1 2003-10-15 Seth Nickell <seth@gnome.org>
5 Include dbus_h_wrapper.h in the dist tarball.
7 2003-10-14 Havoc Pennington <hp@redhat.com>
9 * bus/bus.c (bus_context_check_security_policy): revamp this to
10 work more sanely with new policy-based requested reply setup
12 * bus/connection.c (bus_transaction_send_from_driver): set bus
13 driver messages as no reply
15 * bus/policy.c (bus_client_policy_check_can_receive): handle a
16 requested_reply attribute on allow/deny rules
18 * bus/system.conf: add <allow requested_reply="true"/>
20 * bus/driver.c (bus_driver_handle_message): fix check for replies
21 sent to the bus driver, which was backward. How did this ever work
22 at all though? I think I'm missing something.
24 * dbus/dbus-message.c (decode_header_data): require error and
25 method return messages to have a reply serial field to be valid
26 (_dbus_message_loader_queue_messages): break up this function;
27 validate that reply serial and plain serial are nonzero;
28 clean up the OOM/error handling.
29 (get_uint_field): don't return -1 from this
30 (dbus_message_create_header): fix signed/unsigned bug
32 * bus/connection.c (bus_connections_expect_reply): save serial of
33 the incoming message, not reply serial
35 2003-10-14 Havoc Pennington <hp@redhat.com>
37 * bus/connection.c: implement pending reply tracking using
40 * bus/bus.c (bus_context_check_security_policy): verify that a
41 reply is pending in order to allow a reply to be sent. Deny
42 messages of unknown type.
44 * bus/dbus-daemon-1.1.in: update to mention new resource limits
46 * bus/bus.c (bus_context_get_max_replies_per_connection): new
47 (bus_context_get_reply_timeout): new
49 2003-10-13 Seth Nickell <seth@gnome.org>
53 Pass "make distcheck": remove a couple files from DIST_FILES
54 that weren't included in the final version.
56 2003-10-12 Havoc Pennington <hp@pobox.com>
58 Added test code that 1) starts an actual bus daemon and 2) uses
59 DBusGProxy; fixed bugs that were revealed by the test. Lots
60 more testing possible, but this is the basic framework.
62 * glib/dbus-gproxy.c (dbus_gproxy_manager_unregister): remove
63 empty proxy lists from the proxy list hash
65 * dbus/dbus-message.c (dbus_message_iter_get_args_valist): add a
66 couple of return_if_fail checks
68 * dbus/dbus-pending-call.c (_dbus_pending_call_new): use dbus_new0
69 to allocate, so everything is cleared to NULL as it should be.
71 * glib/dbus-gmain.c (dbus_connection_setup_with_g_main): pass
72 source as data to dbus_connection_set_timeout_functions() as the
73 timeout functions expected
75 * test/glib/run-test.sh: add a little script to start up a message
76 bus and run tests using it
78 * tools/dbus-launch.1: updates
80 * tools/dbus-launch.c (main): add --config-file option
82 * tools/dbus-launch.c (main): remove confusing else if (runprog)
83 that could never be reached.
85 * dbus/dbus-message.c (dbus_message_new_method_return)
86 (dbus_message_new_error, dbus_message_new_signal): set the
87 no-reply-expected flag on all these. Redundant, but may
88 as well be consistent.
90 2003-10-11 Havoc Pennington <hp@pobox.com>
92 * test/decode-gcov.c (function_solve_graph): make broken block
93 graph a nonfatal error since it seems to be broken. Need to debug
96 * dbus/dbus-marshal.c (_dbus_type_is_valid): new function since we
97 can't just check type > INVALID < LAST anymore
99 * dbus/dbus-message.c (dbus_message_get_signature): new function
100 (dbus_message_has_signature): new function
101 (struct DBusMessage): add signature field (right now it isn't sent
102 over the wire, just generated on the fly)
103 (dbus_message_copy): copy the signature, and init strings to
104 proper length to avoid some reallocs
105 (dbus_message_iter_init_array_iterator): return void, since it
107 (dbus_message_iter_init_dict_iterator): return void since it can't fail
108 (_dbus_message_loader_queue_messages): add silly temporary hack to
109 fill in message->signature on load
111 * dbus/dbus-protocol.h: change DBUS_TYPE_* values to be ASCII
112 characters, so they are relatively human-readable.
114 2003-10-11 Havoc Pennington <hp@pobox.com>
116 * dbus/dbus-message.c (_dbus_message_test): add more test
117 coverage, but #if 0 for now since they uncover a bug
118 not fixed yet; I think in re_align_field_recurse()
119 (re_align_field_recurse): add FIXME about broken assertion
121 * dbus/dbus-sysdeps.c (_dbus_sysdeps_test): add more test coverage
123 * bus/connection.c: share a couple code bits with expirelist.c
125 * bus/expirelist.h, bus/expirelist.c: implement a generic
126 expire-items-after-N-seconds facility, was going to share between
127 expiring connections and replies, decided not to use for expiring
130 * COPYING: include AFL 2.0 (still need to change all the file headers)
132 2003-10-09 Havoc Pennington <hp@redhat.com>
134 * configure.in: define DBUS_HAVE_GCC33_GCOV if we have
135 gcc 3.3. Not that we do anything about it yet.
137 * bus/signals.c (bus_match_rule_parse): impose max length on the
140 * dbus/dbus-protocol.h: add DBUS_MAXIMUM_MATCH_RULE_LENGTH
142 2003-10-09 Havoc Pennington <hp@redhat.com>
144 Make matching rules theoretically work (add parser).
146 * bus/bus.c (bus_context_check_security_policy): fix up to handle
147 the case where destination is explicitly specified as bus driver
148 and someone else is eavesdropping.
150 * bus/policy.c (bus_client_policy_check_can_receive): fix up
151 definition of eavesdropping and assertion
153 * tools/dbus-send.c (main): use dbus_message_type_from_string
155 * bus/signals.c (bus_match_rule_parse): implement
157 * dbus/dbus-message.c (dbus_message_type_from_string): new
159 * dbus/dbus-errors.h (DBUS_ERROR_MATCH_RULE_INVALID): add
161 2003-10-02 Havoc Pennington <hp@pobox.com>
163 * glib/dbus-gproxy.c (dbus_gproxy_call_no_reply): rename from
164 dbus_gproxy_oneway_call
166 * glib/dbus-gmain.c (dbus_connection_setup_with_g_main)
167 (dbus_server_setup_with_g_main): fix to allow calling them more
168 than once on the same args
169 (dbus_bus_get_with_g_main): new function
171 2003-10-02 Havoc Pennington <hp@redhat.com>
173 * doc/dbus-tutorial.xml: write some stuff
175 2003-09-29 Havoc Pennington <hp@pobox.com>
177 * configure.in: split checks for Doxygen from XML docs, check for
180 * doc/Makefile.am: XML-ify all the docs, and add a blank
183 2003-09-29 Havoc Pennington <hp@pobox.com>
185 * Merge dbus-object-names branch. To see the entire patch
186 do cvs diff -r DBUS_OBJECT_NAMES_BRANCHPOINT -r dbus-object-names,
188 To revert, I tagged DBUS_BEFORE_OBJECT_NAMES_MERGE.
190 2003-09-28 Havoc Pennington <hp@pobox.com>
192 * HACKING: update to reflect new server
194 2003-09-26 Seth Nickell <seth@gnome.org>
197 * python/examples/example-signals.py:
199 Start implementing some notions of signals. The API
200 is really terrible, but they sort of work (with the
201 exception of being able to filter by service, and to
202 transmit signals *as* a particular service). Need to
203 figure out how to make messages come from the service
206 * python/dbus_bindings.pyx.in:
208 Removed duplicate message_handler callbacks.
210 2003-09-25 Havoc Pennington <hp@redhat.com>
212 * bus/session.conf.in: fix my mess
214 2003-09-25 Havoc Pennington <hp@pobox.com>
216 * bus/session.conf.in: fix security policy, reported by Seth Nickell
218 2003-09-25 Seth Nickell <seth@gnome.org>
220 * python/examples/example-service.py:
222 Johan notices complete wrong code in example-service, but
223 completely wrong in a way that works exactly the same (!).
224 Johan is confused, how could this possibly work? Example
225 code fails to serve purpose of making things clear.
228 2003-09-25 Mark McLoughlin <mark@skynet.ie>
230 * doc/dbus-specification.sgml: don't require header fields
231 to be 4-byte aligned and specify that fields should be
232 distinguished from padding by the fact that zero is not
235 * doc/TODO: remove re-alignment item and add item to doc
236 the OBJECT_PATH type.
238 * dbus/dbus-message.c:
239 (HeaderField): rename the original member to value_offset
240 and introduce a name_offset member to keep track of where
241 the field actually begins.
242 (adjust_field_offsets): remove.
243 (append_int_field), (append_uint_field),
244 (append_string_field): don't align the start of the header
245 field to a 4-byte boundary.
246 (get_next_field): impl finding the next marhsalled field
248 (re_align_field_recurse): impl re-aligning a number of
249 already marshalled fields.
250 (delete_field): impl deleting a field of any type and
251 re-aligning any following fields.
252 (delete_int_or_uint_field), (delete_string_field): remove.
253 (set_int_field), (set_uint_field): no need to re-check
254 that we have the correct type for the field.
255 (set_string_field): ditto and impl re-aligning any
257 (decode_header_data): update to take into account that
258 the fields aren't 4-byte aligned any more and the new
259 way to distinguish padding from header fields. Also,
260 don't exit when there is too much header padding.
261 (process_test_subdir): print the directory.
262 (_dbus_message_test): add test to make sure a following
263 field is re-aligned correctly after field deletion.
265 * dbus/dbus-string.[ch]:
266 (_dbus_string_insert_bytes): rename from insert_byte and
267 allow the insert of multiple bytes.
268 (_dbus_string_test): test inserting multiple bytes.
270 * dbus/dbus-marshal.c: (_dbus_marshal_set_string): add
271 warning note to docs about having to re-align any
272 marshalled values following the string.
274 * dbus/dbus-message-builder.c:
275 (append_string_field), (_dbus_message_data_load):
276 don't align the header field.
278 * dbus/dbus-auth.c: (process_test_subdir): print the
281 * test/break-loader.c: (randomly_add_one_byte): upd. for
284 * test/data/invalid-messages/bad-header-field-alignment.message:
287 * test/data/valid-messages/unknown-header-field.message: shove
288 a dict in the unknown field.
290 2003-09-25 Seth Nickell <seth@gnome.org>
293 * python/dbus_bindings.pyx.in:
295 Handle return values.
297 * python/examples/example-client.py:
298 * python/examples/example-service.py:
300 Pass back return values from the service to the client.
302 2003-09-24 Seth Nickell <seth@gnome.org>
306 Connect Object methods (when you are sharing an object) up... pass
307 in a list of methods to be shared. Sharing all the methods just
308 worked out too weird. You can now create nice Services over the
311 * python/dbus_bindings.pyx.in:
313 Keep references to user_data tuples passed into C functions so
314 Python doesn't garbage collect on us.
316 Implement MethodReturn and Error subclasses of Message for creating
317 DBusMessage's of those types.
319 * python/examples/example-client.py:
320 * python/examples/example-service.py:
322 Simple example code showing both how create DBus services and objects,
325 2003-09-23 Havoc Pennington <hp@pobox.com>
327 * glib/dbus-gproxy.c (dbus_gproxy_manager_filter): implement
329 2003-09-23 Havoc Pennington <hp@redhat.com>
331 * glib/dbus-gproxy.c (dbus_gproxy_connect_signal): implement
332 (dbus_gproxy_disconnect_signal): implement
333 (dbus_gproxy_manager_remove_signal_match): implement
334 (dbus_gproxy_manager_add_signal_match): implement
335 (dbus_gproxy_oneway_call): implement
337 2003-09-23 Havoc Pennington <hp@pobox.com>
339 * glib/dbus-gproxy.c (struct DBusGProxy): convert to a GObject
340 subclass. This means dropping the transparent thread safety of the
341 proxy; you now need a separate proxy per-thread, or your own
342 locking on the proxy. Probably right anyway.
