbus-gproxy.c: Put in a pile of assertions that the proxy name
is not NULL when it shouldn't be. Also a couple of possible fixes
for #4637 though I don't understand why the bug happens, to be
honest... also the object constructor has an assert name != NULL
and the name is only currently NULL for peer-to-peer proxies that
I don't think anyone uses? So it should be asserting.
Anyway, for now at least see if we get an earlier assertion failure.
* glib/dbus-gvalue-utils.c: Put in a couple of assertions for
apparently broken code to be sure the tests fail and someone
will fix them...
Thiago Macieira [Sun, 7 May 2006 09:36:19 +0000 (09:36 +0000)]
* qt/qdbusmarshall.cpp: Fix a problem of demarshalling lists
and arrays when they had a single element: has_next returns
false, even before you read the element. So, instead, check
the array length.
Thiago Macieira [Tue, 2 May 2006 14:00:27 +0000 (14:00 +0000)]
* qt/dbusidl2cpp.cpp: There's no callAsync. Use the correct
call (r535506)
* qt/dbusidl2cpp.cpp:
* qt/qdbusabstractadaptor.cpp:
* qt/qdbusabstractadaptor.h: Make QDBusAdaptorConnector be a
sibling of the QDBusAbstractAdaptor objects instead of the
parent. (r535848)
* qt/dbusidl2cpp.cpp:
* qt/qdbusabstractinterface.cpp:
* qt/qdbusabstractinterface.h:
* qt/qdbusabstractinterface_p.h:
* qt/qdbusinterface.cpp: Make properties in interfaces
actually work. The code that was generated would not compile,
due to moc calls to functions that did not exist. They now
shall. (r536571)
* qt/qdbusinterface.h: Rename QDBusRef to QDBusInterfacePtr
and disable the copy operators. (r533772, r534746)
* qt/qdbuserror.h: Remove the automatic cast to
bool. (r533929)
* qt/qdbusabstractinterface.cpp:
* qt/qdbusabstractinterface.h: Change the default call mode to
not use the event loop. Add convenience call() methods that
take a CallMode parameter. (r534042)
* qt/qdbusconnection.h: Change the default call mode to not
use the event loop. (r534042)
* qt/qdbusinterface.cpp:
* qt/qdbusinterface.h: Add a method to tell us if the
interface is valid (since we don't return a null pointer
anymore) (r534099)
* qt/qdbusinterface_p.h: Don't crash if metaObject is 0
(r534101)
* qt/qdbusinternalfilters.cpp: Decouple the introspection
function in two so taht we get the chance to introspect
without having a QDBusMessage (r534102)
* qt/qdbusbus.h:
* qt/qdbusconnection.cpp:
* qt/qdbusconnection_p.h:
* qt/qdbusintegrator.cpp: Keep a list of our own names to
avoid a round-trip to the server when attempting to introspect
one of our own objects. Also make sure the filter functions
match the empty interface as well. (r534108)
Don't keep the connection names. Instead, trust the unique
connection name (r534111)
Remove event loop usage (r534112)
projects / thirdparty / dbus.git / log