343 (dbus_gproxy_ref, dbus_gproxy_unref): nuke, just use g_object_ref
345 2003-09-22 Havoc Pennington <hp@redhat.com>
347 * glib/dbus-gproxy.c (dbus_gproxy_manager_get): implement
349 2003-09-21 Seth Nickell <seth@gnome.org>
351 First checkin of the Python bindings.
354 * python/Makefile.am:
355 * python/dbus_bindings.pyx.in:
356 * python/dbus_h_wrapper.h:
358 Pieces for Pyrex to operate on, building a dbus_bindings.so
359 python module for low-level access to the DBus APIs.
363 High-level Python module for accessing DBus objects.
368 Build stuff for the python bindings.
372 Extra macro needed for finding the Python C header files.
374 2003-09-21 Havoc Pennington <hp@pobox.com>
376 * glib/dbus-gproxy.c (dbus_gproxy_manager_new): start
377 implementing the proxy manager, didn't get very far.
379 * dbus/dbus-bus.c (dbus_bus_add_match): new
380 (dbus_bus_remove_match): new
382 * glib/dbus-gproxy.c (dbus_gproxy_new_for_service): add a
383 path_name argument; adjust the other not-yet-implemented
384 gproxy constructors to be what I think they should be.
386 2003-09-21 Havoc Pennington <hp@pobox.com>
388 * dbus/dbus-bus.c (dbus_bus_get): set exit_on_disconnect to TRUE
389 by default for message bus connections.
391 * dbus/dbus-connection.c (dbus_connection_dispatch): exit if
392 exit_on_disconnect flag is set and we process the disconnected
394 (dbus_connection_set_exit_on_disconnect): new function
396 2003-09-21 Havoc Pennington <hp@pobox.com>
398 Get matching rules mostly working in the bus; only actually
399 parsing the rule text remains. However, the client side of
400 "signal connections" hasn't been started, this patch is only the
403 * dbus/dispatch.c: fix for the matching rules changes
405 * bus/driver.c (bus_driver_handle_remove_match)
406 (bus_driver_handle_add_match): send an ack reply from these
409 * glib/dbus-gproxy.c (dbus_gproxy_begin_call): fix order of
410 arguments, reported by Seth Nickell
412 * bus/config-parser.c (append_rule_from_element): support
413 eavesdrop=true|false attribute on policies so match rules
414 can be prevented from snooping on the system bus.
416 * bus/dbus-daemon-1.1.in: consistently use terminology "sender"
417 and "destination" in attribute names; fix some docs bugs;
418 add eavesdrop=true|false attribute
420 * bus/driver.c (bus_driver_handle_add_match)
421 (bus_driver_handle_remove_match): handle AddMatch, RemoveMatch
424 * dbus/dbus-protocol.h (DBUS_SERVICE_ORG_FREEDESKTOP_BROADCAST): get
425 rid of broadcast service concept, signals are just always broadcast
427 * bus/signals.c, bus/dispatch.c, bus/connection.c, bus/bus.c:
428 mostly implement matching rules stuff (currently only exposed as signal
431 2003-09-21 Mark McLoughlin <mark@skynet.ie>
433 * doc/dbus-specification.sgml: Change the header field name
434 to be an enum and update the rest of the spec to reference
435 the fields using the conventinal name.
437 * dbus/dbus-protocol.h: update to reflect the spec.
439 * doc/TODO: add item to remove the 4 byte alignment requirement.
441 * dbus/dbus-message.c: Remove the code to generalise the
442 header/body length and serial number header fields as named
443 header fields so we can reference field names using the
445 (append_int_field), (append_uint_field), (append_string_field):
446 Append the field name as a byte rather than four chars.
447 (delete_int_or_uint_field), (delete_string_field): reflect the
448 fact that the field name and typecode now occupy 4 bytes instead
450 (decode_string_field), (decode_header_data): update to reflect
451 protocol changes and move the field specific encoding from
452 decode_string_field() back into decode_header_data().
454 * dbus/dbus-internals.[ch]: (_dbus_header_field_to_string):
455 Add utility to aid debugging.
457 * dbus/dbus-message-builder.c:
458 (append_string_field), (_dbus_message_data_load): Update to
459 reflect protocol changes; Change the FIELD_NAME directive
460 to HEADER_FIELD and allow it to take the field's conventional
461 name rather than the actual value.
463 * test/data/*/*.message: Update to use HEADER_FIELD instead
464 of FIELD_NAME; Always align the header on an 8 byte boundary
465 *before* updating the header length.
467 2003-09-15 Havoc Pennington <hp@pobox.com>
469 * dbus/dbus-pending-call.c: add the get/set object data
470 boilerplate as for DBusConnection, etc. Use generic object data
471 for the notify callback.
473 * glib/dbus-gparser.c (parse_node): parse child nodes
475 * tools/dbus-viewer.c: more hacking on the dbus-viewer
477 * glib/dbus-gutils.c (_dbus_gutils_split_path): add a file to
478 contain functions shared between the convenience lib and the
481 * glib/Makefile.am (libdbus_glib_1_la_LDFLAGS): add
482 -export-symbols-regex to the GLib library
484 * dbus/dbus-object-tree.c (_dbus_object_tree_dispatch_and_unlock):
485 fix the locking in here, and add a default handler for
486 Introspect() that just returns sub-nodes.
488 2003-09-14 Havoc Pennington <hp@pobox.com>
490 * glib/dbus-gthread.c (dbus_g_thread_init): rename to make g_foo
491 rather than gfoo consistent
493 * glib/dbus-gproxy.h: delete for now, move contents to
494 dbus-glib.h, because the include files don't work right since we
495 aren't in the dbus/ subdir.
497 * glib/dbus-gproxy.c (dbus_gproxy_send): finish implementing
498 (dbus_gproxy_end_call): finish
499 (dbus_gproxy_begin_call): finish
501 * glib/dbus-gmain.c (dbus_set_g_error): new
503 * glib/dbus-gobject.c (handle_introspect): include information
504 about child nodes in the introspection
506 * dbus/dbus-connection.c (dbus_connection_list_registered): new
507 function to help in implementation of introspection
509 * dbus/dbus-object-tree.c
510 (_dbus_object_tree_list_registered_and_unlock): new function
512 2003-09-12 Havoc Pennington <hp@pobox.com>
514 * glib/dbus-gidl.h: add common base class for all the foo_info
517 * tools/dbus-viewer.c: add GTK-based introspection UI thingy
520 * test/Makefile.am: try test srcdir -ef . in addition to test
521 srcdir = ., one of them should work (yeah lame)
523 * glib/Makefile.am: build the "idl" parser stuff as a convenience
526 * glib/dbus-gparser.h: make description_load routines return
527 NodeInfo* not Parser*
529 * Makefile.am (SUBDIRS): build test dir after all library dirs
531 * configure.in: add GTK+ detection
533 2003-09-07 Havoc Pennington <hp@pobox.com>
535 * Make Doxygen contented.
537 2003-09-07 Havoc Pennington <hp@pobox.com>
539 * doc/dbus-specification.sgml: more updates
541 2003-09-06 Havoc Pennington <hp@pobox.com>
543 * doc/dbus-specification.sgml: partial updates
545 * bus/dbus-daemon-1.1.in: fix the config file docs for the
546 zillionth time; hopefully I edited the right file this time.
548 * bus/config-parser.c (append_rule_from_element): support
549 send_type, send_path, receive_type, receive_path
551 * bus/policy.c: add message type and path to the list of things
552 that can be "firewalled"
554 2003-09-06 Havoc Pennington <hp@pobox.com>
556 * dbus/dbus-connection.c (dbus_connection_register_fallback): add this
557 (dbus_connection_register_object_path): make this not handle
558 messages to paths below the given path
560 2003-09-03 Havoc Pennington <hp@pobox.com>
562 * test/glib/Makefile.am: add this with random glib-linked test
565 * glib/Makefile.am: remove the random test programs from here,
566 leave only the unit tests
568 * glib/dbus-gobject.c (_dbus_gobject_test): add test for
569 uscore/javacaps conversion, and fix
570 (get_object_property, set_object_property): change to .NET
571 convention for mapping props to methods, set_FooBar/get_FooBar,
572 since one language has such a convention we may as well copy it.
573 Plus real methods in either getFooBar or get_foo_bar style won't
574 collide with this convention.
576 2003-09-01 Havoc Pennington <hp@pobox.com>
578 * glib/dbus-gparser.c: implement
580 * glib/dbus-gobject.c: start implementing skeletons support
582 * configure.in: when disabling checks/assert, also define
583 G_DISABLE_ASSERT and G_DISABLE_CHECKS
585 2003-09-01 Havoc Pennington <hp@pobox.com>
587 * glib/Makefile.am: rearrange a bunch of files and get "make
588 check" framework set up
590 2003-08-31 Havoc Pennington <hp@pobox.com>
592 * fix build with --disable-tests
594 2003-08-30 Havoc Pennington <hp@pobox.com>
596 * dbus/dbus-connection.c: purge DBusMessageHandler
598 * dbus/dbus-message-handler.c: remove DBusMessageHandler, just
599 use callbacks everywhere
601 2003-08-30 Havoc Pennington <hp@pobox.com>
603 * test/data/valid-config-files/system.d/test.conf: change to
604 root for the user so warnings don't get printed
606 * dbus/dbus-message.c: add dbus_message_get_path,
607 dbus_message_set_path
609 * dbus/dbus-object-tree.c (do_test_dispatch): add test of
610 dispatching to a path
612 * dbus/dbus-string.c (_dbus_string_validate_path): add
614 * dbus/dbus-marshal.c (_dbus_demarshal_object_path): implement
615 (_dbus_marshal_object_path): implement
617 * dbus/dbus-protocol.h (DBUS_HEADER_FIELD_PATH): new header field
618 to contain the path to the target object
619 (DBUS_HEADER_FIELD_SENDER_SERVICE): rename
620 DBUS_HEADER_FIELD_SENDER to explicitly say it's the sender service
622 2003-08-30 Havoc Pennington <hp@pobox.com>
624 * dbus/dbus-object-tree.c: write tests and fix the discovered bugs
626 2003-08-29 Havoc Pennington <hp@pobox.com>
628 * dbus/dbus-object-tree.c: modify to allow overlapping paths to be
630 (struct DBusObjectSubtree): shrink this
631 a lot, since we may have a lot of them
632 (_dbus_object_tree_free_all_unlocked): implement
633 (_dbus_object_tree_dispatch_and_unlock): implement
635 2003-08-29 Havoc Pennington <hp@pobox.com>
637 * dbus/dbus-internals.h: fix _DBUS_N_GLOBAL_LOCKS
639 2003-08-28 Havoc Pennington <hp@pobox.com>
643 * dbus/dbus-connection.c: port to no ObjectID, create a
644 DBusObjectTree, rename ObjectTree to ObjectPath in public API
646 * dbus/dbus-connection.h (struct DBusObjectTreeVTable): delete
647 everything except UnregisterFunction and MessageFunction
649 * dbus/dbus-marshal.c: port away from DBusObjectID,
650 add DBUS_TYPE_OBJECT_PATH
652 * dbus/dbus-object-registry.[hc], dbus/dbus-object.[hc],
653 dbus/dbus-objectid.[hc]: remove these, we are moving to
654 path-based object IDs
656 2003-08-25 Havoc Pennington <hp@pobox.com>
658 Just noticed that dbus_message_test is hosed, I wonder when I
659 broke that. I thought make check was passing earlier...
661 * dbus/dbus-object-tree.c: add new "object tree" to match DCOP
662 container tree, will replace most of dbus-object-registry
664 * dbus/dbus-string.c (_dbus_string_append_printf_valist): fix C99
667 2003-08-19 Havoc Pennington <hp@pobox.com>
669 * dbus/dbus-message.c (decode_string_field): support FIELD_SENDER
670 (dbus_message_is_error): fix this function
672 * bus/dbus-daemon-1.1: clarify logic on when <deny>/<allow> rules
675 * bus/policy.c (bus_client_policy_check_can_receive): fix code to
676 reflect clarified man page
677 (bus_client_policy_check_can_send): ditto
679 * bus/session.conf.in: fixup
681 * bus/system.conf.in: fixup
683 2003-08-18 Havoc Pennington <hp@redhat.com>
685 * dbus/dbus-hash.c (_dbus_hash_table_insert_two_strings): fix
687 * dbus/dbus-message.c (_dbus_message_loader_queue_messages): fix
688 dumb bug created earlier (wrong order of args to
689 decode_header_data())
691 * tools/dbus-send.c: port
693 * tools/dbus-print-message.c (print_message): port
695 * test/data/*messages: port all messages over
697 * dbus/dbus-message-builder.c: support including
700 * bus/driver.c: port over
702 * bus/dispatch.c: port over to new stuff
704 * dbus/dbus-connection.c (_dbus_connection_new_for_transport):
705 rename disconnect signal to "Disconnected"
707 2003-08-17 Havoc Pennington <hp@pobox.com>
709 This doesn't compile yet, but syncing up so I can hack on it from
710 work. What are branches for if not broken code? ;-)
712 * dbus/dbus-protocol.h: remove DBUS_HEADER_FIELD_NAME, add
713 DBUS_HEADER_FIELD_INTERFACE, DBUS_HEADER_FIELD_MEMBER,
714 DBUS_HEADER_FIELD_ERROR_NAME
716 * dbus/dbus-hash.c: Introduce DBUS_HASH_TWO_STRINGS as hack to use
717 for the interface+member pairs
718 (string_hash): change to use g_str_hash algorithm
719 (find_direct_function, find_string_function): refactor these to
722 * dbus/dbus-message.c: port all of this over to support
723 interface/member fields instead of name field
725 * dbus/dbus-object-registry.c: port over
727 * dbus/dbus-string.c (_dbus_string_validate_interface): rename
728 from _dbus_string_validate_name
730 * bus/dbus-daemon-1.1: change file format for the
731 <deny>/<allow> stuff to match new message naming scheme
733 * bus/policy.c: port over
735 * bus/config-parser.c: parse new format
737 2003-08-16 Havoc Pennington <hp@pobox.com>
739 * dbus/dbus-object-registry.c (add_and_remove_objects): remove
742 * glib/dbus-gproxy.c: some hacking
744 2003-08-15 Havoc Pennington <hp@redhat.com>
746 * dbus/dbus-pending-call.c (dbus_pending_call_block): implement
748 * dbus/dbus-connection.c
749 (dbus_connection_send_with_reply_and_block): factor out internals;
750 change to convert any error replies to DBusError instead of
751 returning them as a message
753 2003-08-15 Havoc Pennington <hp@pobox.com>
755 * dbus/dbus-connection.c,
756 dbus/dbus-pending-call.c: Finish the pending call stuff
758 2003-08-14 Havoc Pennington <hp@redhat.com>
760 * dbus/dbus-pending-call.c: start on new object that will replace
761 DBusMessageHandler and ReplyHandlerData for tracking outstanding
764 * dbus/dbus-gproxy.c: start on proxy object used to communicate
765 with remote interfaces
767 * dbus/dbus-gidl.c: do the boring boilerplate in here
769 2003-08-12 Havoc Pennington <hp@pobox.com>
771 * bus/dispatch.c (bus_dispatch): make this return proper
772 DBusHandlerResult to avoid DBUS_ERROR_UNKNOWN_METHOD
774 * dbus/dbus-errors.c (dbus_set_error): use
775 _dbus_string_append_printf_valist
777 * dbus/dbus-string.c (_dbus_string_append_printf_valist)
778 (_dbus_string_append_printf): new
780 * dbus/dbus-errors.h (DBUS_ERROR_UNKNOWN_MESSAGE): change to
783 * dbus/dbus-connection.c (dbus_connection_dispatch): handle
784 DBUS_HANDLER_RESULT_NEED_MEMORY; send default error reply if a
785 message is unhandled.
787 2003-08-11 Havoc Pennington <hp@pobox.com>
789 * bus/test.c (client_disconnect_handler): change to return
790 HANDLED (would have been REMOVE_MESSAGE)
792 * dbus/dbus-object.h (enum DBusHandlerResult): rename to
793 HANDLED/NOT_YET_HANDLED instead of
794 REMOVE_MESSAGE/ALLOW_MORE_HANDLERS to make it clearer how it
797 2003-08-10 Havoc Pennington <hp@pobox.com>
799 * tools/dbus-send.c (main): add --type argument, for now
800 supporting only method_call and signal types.
802 * tools/dbus-print-message.c: print message type
804 * dbus/dbus-connection.c (_dbus_connection_new_for_transport):
805 init connection->objects
807 * doc/dbus-specification.sgml: fix sgml
809 * bus/*.c: port over to object-instance API changes
811 * test/test-service.c: ditto
813 * dbus/dbus-message.c (dbus_message_create_header): allow #NULL
814 name, we will have to fix up the rest of the code to also handle
816 (dbus_message_new): generic message-creation call
817 (set_string_field): allow appending name field
819 2003-08-06 Havoc Pennington <hp@pobox.com>
821 * dbus/dbus-object-registry.c: implement signal connection
824 * dbus/dbus-connection.c (_dbus_connection_unref_unlocked): new
826 * dbus/dbus-internals.c (_dbus_memdup): new function
828 2003-08-02 Havoc Pennington <hp@pobox.com>
830 * dbus/dbus-message.c (dbus_message_get_no_reply)
831 (dbus_message_set_no_reply): add these and remove
832 set_is_error/get_is_error
834 * dbus/dbus-protocol.h, doc/dbus-specification.sgml:
835 remove the ERROR flag, since there's now an ERROR type
837 2003-08-01 Havoc Pennington <hp@pobox.com>
839 * dbus/dbus-object-registry.c (_dbus_object_registry_handle_and_unlock):
842 * dbus/dbus-message.c (dbus_message_get_type): new function
844 * doc/dbus-specification.sgml: add "type" byte to messages
846 2003-08-01 Havoc Pennington <hp@pobox.com>
848 * dbus/dbus-protocol.h (DBUS_MESSAGE_TYPE_*): introduce
849 a message type enum to distinguish kinds of message
850 (DBUS_HEADER_FLAG_NO_REPLY_EXPECTED): flag for a message
851 that need not be replied to
853 2003-08-01 Havoc Pennington <hp@pobox.com>
855 * dbus/dbus-marshal.c: adapt to DBusObjectID changes
856 (unpack_8_octets): fix no-64-bit-int bug
858 * dbus/dbus-object-registry.c (validate_id): validate the
859 connection ID bits, not just the instance ID.
861 * dbus/dbus-connection.c (_dbus_connection_init_id): initialize
862 the connection-global 33 bits of the object ID
864 * dbus/dbus-object-registry.c (info_from_entry): fill in
865 object ID in the new way
867 * dbus/dbus-objectid.h: rather than high/low bits, specifically
868 define server/client/instance bits.
870 2003-07-30 Havoc Pennington <hp@pobox.com>
872 * dbus/dbus-connection.c (dbus_connection_register_object): fix
875 2003-07-13 Havoc Pennington <hp@pobox.com>
877 * dbus/dbus-object.h (struct DBusObjectVTable): add padding
878 fields to DBusObjectVTable and DBusObjectInfo
880 2003-07-12 Havoc Pennington <hp@pobox.com>
882 * dbus/dbus-object-registry.c: implement unit test,
883 fix bugs discovered in process
885 * dbus/dbus-connection.c: remove handler_table and
886 register_handler(), add DBusObjectRegistry usage
888 * dbus/dbus-objectid.c (dbus_object_id_is_null)
889 (dbus_object_id_set_null): new functions
891 2003-07-08 Havoc Pennington <hp@pobox.com>
893 * dbus/dbus-object.c: implement some of this
895 * dbus/dbus-object-registry.c
896 (_dbus_object_registry_add_and_unlock): fill in the object_id out
898 (_dbus_object_registry_new): handle OOM
900 2003-07-08 Havoc Pennington <hp@pobox.com>
902 * dbus/dbus-object.h: sketch out an API for registering objects
903 with a connection, that allows us to use as little as 24 bytes
904 per object and lets application code represent an object in
907 * dbus/dbus-object-registry.c: implement the hard bits of the
908 DBusConnection aspect of object API. Not yet wired up.
910 2003-07-06 Havoc Pennington <hp@pobox.com>
912 * dbus/dbus-marshal.c (_dbus_marshal_set_object_id): new function
913 (_dbus_marshal_object_id): new
914 (_dbus_demarshal_object_id): new
915 (_dbus_marshal_get_arg_end_pos): support object ID type, and
916 consolidate identical switch cases. Don't conditionalize handling
917 of DBUS_TYPE_UINT64, need to handle the type always.
918 (_dbus_marshal_validate_arg): consolidate identical cases, and
919 handle DBUS_TYPE_OBJECT_ID
921 * dbus/dbus-objectid.c: new file with DBusObjectID data type.
923 * dbus/dbus-protocol.h: add DBUS_TYPE_OBJECT_ID
925 2003-09-28 Havoc Pennington <hp@pobox.com>
929 2003-09-28 Havoc Pennington <hp@pobox.com>
931 * doc/Makefile.am (dbus-specification.html): testing a funky hack
932 to work with Debian db2html
934 2003-09-28 Havoc Pennington <hp@pobox.com>
938 * doc/Makefile.am (dbus-test-plan.html): accept nonexistence of
939 stylesheet-images for benefit of Debian
941 Change back to using filesystem-linked sockets for the system
942 bus, so only root can create the default system bus address.
944 * bus/system.conf.in: change to use
945 DBUS_SYSTEM_BUS_DEFAULT_ADDRESS
947 * dbus/Makefile.am (INCLUDES): remove DBUS_SYSTEM_BUS_PATH define
950 * configure.in: define DBUS_SYSTEM_BUS_DEFAULT_ADDRESS
951 here, and AC_DEFINE DBUS_SYSTEM_PATH
953 2003-08-09 Anders Carlsson <andersca@codefactory.se>
959 2003-08-09 Anders Carlsson <andersca@codefactory.se>
961 * doc/dbus-specification.sgml:
962 Add activation reply values.
964 2003-08-05 Havoc Pennington <hp@redhat.com>
968 2003-08-05 Anders Carlsson <andersca@codefactory.se>
970 * glib/dbus-gmain.c: (watch_fd_new), (watch_fd_ref),
971 (watch_fd_unref), (dbus_gsource_check), (dbus_gsource_dispatch),
972 (add_watch), (remove_watch), (create_source):
973 Refcount fds, fixes some reentrancy issues.
975 2003-07-30 Havoc Pennington <hp@redhat.com>
977 * dbus/dbus-bus.c (init_connections_unlocked): fix default system
978 bus address to be abstract if we have abstract sockets
982 2003-07-28 Havoc Pennington <hp@redhat.com>
984 * bus/messagebus.in: fix to avoid processname/servicename
985 confusion, from Michael Kearey
986 https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=100965
988 2003-07-23 Havoc Pennington <hp@pobox.com>
990 * dbus/dbus-message.c (dbus_message_iter_get_named):
991 fix from Andy Hanton to remove broken "+1"
993 2003-07-16 Havoc Pennington <hp@pobox.com>
995 * tools/dbus-launch.c (babysit): close stdout/stderr in the
996 babysitter process, as suggested by Thomas Leonard, so
997 an "eval `dbus-launch --exit-with-session`" will actually
1000 2003-07-16 Havoc Pennington <hp@pobox.com>
1002 * configure.in: print out EXPANDED_* variables in the summary at
1003 the end; clean up the code that computes EXPANDED_ variables and
1004 get the ones using exec_prefix right. Should make things work
1005 when you build without --prefix
1007 2003-06-29 Havoc Pennington <hp@pobox.com>
1009 * mono/Test.cs (class Test): fire up a main loop and run it
1011 * mono/DBus.cs (DBus): don't g_thread_init since it can only be
1012 done once, the app has to do it
1014 2003-06-26 Havoc Pennington <hp@pobox.com>
1016 * mono/Connection.cs: set up connection with the glib main loop
1018 2003-07-01 Havoc Pennington <hp@redhat.com>
1020 * doc/dbus-specification.sgml: clarify the format of a type code,
1021 change suggested by Jim Blandy
1023 2003-06-29 Miloslav Trmac <mitr@volny.cz>
1026 * tools/Makefile.am: Don't assume srcdir == builddir.
1028 * dbus/dbus-memory.c (dbus_realloc): Don't check guards after shrinking
1029 the allocated block.
1030 (_dbus_memory_test): New function.
1031 * dbus/dbus-test.h: Add _dbus_memory_test ().
1032 * dbus/dbus-test.c (dbus_internal_do_not_use_run_tests): Call it.
1034 * dbus/dbus-message.c (decode_header_data): Use %.4s instead
1036 (dbus_message_new): Remove obsolete @todo.
1038 * dbus/dbus-marshal.c (_dbus_marshal_set_int64)
1039 (_dbus_marshal_set_uint64): Fix comment.
1041 * dbus/dbus-message.c (append_int_field, append_uint_field): Don't
1042 hardcode FIELD_REPLY_SERIAL.
1044 * dbus/dbus-mainloop.c (_dbus_loop_remove_watch)
1045 (_dbus_loop_remove_timeout): Cast function pointers to (void *) for %p
1047 * configure.in: Add -D_POSIX_C_SOURCE=199309L -DBSD_SOURCE to CFLAGS
1048 and disable DBUS_USE_ATOMIC_INT_486 when --enable-ansi is used
1050 2003-06-24 Havoc Pennington <hp@pobox.com>
1052 * mono/*.cs: Use IntPtr.Zero instead of ((IntPtr) 0)
1054 2003-06-23 Anders Carlsson <andersca@codefactory.se>
1060 * gcj/TestMessage.java: (TestMessage), (TestMessage.main):
1061 * gcj/org/.cvsignore:
1062 * gcj/org/Makefile.am:
1063 * gcj/org/freedesktop/.cvsignore:
1064 * gcj/org/freedesktop/Makefile.am:
1065 * gcj/org/freedesktop/dbus/.cvsignore:
1066 * gcj/org/freedesktop/dbus/Makefile.am:
1067 * gcj/org/freedesktop/dbus/Message.java: (Message),
1069 * gcj/org/freedesktop/dbus/natMessage.cc:
1070 Fix the build system.
1072 2003-06-22 Havoc Pennington <hp@pobox.com>
1074 * mono/Connection.cs: add more bindings
1076 * dbus/dbus-threads.c (dbus_threads_init): allow calling this
1079 2003-06-22 Havoc Pennington <hp@pobox.com>
1081 * mono/Connection.cs, mono/DBus.cs, mono/Error.cs:
1082 Start wrapping more stuff.
1084 2003-06-22 Havoc Pennington <hp@pobox.com>
1086 * mono/Message.cs: implement Message.Wrap() that ensures we only
1087 have a single C# wrapper per DBusMessage, assuming it works which
1088 it probably doesn't.
1090 * dbus/dbus-message.c (dbus_message_allocate_data_slot): new
1091 (dbus_message_free_data_slot): new
1092 (dbus_message_set_data): new
1093 (dbus_message_get_data): new
1095 2003-06-22 Havoc Pennington <hp@pobox.com>
1097 * dbus/dbus-dataslot.c (_dbus_data_slot_allocator_unref)
1098 (_dbus_data_slot_allocator_alloc): rework these to keep a
1099 reference count on each slot and automatically manage a global
1100 slot ID variable passed in by address
1102 * bus/bus.c: convert to new dataslot API
1104 * dbus/dbus-bus.c: convert to new dataslot API
1106 * dbus/dbus-connection.c: convert to new dataslot API
1108 * dbus/dbus-server.c: convert to new dataslot API
1110 * glib/dbus-gmain.c: ditto
1114 * bus/connection.c: ditto
1116 2003-06-22 Anders Carlsson <andersca@codefactory.se>
1118 * configure.in: Add AM_PROG_GCJ and move AM_PROG_LIBTOOL
1119 after the gcj checks so that the correct configuration tags
1120 will be added to libtool.
1122 * dbus-glib-1.pc.in: No need to specify any includes since
1123 dbus-1.pc.in has those.
1125 2003-06-22 Havoc Pennington <hp@pobox.com>
1127 * mono/*, gcj/*, configure.in, Makefile.am:
1128 Check in makefiles and subdirs for mono and gcj bindings.
1129 Neither binding actually exists, just trying to get through
1130 all the build and other boring bits.
1132 2003-06-21 Philip Blundell <philb@gnu.org>
1134 * tools/dbus-monitor.1: Updated.
1136 * tools/dbus-send.1: Likewise.
1138 2003-06-20 Anders Carlsson <andersca@codefactory.se>
1140 * dbus/dbus-transport-unix.c (unix_handle_watch): Check
1141 for hangup and error after checking read so we won't discard
1142 pending data if both hangup and read are set.
1144 2003-06-19 Philip Blundell <philb@gnu.org>
1146 * tools/dbus-print-message.c (print_message): Handle BOOLEAN.
1148 * tools/dbus-send.c: Accept both --system and --session.
1150 * tools/dbus-monitor.c: Same here.
1152 2003-06-19 Anders Carlsson <andersca@codefactory.se>
1154 * glib/dbus-glib.h: Fix so that dbus-glib.h can be used
1155 from C++ (Patch by Miloslav Trmac).
1157 2003-06-15 Joe Shaw <joe@assbarn.com>
1159 * configure.in: Check for socklen_t.
1161 * dbus/dbus-sysdeps.c: Define socklen_t if it's not defined.
1163 * test/test-segfault.c: Add #include <sys/time.h>
1165 * tools/Makefile.am: Add DBUS_X_CFLAGS to the INCLUDES since
1166 dbus-launch needs it.
1168 2003-06-09 Havoc Pennington <hp@redhat.com>
1170 * dbus/dbus-sysdeps.c (_dbus_listen_unix_socket): don't use
1171 SUN_LEN, it breaks abstract socket usage
1173 * dbus/dbus-internals.c (_dbus_verbose_real): only print PID at
1176 2003-06-04 Havoc Pennington <hp@pobox.com>
1178 * dbus/dbus-server.c (dbus_server_listen): allow abstract sockets
1179 using unix:abstract=/foo, and when listening in a tmpdir
1180 i.e. unix:tmpdir=/tmp, always use abstract sockets if we can.
1182 * dbus/dbus-transport.c (_dbus_transport_open): support
1185 * dbus/dbus-server-unix.c (_dbus_server_new_for_domain_socket):
1186 support abstract sockets
1188 * dbus/dbus-transport-unix.c
1189 (_dbus_transport_new_for_domain_socket): support abstract sockets
1191 * dbus/dbus-sysdeps.c (_dbus_connect_unix_socket): add "abstract"
1192 toggle as an argument, implement abstract namespace support
1193 (_dbus_listen_unix_socket): ditto
1195 * configure.in: add --enable-abstract-sockets and implement
1196 a configure check for autodetection of the right value.
1198 2003-06-01 Havoc Pennington <hp@pobox.com>
1200 * tools/dbus-cleanup-sockets.c: add utility to clean up sockets
1201 in /tmp (though on Linux this will end up being useless,
1202 when we add abstract namespace support)
1204 * configure.in: define DBUS_SESSION_SOCKET_DIR in addition to
1207 2003-05-28 Colin Walters <walters@verbum.org>
1209 * tools/dbus-monitor.c (main): Fix silly typo (s/--session/--system/).
1211 2003-05-18 Anders Carlsson <andersca@codefactory.se>
1213 * dbus/dbus-message.c (dbus_message_new): Remove @todo.
1215 2003-05-17 Colin Walters <walters@gnu.org>
1217 * tools/dbus-send.c: Don't exit with an error code if --help was
1218 passed. Default to using the session bus instead of the system
1221 * tools/dbus-launch.c: Ditto.
1223 * tools/dbus-monitor.c: Ditto.
1225 * tools/dbus-send.1: Update with new arguments.
1227 * tools/dbus-launch.c: Emit code to export variables. New
1228 arguments -s and -c to specify shell syntax, and a bit of code to
1229 autodetect syntax. Also, allow specifying a program to run.
1231 * tools/dbus-launch.1: Update with new arguments.
1233 * tools/dbus-send.1: Ditto.
1235 * tools/dbus-monitor.1: Ditto.
1237 2003-05-17 Havoc Pennington <hp@pobox.com>
1239 * bus/config-parser.c (merge_included): merge in policies from
1240 child configuration file.
1242 * bus/policy.c (bus_policy_merge): function to merge two policies
1245 2003-05-16 Havoc Pennington <hp@redhat.com>
1247 * dbus/dbus-connection.c: disable verbose lock spew
1249 * tools/dbus-send.c: add --print-reply command line option
1251 * tools/dbus-print-message.h (print_message): new util function
1252 shared by dbus-send and dbus-monitor
1254 * tools/dbus-monitor.c (handler_func): exit on disconnect
1256 * dbus/dbus-transport-unix.c (do_reading): if the transport is
1257 disconnected, don't try to use the read_watch
1259 * dbus/dbus-watch.c (dbus_watch_get_enabled): assert watch != NULL
1260 so we can find this bug more easily
1262 2003-05-16 Havoc Pennington <hp@redhat.com>
1264 * bus/policy.c (free_rule_list_func): avoid a crash when passed
1265 NULL as DBusHashTable is annoyingly likely to do.
1267 2003-05-16 Colin Walters <walters@verbum.org>
1269 * tools/dbus-monitor.c: Add --session argument and usage()
1272 * tools/dbus-monitor.1: Update with new --session arg.
1274 * bus/Makefile.am (install-data-hook): Create
1275 $(libdir)/dbus-1.0/services so that the session bus is happy.
1277 2003-05-15 Havoc Pennington <hp@redhat.com>
1279 * dbus/dbus-sysdeps.c (_dbus_atomic_dec, _dbus_atomic_inc): work
1280 on non-x86. ifdef's are evil.
1282 2003-05-15 Havoc Pennington <hp@redhat.com>
1284 * configure.in: 0.11
1288 * bus/Makefile.am (initddir): apparently we are supposed to put
1289 init scripts in /etc/rc.d/init.d not /etc/init.d
1291 * bus/Makefile.am: remove the "you must --enable-tests to make
1292 check" as it broke distcheck
1294 * bus/Makefile.am (install-data-hook): create /etc/dbus-1/system.d
1296 2003-05-13 James Willcox <jwillcox@gnome.org>
1299 * bus/activation.c: (bus_activation_service_created),
1300 (bus_activation_activate_service):
1301 * bus/driver.c: (bus_driver_send_service_deleted),
1302 (bus_driver_send_service_created), (bus_driver_send_service_lost),
1303 (bus_driver_send_service_acquired),
1304 (bus_driver_send_welcome_message),
1305 (bus_driver_handle_list_services):
1306 * bus/session.conf.in:
1307 * dbus/dbus-bus.c: (dbus_bus_acquire_service),
1308 (dbus_bus_service_exists), (dbus_bus_activate_service):
1311 Add some convenience API which lets you activate a service, and did a
1312 bunch of s/0/DBUS_TYPE_INVALID/ in calls to dbus_message_append_args()
1313 and dbus_message_get_args()
1315 2003-05-11 Havoc Pennington <hp@pobox.com>
1317 * dbus/dbus-marshal.c (_dbus_marshal_validate_arg): fix to avoid
1318 calling _dbus_marshal_validate_arg() for every byte in a byte
1321 * dbus/dbus-message-handler.c: use atomic reference counting to
1322 reduce number of locks slightly; the global lock in here sucks
1324 * dbus/dbus-connection.c
1325 (_dbus_connection_update_dispatch_status_and_unlock): variant of
1326 update_dispatch_status that can be called with lock held; then use
1327 in a couple places to reduce locking/unlocking
1328 (dbus_connection_send): hold the lock over the whole function
1329 instead of acquiring it twice.
1331 * dbus/dbus-timeout.c (_dbus_timeout_new): handle OOM
1333 * bus/connection.c (bus_connections_setup_connection): fix access
1334 to already-freed memory.
1336 * dbus/dbus-connection.c: keep a little cache of linked list
1337 nodes, to avoid using the global linked list alloc lock in the
1338 normal send-message case. Instead we just use the connection lock
1339 that we already have to take.
1341 * dbus/dbus-list.c (_dbus_list_find_last): new function
1343 * dbus/dbus-sysdeps.c (_dbus_atomic_inc, _dbus_atomic_dec):
1344 change to use a struct for the atomic type; fix docs,
1345 they return value before increment, not after increment.
1347 * dbus/dbus-string.c (_dbus_string_append_4_aligned)
1348 (_dbus_string_append_8_aligned): new functions to try to
1349 microoptimize this operation.
1350 (reallocate_for_length): break this out of set_length(), to
1351 improve profile info, and also so we can consider inlining the
1354 * dbus/dbus-message.c (dbus_message_new_empty_header): init data
1355 strings with some preallocation, cuts down on our calls to realloc
1356 a fair bit. Though if we can get the "move entire string to empty
1357 string" optimization below to kick in here, it would be better.
1359 * dbus/dbus-string.c (_dbus_string_move): just call
1360 _dbus_string_move_len
1361 (_dbus_string_move_len): add a special case for moving
1362 an entire string into an empty string; we can just
1363 swap the string data instead of doing any reallocs.
1364 (_dbus_string_init_preallocated): new function
1366 2003-05-11 Havoc Pennington <hp@pobox.com>
1368 Write a "test-profile" that does echo client-server with threads;
1369 profile reveals lock contention, memcpy/realloc of buffers, and
1370 UTF-8 validation as hot spots. 20% of lock contention eliminated
1371 with dbus_atomic_inc/dec implementation on x86. Much remaining
1372 contention is global mempool locks for GList and DBusList.
1374 * dbus/dbus-sysdeps.c (_dbus_atomic_inc, _dbus_atomic_dec): add
1377 * dbus/dbus-connection.c (struct DBusConnection): use
1378 dbus_atomic_t for the reference count
1380 * dbus/dbus-message.c (struct DBusMessage): declare
1381 dbus_atomic_t values as volatile
1383 * configure.in: code to detect ability to use atomic integer
1384 operations in assembly, from GLib patch
1386 * dbus/dbus-internals.c (_dbus_verbose_real): call getpid every
1387 time, tired of it being wrong in threads and forked processes
1389 * glib/test-profile.c: a little program to bounce messages back
1390 and forth between threads and eat CPU
1392 * dbus/dbus-connection.c: add debug spew macros for debugging
1393 thread locks; include config.h at top; fix deadlock in
1394 dbus_connection_flush()
1396 2003-05-08 Havoc Pennington <hp@pobox.com>
1398 * dbus/dbus-spawn.c: s/_exit/exit/ because it was keeping gcov
1399 data from getting written, and there wasn't a good reason to
1402 * test/decode-gcov.c (mark_inside_dbus_build_tests): don't count
1403 dbus_verbose lines in test coverage
1404 (main): add list of functions sorted by # of untested blocks
1405 to the coverage report
1407 * dbus/dbus-mempool.c: put some test-only code in DBUS_BUILD_TESTS
1409 * dbus/dbus-marshal.c (_dbus_marshal_test): extend test coverage
1411 * dbus/dbus-message-handler.c (_dbus_message_handler_test):
1412 extend test coverage
1414 * test/data/auth/cancel.auth-script: test canceling an
1417 * dbus/Makefile.am: remove dbus-server-debug.[hc] for now, as they
1418 aren't used. in CVS history if we end up needing them.
1420 2003-05-04 Havoc Pennington <hp@pobox.com>
1422 * dbus/dbus-message-handler.c (_dbus_message_handler_test): add
1425 * dbus/dbus-marshal.c (_dbus_demarshal_string_array): fix this
1426 function, which assumed length was in # of strings, not bytes
1428 * dbus/dbus-message.c (_dbus_message_test): add tests for some
1431 * dbus/dbus-connection.c
1432 (_dbus_connection_queue_received_message): disable function for
1433 now, we are only using it in test mode
1435 * dbus/dbus-message.c (_dbus_message_loader_queue_messages):
1436 remove a mistaken FIXME
1438 2003-05-04 Havoc Pennington <hp@pobox.com>
1440 * dbus/dbus-connection.c (dbus_connection_preallocate_send):
1441 unlock mutex on successful return, patch from Anders Gustafsson
1443 2003-05-04 Havoc Pennington <hp@pobox.com>
1445 * dbus-glib-1.pc.in (Requires): fix dependencies, from
1448 2003-05-04 Havoc Pennington <hp@pobox.com>
1450 * tools/dbus-launch.c: implement
1452 * bus/main.c (main), bus/bus.c (bus_context_new):
1453 implement --print-pid and --fork
1455 2003-05-03 Havoc Pennington <hp@redhat.com>
1457 * dbus/dbus-address.c (dbus_parse_address): fix bug when a key in
1458 the address had no value, and add to test suite. Fix and
1459 regression test from Miloslav Trmac
1461 2003-05-03 Havoc Pennington <hp@pobox.com>
1463 * dbus/dbus-watch.c (dbus_watch_handle): warn and return if a
1464 watch is invalid when handled
1466 * tools/Makefile.am, tools/dbus-launch.c, tools/dbus-launch.1: add
1467 dbus-launch utility to launch the bus from a shell script. Didn't
1468 actually implement dbus-launch yet, it's just a placeholder still.
1470 2003-05-03 Havoc Pennington <hp@pobox.com>
1472 * bus/Makefile.am, bus/dbus-daemon-1.1.in: man page for the
1473 daemon; also documents daemon config file, so replaces
1474 doc/config-file.txt. Corrected some stuff from config-file.txt in
1475 the process of moving it.
1477 2003-05-03 Havoc Pennington <hp@pobox.com>
1479 * tools/Makefile.am, tools/dbus-send.1, tools/dbus-monitor.1:
1482 2003-05-03 Colin Walters <walters@verbum.org>
1484 * dbus/dbus-sysdeps.c (fill_user_info): Test against
1485 DBUS_UID_UNSET to determine whether to do a uid lookup or not.
1487 * Makefile.am: Update to use new .pc versioning scheme.
1489 2003-05-02 Havoc Pennington <hp@redhat.com>
1491 * bus/system.conf.in: allow send/receive to/from message bus
1494 2003-04-30 Havoc Pennington <hp@redhat.com>
1496 * configure.in: print a note when building with unit tests and
1499 2003-04-30 Havoc Pennington <hp@redhat.com>
1501 * Makefile.am: add a check-local that complains if you didn't
1502 configure with --enable-tests
1504 2003-04-29 Havoc Pennington <hp@redhat.com>
1506 * glib/dbus-gmain.c: docs cleanups
1508 * dbus/dbus-types.h: add docs on int64 types
1510 * dbus/dbus-memory.c: fix docs to avoid putting private API in
1511 public API docs section
1513 2003-04-29 Havoc Pennington <hp@redhat.com>
1515 * dbus-1.pc.in, dbus-glib-1.pc.in: rename these from
1516 dbus-1.0.pc.in, dbus-glib-1.0.pc.in. As these change with the
1517 parallel install API version, not with the D-BUS package version.
1519 * HACKING: move some of README over here
1521 * README: updates, and document API/ABI policy
1523 * configure.in: reindentation
1525 2003-04-29 Havoc Pennington <hp@redhat.com>
1527 * dbus/dbus.h: add "you have to define DBUS_API_SUBJECT_TO_CHANGE
1528 to use this library" to be sure people have the right
1531 2003-04-28 Havoc Pennington <hp@redhat.com>
1533 * configure.in: add --enable-docs which by default is auto yes if
1534 doxygen and db2html found, no otherwise; but can be forced on/off
1536 * doc/Makefile.am: conditionalize whether to build docs on
1539 2003-04-28 Havoc Pennington <hp@redhat.com>
1541 * configure.in: 0.10
1545 * bus/system.conf.in: add <includedir>system.d</includedir>
1547 * dbus/dbus-userdb.c (_dbus_user_database_lookup): fix bug when
1548 username was provided but not uid
1550 * bus/config-parser.c (struct BusConfigParser): keep track of
1551 whether the parser is toplevel or was included; change some
1552 of the error handling if it's included.
1554 2003-04-27 Havoc Pennington <hp@pobox.com>
1558 * dbus/dbus-transport.c (_dbus_transport_get_dispatch_status):
1559 report correct status if we finish processing authentication
1560 inside this function.
1562 * bus/activation.c (try_send_activation_failure): use
1563 bus_transaction_send_error_reply
1565 * bus/connection.c (bus_connection_get_groups): return an error
1566 explaining the problem
1568 * bus/bus.c (bus_context_check_security_policy): implement
1569 restriction here that inactive connections can only send the
1570 hello message. Also, allow bus driver to send anything to
1573 * bus/connection.c (bus_connection_complete): create the
1574 BusClientPolicy here instead of on-demand.
1575 (bus_connection_get_policy): don't return an error
1577 * dbus/dbus-message.c (dbus_message_new_error_reply): allow NULL
1578 sender field in message being replied to
1580 * bus/bus.c (bus_context_check_security_policy): fix silly typo
1581 causing it to return FALSE always
1583 * bus/policy.c (bus_client_policy_check_can_send): fix bug where
1584 we checked sender rather than destination
1586 2003-04-25 Havoc Pennington <hp@redhat.com>
1588 test suite is slightly hosed at the moment, will fix soon
1590 * bus/connection.c (bus_connections_expire_incomplete): fix to
1591 properly disable the timeout when required
1592 (bus_connection_set_name): check whether we can remove incomplete
1593 connections timeout after we complete each connection.
1595 * dbus/dbus-mainloop.c (check_timeout): fix this up a bit,
1596 probably still broken.
1598 * bus/services.c (bus_registry_acquire_service): implement max
1599 number of services owned, and honor allow/deny rules on which
1600 services a connection can own.
1602 * bus/connection.c (bus_connection_get_policy): report errors here
1604 * bus/activation.c: implement limit on number of pending
1607 2003-04-25 Havoc Pennington <hp@redhat.com>
1609 * dbus/dbus-transport.c (_dbus_transport_get_unix_user): fix bug
1610 where we used >= 0 instead of != DBUS_UID_UNSET.
1612 2003-04-25 Havoc Pennington <hp@redhat.com>
1614 * glib/dbus-gmain.c (remove_watch): fix for a crash when watches
1615 were toggled without add/remove, fix from Anders Gustafsson
1617 2003-04-24 Havoc Pennington <hp@redhat.com>
1619 * test/data/valid-config-files/basic.conf: add <limit> tags to
1622 * bus/config-parser.h, bus/config-parser.c, bus/bus.c: Implement
1623 <limit> tag in configuration file.
1625 2003-04-24 Havoc Pennington <hp@redhat.com>
1627 * bus/dispatch.c: somehow missed some name_is
1629 * dbus/dbus-timeout.c (_dbus_timeout_set_enabled)
1630 (_dbus_timeout_set_interval): new
1632 * bus/connection.c (bus_connections_setup_connection): record time
1633 when each connection is first set up, and expire them after the
1634 auth timeout passes.
1636 2003-04-24 Havoc Pennington <hp@redhat.com>
1638 * dbus/dbus-message.c (dbus_message_name_is): rename
1639 (dbus_message_service_is): rename
1640 (dbus_message_sender_is): rename
1641 (dbus_message_get_service): rename
1643 2003-04-24 Havoc Pennington <hp@redhat.com>
1645 * configure.in: add --enable-checks
1647 * dbus/dbus-message.c (dbus_message_new): reverse name/service arguments
1649 * dbus/dbus-connection.c (dbus_connection_preallocate_send): fix
1650 to use thread locks.
1651 (_dbus_connection_handler_destroyed_locked): move some private
1652 functions into proper docs group
1654 * dbus/dbus-internals.h: add _dbus_return_if_fail,
1655 _dbus_return_val_if_fail
1657 Throughout: use dbus_return_if_fail
1659 2003-04-23 James Willcox <jwillcox@gnome.org>
1662 * glib/dbus-gmain.c: (add_timeout), (wakeup_main), (create_source),
1663 (dbus_connection_setup_with_g_main),
1664 (dbus_server_setup_with_g_main):
1665 * glib/test-dbus-glib.c: (main):
1666 * glib/test-thread-client.c: (main):
1667 * glib/test-thread-server.c: (new_connection_callback), (main):
1668 * tools/dbus-monitor.c: (main):
1670 Added a GMainContext argument to dbus_connection_setup_with_g_main()
1671 and dbus_server_setup_with_g_main().
1673 2003-04-20 Havoc Pennington <hp@pobox.com>
1675 * doc/dbus-specification.sgml: document the restrictions on
1676 message and service names
1678 2003-04-22 Havoc Pennington <hp@redhat.com>
1680 * dbus/dbus-message.c, dbus/dbus-marshal.c: add 64-bit integer
1681 support, and do some code cleanups to share more code and
1682 speed up array marshal/demarshal.
1684 * dbus-1.0.pc.in (Cflags): put libdir include file in cflags
1686 * configure.in: generate dbus-arch-deps.h
1688 * dbus/dbus-protocol.h (DBUS_TYPE_INT64, DBUS_TYPE_UINT64): add
1691 2003-04-22 Havoc Pennington <hp@redhat.com>
1693 * test/data/valid-messages/opposite-endian.message: fix test
1694 to use proper type for rply field
1696 * test/data/invalid-messages: add tests for below validation
1698 * dbus/dbus-message.c (decode_header_data): validate field types,
1699 and validate that named fields are valid names
1700 (decode_name_field): consider messages in the
1701 org.freedesktop.Local. namespace to be invalid.
1703 * dbus/dbus-string.c (_dbus_string_validate_name): new
1705 2003-04-19 Havoc Pennington <hp@pobox.com>
1707 * bus/driver.c (bus_driver_handle_hello): check limits and
1708 return an error if they are exceeded.
1710 * bus/connection.c: maintain separate lists of active and inactive
1711 connections, and a count of each. Maintain count of completed
1712 connections per user. Implement code to check connection limits.
1714 * dbus/dbus-list.c (_dbus_list_unlink): export
1716 * bus/bus.c (bus_context_check_security_policy): enforce a maximum
1717 number of bytes in the message queue for a connection
1719 2003-04-18 Havoc Pennington <hp@pobox.com>
1721 * dbus/dbus-auth.c (record_mechanisms): memleak fixes
1723 * dbus/dbus-sysdeps.c (_dbus_string_save_to_file): fix some
1726 * dbus/dbus-keyring.c (add_new_key): fix a memleak, and
1727 on realloc be sure to update the pointer in the keyring
1729 * dbus/dbus-string.c (_dbus_string_zero): compensate for align
1730 offset to avoid writing to unallocated memory
1732 * dbus/dbus-auth.c (process_rejected): return FALSE if we fail to
1733 try the next mechanism, so we properly handle OOM
1735 * dbus/dbus-keyring.c (_dbus_keyring_new_homedir): fix double-free
1737 (_dbus_keyring_new): fix OOM bug
1738 (_dbus_keyring_new_homedir): always set error; impose a maximum
1739 number of keys we'll load from the file, mostly to speed up the
1740 test suite and make its OOM checks more useful, but also for
1743 * dbus/dbus-auth.c (process_error_server): reject authentication
1744 if we get an error from the client
1745 (process_cancel): on cancel, send REJECTED, per the spec
1746 (process_error_client): send CANCEL if we get an error from the
1749 2003-04-18 Havoc Pennington <hp@pobox.com>
1751 * dbus/dbus-mainloop.c (_dbus_loop_iterate): fix UMR in verbose
1754 * dbus/dbus-auth.c (handle_client_data_cookie_sha1_mech): fix OOM
1757 * dbus/dbus-keyring.c (_dbus_keyring_new_homedir): only whine
1758 about DBUS_TEST_HOMEDIR once
1760 * bus/Makefile.am (TESTS_ENVIRONMENT): put DBUS_TEST_HOMEDIR in
1763 * bus/dispatch.c (bus_dispatch_sha1_test): actually load sha1
1764 config file so we test the right thing
1766 Throughout: assorted docs improvements
1768 2003-04-18 Havoc Pennington <hp@pobox.com>
1770 * glib/dbus-gmain.c: adapt to watch changes
1772 * bus/bus.c, bus/activation.c, etc.: adjust to watch changes
1774 * dbus/dbus-server.h: remove dbus_server_handle_watch
1776 * dbus/dbus-connection.h: remove dbus_connection_handle_watch
1778 * dbus/dbus-watch.c (dbus_watch_handle): change DBusWatch to work
1779 like DBusTimeout, so we don't need dbus_connection_handle_watch
1782 2003-04-17 Havoc Pennington <hp@redhat.com>
1784 * dbus/dbus-userdb.c, dbus/dbus-sysdeps.c: redo all the passwd
1785 database usage so it all goes via the DBusUserDatabase cache.
1787 2003-04-17 Havoc Pennington <hp@redhat.com>
1789 * dbus/dbus-mainloop.c (_dbus_loop_iterate): fix logic so that if
1790 there was an OOM watch we skipped, we always return TRUE so we
1791 iterate again to have a look at it again. Fixes test suite hang.
1792 Code rearrangement also lets us lose some memset and only iterate
1793 over callbacks once.
1795 * bus/driver.c (bus_driver_handle_message): sense of test for
1798 2003-04-16 Havoc Pennington <hp@pobox.com>
1800 * doc/dbus-specification.sgml: make spec say serials are unsigned
1802 * dbus/dbus-message.h: change message serials to unsigned
1804 * dbus/dbus-connection.c: adapt to message serials being unsigned
1806 2003-04-15 Havoc Pennington <hp@pobox.com>
1808 * bus/bus.c: create and keep around a shared DBusUserDatabase
1811 * bus/connection.c (bus_connection_get_groups): don't cache
1812 groups for user in the connection object, since user database
1813 object now does that.
1815 2003-04-16 Havoc Pennington <hp@redhat.com>
1817 * dbus/dbus-message.c (_dbus_message_add_size_counter): keep a
1818 list of size counters
1819 (_dbus_message_loader_putback_message_link): put back a popped link
1821 * dbus/dbus-connection.c
1822 (dbus_connection_set_max_live_messages_size): rename
1824 (dbus_connection_get_outgoing_size): get size of outgoing
1826 (_dbus_connection_set_connection_counter): remove this cruft
1828 2003-04-14 Havoc Pennington <hp@redhat.com>
1830 * dbus/dbus-userdb.c: user database abstraction, mostly to get
1831 caching, but at some point we might want to be able to use a
1834 * bus/dispatch.c (bus_dispatch_sha1_test): add a test that uses
1835 SHA1 conf file to test the sha1 auth mechanism, since the regular
1836 test always uses EXTERNAL when available.
1839 test/data/valid-config-files/debug-allow-all-sha1.conf.in:
1840 add conf file that requires use of sha1 auth
1842 2003-04-13 Havoc Pennington <hp@pobox.com>
1844 * tools/dbus-send.c, tools/dbus-monitor.c: two utility programs
1845 from Philip Blundell to send messages and monitor them.
1847 2003-04-13 Havoc Pennington <hp@pobox.com>
1849 * dbus/dbus-mainloop.c: fix some reentrancy issues by refcounting
1852 * test/data/valid-config-files/debug-allow-all.conf.in: allow all
1855 * dbus/dbus-transport.c (_dbus_transport_get_dispatch_status):
1856 fix to only recover unused bytes if we're already authenticated
1857 (_dbus_transport_get_is_authenticated): fix to still mark us
1858 authenticated if there are unused bytes.
1860 * bus/dispatch.c: implement security policy checking
1862 * bus/connection.c (bus_transaction_send_from_driver): new
1864 * bus/bus.c (bus_context_check_security_policy): new
1866 * bus/dispatch.c (send_service_nonexistent_error): delete this,
1867 now we just set the DBusError and it gets converted to an error
1870 * bus/connection.c (allow_user_function): enable code using actual
1871 data from the config file
1873 * bus/policy.c (list_allows_user): handle wildcard rules for
1874 user/group connection perms
1876 2003-04-13 Havoc Pennington <hp@pobox.com>
1878 * bus/config-parser.c: Load up the BusPolicy and BusPolicyRules
1880 * dbus/dbus-sysdeps.c (_dbus_get_user_id): new function
1882 * bus/policy.c (bus_policy_append_mandatory_rule)
1883 (bus_policy_append_default_rule, bus_policy_append_user_rule)
1884 (bus_policy_append_group_rule): new functions
1886 2003-04-12 Havoc Pennington <hp@pobox.com>
1888 * bus/config-parser.c (bus_config_parser_new): fix a memleak
1890 * dbus/dbus-sysdeps.c: change DBusCredentials to use longs for
1891 the pid/gid/uid, just for paranoia.
1893 * test/break-loader.c (randomly_do_n_things): find a byte
1894 containing a type code, and randomly change it to a different
1897 2003-04-12 Havoc Pennington <hp@pobox.com>
1899 * bus/policy.h: change BusPolicy to be the thing from the config
1900 file, and rename old BusPolicy to BusClientPolicy
1902 * bus/bus.c, bus/connection.c, bus/config-parser.c: change to
1903 match change in how policy works
1905 * dbus/dbus-internals.h: mark assert_not_reached as
1906 __attribute((noreturn))__
1908 2003-04-11 Havoc Pennington <hp@redhat.com>
1910 * doc/dbus-specification.sgml: fix a spot with the wrong name for
1911 the broadcast service. Use boolean return for ServiceExists.
1913 2003-04-11 Havoc Pennington <hp@redhat.com>
1915 * configure.in: add another directory to look for qt in.
1917 2003-04-11 Havoc Pennington <hp@redhat.com>
1919 * AUTHORS: add Colin Walters
1921 2003-04-11 Havoc Pennington <hp@redhat.com>
1927 2003-04-11 Havoc Pennington <hp@redhat.com>
1929 * bus/messagebus.in: remove pid file when stopping the
1930 message bus, since the bus won't have privileges to remove it
1933 2003-04-11 Havoc Pennington <hp@redhat.com>
1935 * bus/bus.c (bus_context_new): move credentials change after
1938 2003-04-11 Havoc Pennington <hp@pobox.com>
1940 * test/decode-gcov.c: add "below average functions" to the
1941 coverage report, and change how some of the code works.
1943 * bus/test-main.c: bracket some stuff in DBUS_BUILD_TESTS so it's
1944 not in the coverage stats.
1946 * test/test-service.c (main): use _dbus_verbose not fprintf in a
1947 couple places so running the test suite doesn't result in megaspam.
1949 2003-04-11 Havoc Pennington <hp@pobox.com>
1951 * bus/dispatch.c (check_existent_service_activation): accept a no
1952 memory error in a place we didn't before
1954 * bus/test.c (bus_test_run_everything): remove hacky "do it twice
1955 in case the first one failed," since the test suite is less
1958 2003-04-10 Havoc Pennington <hp@pobox.com>
1960 * bus/dispatch.c (check_segfault_service_activation): add test
1961 for launching an executable that just crashes.
1963 * test/test-segfault.c (main): try setting coredumpsize to 0 so we
1964 don't leave a million cores. We'll see how portable this is.
1966 2003-04-10 Havoc Pennington <hp@pobox.com>
1968 * dbus/dbus-spawn.c (_dbus_spawn_async_with_babysitter): move all
1969 the possible parent failures before we fork, so that we don't
1970 fail to create a babysitter after creating the child.
1972 * bus/activation.c (bus_activation_activate_service): kill child
1973 if we don't successfully complete the activation.
1975 2003-04-10 Havoc Pennington <hp@redhat.com>
1977 * dbus/dbus-connection.c (dbus_connection_flush): don't spin on
1978 the connection if it's disconnected
1980 * bus/activation.c (bus_activation_service_created): use new
1981 transaction features to roll back removal of pending activation if
1982 we don't successfully create the service after all. Don't remove
1983 pending activation if the function fails.
1985 * dbus/dbus-list.c (_dbus_list_insert_before_link)
1986 (_dbus_list_insert_after_link): new code to facilitate
1989 * dbus/dbus-hash.c (_dbus_hash_table_insert_string_preallocated):
1990 new functionality, so we can preallocate the ability to insert
1993 * bus/connection.c (bus_transaction_add_cancel_hook): new function
1994 allowing us to put custom hooks in a transaction to be used for
1995 cancelling said transaction
1997 * doc/dbus-specification.sgml: add some discussion of secondary
1998 service owners, and disallow zero-length service names
2000 * bus/services.c (bus_registry_acquire_service): new function,
2001 splits out part of bus_driver_handle_acquire_service() and fixes
2002 a bug where we didn't remove the service doing the acquiring
2003 from the secondary queue if we failed to remove the current owner
2004 from the front of the queue.
2006 2003-04-10 Alexander Larsson <alexl@redhat.com>
2008 * doc/dbus-specification.sgml:
2009 s/org.freedesktop.Broadcast/org.freedesktop.DBus.Broadcast/
2011 2003-04-10 Alexander Larsson <alexl@redhat.com>
2016 Added files to cvsignore
2018 * dbus/dbus-message.h:
2019 * dbus/dbus-message.c: (dbus_message_iter_get_named):
2020 Make get_named() take two out argument and return a boolean.
2021 (dbus_message_iter_get_args_valist):
2022 Update usage of get_named().
2023 (dbus_message_iter_append_byte):
2025 (dbus_message_iter_append_named)
2027 (message_iter_test), (check_message_handling_type), (_dbus_message_test):
2030 2003-04-10 Alexander Larsson <alexl@redhat.com>
2032 * dbus/dbus-marshal.[ch]:
2033 Add array_type_pos argument to _dbus_marshal_validate_arg.
2034 Let you pass a NULL end_pos to _dbus_marshal_validate_type.
2036 * dbus/dbus-message.[ch]:
2037 Multi-dimensional arrays have full type specification in the
2038 outermost array. Iter code re-arranged to handle this.
2039 Added some more iter tests.
2041 * doc/dbus-specification.sgml:
2044 Update new array encoding description.
2045 Correct DBUS_SERVICE_FLAGS_REPLACE_EXISTING description.
2047 * test/data/invalid-messages/array-with-mixed-types.message:
2048 * test/data/valid-messages/array-of-array-of-uint32.message:
2049 Change to the new array format.
2051 * test/data/invalid-messages/too-short-dict.message:
2054 * test/data/valid-messages/recursive-types.message:
2055 Fix up and extend test.
2057 2003-04-10 Havoc Pennington <hp@pobox.com>
2059 * bus/dispatch.c: lots of fixes
2061 * dbus/dbus-mainloop.c (_dbus_loop_dispatch): export
2062 (_dbus_loop_iterate): remove old "quit if no callbacks" code,
2063 that was crack, broke the test service.
2065 * dbus/dbus-transport.c (_dbus_transport_open): fix error
2066 handling to avoid piling up errors if we get a failure on the
2069 * dbus/dbus-internals.c (_dbus_real_assert_not_reached): include
2070 pid in assertion failures.
2072 * dbus/dbus-mainloop.c (_dbus_loop_iterate): use static arrays up
2073 to some fixed size of file descriptor array. Don't return TRUE
2074 anytime a timeout exists, that led to lots of busy loop silliness
2077 2003-04-09 Havoc Pennington <hp@redhat.com>
2079 * dbus/dbus-mainloop.c (check_timeout): fix timeouts, I thought
2080 I'd checked this in earlier but hadn't.
2082 2003-04-09 Havoc Pennington <hp@redhat.com>
2084 * bus/dispatch.c (bus_dispatch_test): get a bit further through
2085 the activation test (man this is getting old!)
2087 2003-04-09 Havoc Pennington <hp@redhat.com>
2089 * test/test-utils.c: use dispatch status function to fix this up
2091 * bus/connection.c (connection_watch_callback): don't dispatch
2093 (connection_timeout_callback): don't dispatch from here
2094 (bus_connections_setup_connection): set the dispatch status function
2095 (bus_connection_disconnected): unset it
2097 * dbus/dbus-mainloop.c (_dbus_loop_queue_dispatch): new function
2098 used to add a connection to be dispatched
2099 (_dbus_loop_iterate): do the dispatching at the end of each
2102 * dbus/dbus-connection.c
2103 (dbus_connection_set_dispatch_status_function): new function
2104 allowing us to fix up main loop usage
2105 (_dbus_connection_last_unref): free all the various function
2107 (dbus_connection_dispatch): call the DispatchStatusFunction
2108 whenever this function returns
2109 (dbus_connection_handle_watch): call DispatchStatusFunction
2110 (dbus_connection_send_with_reply_and_block): call DispatchStatusFunction
2111 (reply_handler_timeout): call DispatchStatusFunction
2112 (dbus_connection_flush): call DispatchStatusFunction
2114 2003-04-09 Havoc Pennington <hp@redhat.com>
2116 * dbus/dbus-bus.c (dbus_bus_register): fix up error handling and
2119 * bus/dispatch.c (check_service_activated): fix bug in test
2121 * dbus/dbus-mainloop.c (check_timeout): fix this up
2123 * dbus/dbus-internals.c (_dbus_verbose_real): include PID in
2124 verbose output so we can sort out output from different processes,
2125 e.g. in the activation case.
2127 2003-04-08 Colin Walters <walters@gnu.org>
2129 * bus/bus.c (struct BusContext) [pidfile]: New member, to store
2131 (bus_context_new): Set it.
2132 (bus_context_unref): Use it to delete the pid file.
2134 2003-04-08 Havoc Pennington <hp@redhat.com>
2136 * test/data/invalid-messages/array-with-mixed-types.message:
2137 regression test that fails for the moment
2139 * dbus/dbus-test.c (dbus_internal_do_not_use_run_tests): reorder
2140 tests for convenience
2142 * dbus/dbus-marshal.c (_dbus_marshal_validate_arg): don't allow
2143 array of nil, it broke things.
2145 * test/data/invalid-messages/array-of-nil.message: regression test
2147 * test/data/valid-messages/array-of-array-of-uint32.message:
2148 happened to write this so added it to suite
2150 2003-04-08 Havoc Pennington <hp@redhat.com>
2152 * bus/driver.c (bus_driver_handle_acquire_service): init
2153 retval/reply before checking name
2155 * dbus/dbus-marshal.c (_dbus_marshal_validate_arg): add a
2156 recursion depth argument
2158 * dbus/dbus-message.h (struct DBusMessageIter): put some padding
2159 in the public struct for future extension
2161 * dbus/dbus-message-builder.c (_dbus_message_data_load): fix
2164 * dbus/dbus-marshal.c (_dbus_marshal_validate_arg): fix a verbose
2167 * doc/dbus-specification.sgml: fix typo
2169 2003-04-08 Alexander Larsson <alexl@redhat.com>
2171 Implemented recursive types, named types and new-style iters
2174 * glib/test-thread-client.c: (thread_func):
2175 * glib/test-thread-server.c: (handle_test_message):
2176 * test/test-service.c: (handle_echo):
2185 * dbus/dbus-internals.c: (_dbus_type_to_string):
2186 Update for new types.
2188 * dbus/dbus-marshal.[ch]:
2189 Implement recursive types and the new marshalling format.
2190 Remove hardcoded dict marshalling.
2191 Marshal named types.
2193 * dbus/dbus-message-builder.c:
2195 Remove references to old types
2197 * dbus/dbus-message.[ch]:
2198 New non-refcounted iter API that supports recursive iters.
2199 Use iters for appending, including support for recursive
2201 Add byte and named type support.
2202 Update everything to new marshalling formats.
2203 Add tests for new API.
2205 * dbus/dbus-protocol.h:
2206 Remove old array types.
2207 Add types: BYTE, ARRAY, DICT, NAMED
2209 * dbus/dbus-string.c:
2210 * dbus/dbus-sysdeps.c:
2211 Make parse_double locale safe.
2213 * dbus/dbus-test-main.c:
2219 * doc/dbus-specification.sgml:
2222 * test/data/incomplete-messages/missing-body.message:
2223 * test/data/invalid-messages/bad-boolean.message:
2224 * test/data/invalid-messages/bad-boolean-array.message:
2225 * test/data/invalid-messages/boolean-array-length-too-long.message-raw:
2226 * test/data/invalid-messages/boolean-has-no-value.message-raw:
2227 * test/data/invalid-messages/too-short-dict.message:
2228 * test/data/valid-messages/dict-simple.message:
2229 * test/data/valid-messages/dict.message:
2230 * test/data/valid-messages/emptiness.message:
2231 * test/data/valid-messages/lots-of-arguments.message:
2232 * test/data/valid-messages/no-padding.message:
2233 * test/data/valid-messages/recursive-types.message:
2234 Add missing NAME fields
2235 Fix up dicts & arrays
2237 * test/data/invalid-messages/dict-with-nil-value.message:
2238 Removed, this is not invalid anymore.
2240 * test/data/valid-messages/recursive-types.message:
2241 Add new test for deeply recursive types.
2243 2003-04-07 Havoc Pennington <hp@pobox.com>
2245 * bus/driver.c (bus_driver_handle_acquire_service): return an
2246 error if you try to acquire a service that starts with ':'
2248 2003-04-07 Havoc Pennington <hp@redhat.com>
2250 * doc/dbus-specification.sgml: require that base service names
2251 start with ':' and that the base service is created/deleted
2252 as first and last things a connection does on the bus
2254 * bus/dispatch.c (check_existent_service_activation): lots more
2255 work on the activation test; it doesn't fully pass yet...
2257 * test/test-service.c (main): fix so we don't memleak the
2258 connection to the message bus
2259 (filter_func): accept a message asking us to exit
2261 2003-04-06 Havoc Pennington <hp@pobox.com>
2263 * qt/Makefile.am (dbusinclude_HEADERS): install dbus-qt.h,
2266 * configure.in: fixes to Qt detection from Colin Walters
2268 * doc/Makefile.am: Only remove generated docbook dirs if they
2269 exist, from Colin Walters
2271 * dbus/dbus-bus.c: change how we set well-known connections to
2272 NULL, so that it works if a single connection is stored in
2273 two well-known array slots.
2275 * test/Makefile.am: remove a lot of stuff that isn't immediately
2276 useful, it's in CVS history if we want it.
2278 * test/test-service.c: use dbus-mainloop instead of that
2281 2003-04-06 Havoc Pennington <hp@pobox.com>
2283 * dbus/Makefile.am: split lists of sources into stuff that goes in
2284 the library, util functions that go in the lib and are also used
2285 elsewhere, and util functions that are used in tests/daemon but
2286 don't go in the lib.
2288 * dbus/dbus-mainloop.h, dbus/dbus-mainloop.c: move bus/loop.[hc]
2289 here so it can be used in test binaries also
2291 2003-04-06 Havoc Pennington <hp@pobox.com>
2293 * dbus/dbus-sysdeps.c (_dbus_become_daemon): write the pidfile
2294 here in the parent process, so we can return an error if it
2295 fails. Also, move some of the code into the child so the parent
2296 is less hosed if we fail midway through.
2298 * bus/bus.c (bus_context_new): move pidfile detection further up
2299 in the function, before we start overwriting sockets and such.
2301 * bus/messagebus.in: adjust this a bit, not sure if it will work.
2303 * configure.in: add --with-system-pid-file and --with-system-socket
2305 2003-04-06 Colin Walters <walters@verbum.org>
2307 * configure.in (DBUS_SYSTEM_PID_FILE): New variable.
2309 * bus/system.conf.in: Declare a pidfile.
2311 * bus/bus.c (bus_context_new): Test for an existing pid file, and
2312 create one (if appropriate).
2314 * bus/config-parser.c (enum ElementType) [ELEMENT_PIDFILE]: New.
2315 (struct BusConfigParser) [pidfile]: New.
2316 (element_type_to_name, merge_included, start_busconfig_child)
2317 (bus_config_parser_end_element, bus_config_parser_content): Handle it.
2318 (bus_config_parser_unref): Free it.
2319 (bus_config_parser_get_pidfile): New function.
2321 * bus/config-parser.h (_dbus_write_pid_file): Prototype.
2323 * dbus/dbus-errors.h (DBUS_ERROR_PIDFILE_EXISTS): New error.
2325 * dbus/dbus-sysdeps.c (_dbus_write_pid_file): New function.
2327 * dbus/dbus-sysdeps.h: Prototype it.
2329 2003-04-06 Havoc Pennington <hp@pobox.com>
2331 * bus/bus.c (bus_context_new): print the address in here, rather
2332 than in main(), because we need to do it before forking the daemon
2334 * bus/dispatch.c (send_service_nonexistent_error): set the sender
2335 on the service nonexistent error
2337 * bus/driver.c (bus_driver_handle_acquire_service): set the
2338 sender on the AcquireService reply
2340 * test/data/valid-config-files/debug-allow-all.conf.in: Make test
2341 server also listen on a UNIX socket so services can connect to it.
2343 2003-04-06 Havoc Pennington <hp@pobox.com>
2345 * dbus/dbus-threads.c: Redo how the fake debug mutexes are done
2346 so it detects deadlocks and also we actually init threads when
2349 2003-04-06 Havoc Pennington <hp@pobox.com>
2351 * dbus/dbus-server-unix.c (_dbus_server_new_for_domain_socket):
2352 save the domain socket name, and unlink it when we disconnect the
2353 server. Means that at least when we exit normally, we won't leave
2354 a bunch of junk in /tmp
2356 * dbus/dbus-transport-unix.c
2357 (_dbus_transport_new_for_domain_socket): code cleanup (nicer
2358 memory management). (I was making a real change here but then
2361 2003-04-06 Havoc Pennington <hp@pobox.com>
2363 * bus/bus.c (bus_context_new): fix wrong handling of
2364 server_data_slot_unref() in the error case.
2366 * dbus/dbus-internals.h (_dbus_assert): change so it passes
2367 "(condition) != 0" to _dbus_real_assert so that
2368 "_dbus_assert (pointer)" doesn't cause a warning
2370 * bus/main.c (main): accept --print-address option to print out
2371 the message bus address
2373 * dbus/dbus-sysdeps.c (_dbus_generate_random_ascii): export this
2375 * dbus/dbus-transport.c (_dbus_transport_open): special error for
2376 "tmpdir" option to unix: address on client side
2378 * dbus/dbus-server.c (dbus_server_listen): handle "tmpdir" option
2381 * configure.in (TEST_SOCKET_DIR): locate a temporary directory
2382 we can use to create sockets in the test suite.
2384 * bus/main.c (signal_handler): on SIGTERM, exit the daemon
2385 cleanly. To be used for testing.
2387 * dbus/dbus-spawn.c (babysit): use _dbus_set_signal_handler()
2389 * dbus/dbus-sysdeps.c (_dbus_set_signal_handler): new
2391 * dbus/dbus-server-debug-pipe.c (_dbus_transport_debug_pipe_new):
2392 handle trying to call this when there's no servers active
2394 2003-04-05 Havoc Pennington <hp@pobox.com>
2400 2003-04-05 Havoc Pennington <hp@pobox.com>
2402 * bus/bus.c (setup_server): fix this so dbus-daemon-1 doesn't
2403 crash on startup. Need to get "try starting the daemon"
2404 in the test suite I guess. ;-)
2406 * dbus/dbus-server.h, dbus/dbus-server.c: remove the stuff that
2407 tracked the number of open connections; it's better done in
2408 application-specific code as you want it to span all servers etc.
2410 2003-04-05 Havoc Pennington <hp@pobox.com>
2412 * bus/Makefile.am (install-data-hook): add missing DESTDIR,
2413 patch from Colin Walters
2415 2003-04-05 Havoc Pennington <hp@pobox.com>
2417 * doc/config-file.txt (Elements): fix docs of <auth> to reflect
2418 reality; in fact multiple mechanisms are allowed.
2420 * dbus/dbus-internals.c (_dbus_real_assert)
2421 (_dbus_real_assert_not_reached): move guts of _dbus_assert() and
2422 _dbus_assert_not_reached() into functions, so that they don't show
2423 up in basic block counts for test coverage, and don't use up as
2424 much disk space. Does mean slower execution speed though, so
2425 assumes --disable-asserts is the normal production case.
2427 2003-04-05 Havoc Pennington <hp@pobox.com>
2429 * test/Makefile.am (dist-hook): also dist *.in files
2435 2003-04-05 Havoc Pennington <hp@pobox.com>
2437 * dbus/dbus-string.c: docs warning
2439 * dbus/dbus-spawn.c: missing docs
2441 * dbus/dbus-memory.c (struct ShutdownClosure): missing docs
2443 2003-04-05 Havoc Pennington <hp@pobox.com>
2445 * bus/loop.c (bus_loop_iterate): fix the timeout code, using
2448 * dbus/dbus-spawn.c (_dbus_babysitter_unref): set sitter_pid
2449 to -1 once we've reaped the babysitter
2450 (_dbus_babysitter_handle_watch): do as much work as we can, not
2453 * bus/activation.c: add code using DBusBabysitter so that we
2454 handle it when a service fails to start up properly.
2455 (bus_activation_service_created): don't remove the activation
2456 entries as we go, just let them get removed when we free the pending
2457 activation. Unref reply messages after sending them.
2459 2003-04-05 Havoc Pennington <hp@pobox.com>
2461 * test/decode-gcov.c (main): print per-directory stats in the report
2463 * Makefile.am (coverage-report.txt): don't include test/* in gcov stats
2465 2003-04-05 Havoc Pennington <hp@pobox.com>
2467 * Makefile.am (coverage-report.txt): add target "coverage-report.txt"
2469 * test/decode-gcov.c: hack up a little program to suck data
2470 out of gcov files. Yes this is sort of silly.
2472 * configure.in: define something in config.h and do an
2473 AM_CONDITIONAL when gcov is enabled
2475 2003-04-04 Havoc Pennington <hp@redhat.com>
2477 * dbus/dbus-spawn.c, dbus/dbus-spawn.h: Change dbus_spawn to
2478 return a "babysitter" object that is used to monitor the status of
2479 the spawned process and reap it when required.
2481 * test/test-segfault.c, test/test-exit.c,
2482 test/test-sleep-forever.c: binaries that do various lame things,
2483 used in the test suite.
2485 * dbus/dbus-sysdeps.c: kill _dbus_errno_to_string()
2487 2003-04-03 Havoc Pennington <hp@pobox.com>
2489 * dbus/dbus-spawn.c: Move dbus-spawn into a separate file
2490 in preparation for modifying it, dbus-sysdeps is getting
2493 2003-04-03 Havoc Pennington <hp@redhat.com>
2495 * bus/loop.h, bus/loop.c: make the mainloop an object so we can
2498 * bus/*.[hc]: adapt to mainloop change
2500 2003-04-03 Havoc Pennington <hp@redhat.com>
2502 * bus/activation.c (load_directory): fix up memleaks
2503 (bus_activation_entry_free): free the entry
2505 * dbus/dbus-bus.c (dbus_bus_acquire_service): return an error if
2506 we get one from the message bus; fix memleaks.
2508 * dbus/dbus-message.c (dbus_set_error_from_message): new function
2510 2003-04-03 Havoc Pennington <hp@pobox.com>
2512 * bus/config-parser.c (bus_config_parser_unref): free
2513 list of mechanisms, bug discovered by test suite enhancements
2514 (putting system.conf and session.conf into suite)
2516 * test/Makefile.am, test/test-service.c: add placeholder for a
2517 test service that we'll activate as part of test suite. Doesn't
2520 * dbus/dbus-sysdeps.c (_dbus_setenv): support unsetenv by
2521 setting NULL value, and use system malloc not dbus_malloc()
2522 when we have unavoidable memleakage.
2524 * dbus/dbus-bus.c (dbus_bus_get): fix bug where bus type of 0
2525 didn't work, and support DBUS_BUS_ACTIVATION.
2527 * bus/activation.c (child_setup): pass our well-known bus type to
2530 * bus/config-parser.c: support <type> to specify well-known type
2532 * doc/dbus-specification.sgml: document the env variables to
2533 locate well-known buses and find service activator
2535 2003-04-02 Havoc Pennington <hp@redhat.com>
2537 * test/Makefile.am (all-local): add a rule to copy tests to
2538 builddir, so we can have generated tests. Use this to remove the
2539 silly hack for testing system.conf and session.conf. Will use this
2540 shortly to generate .service files pointing to test binaries.
2542 2003-04-02 Havoc Pennington <hp@redhat.com>
2544 * dbus/dbus-string.c (set_length): fix a bug - we allocated max of
2545 current alloc and needed new length, not max of the doubled
2546 allocation and needed new length. Also, when building tests,
2547 don't do the double-allocation stuff, just realloc every time.
2549 2003-04-02 Havoc Pennington <hp@redhat.com>
2551 * dbus/dbus-sysdeps.c (_dbus_file_get_contents): include filenames
2553 (_dbus_string_get_dirname): new
2554 (_dbus_sysdeps_test): new
2555 (_dbus_directory_open): include dirnames in error messages
2557 * bus/config-parser.c: interpret <include> and <includedir> and
2558 <servicedir> relative to config file location if the given
2559 filename is not absolute.
2561 * dbus/dbus-string.c (_dbus_string_find_byte_backward): new
2563 2003-04-02 Havoc Pennington <hp@redhat.com>
2565 * bus/connection.c (bus_transaction_send_error_reply): set sender
2566 service for the error, and unref the reply on success
2568 * bus/activation.c: convert to use BusTransaction so OOM can be
2570 (bus_activation_service_created): set sender of the message
2572 2003-04-01 Havoc Pennington <hp@redhat.com>
2574 * bus/config-parser.c, bus/bus.c: implement <servicedir> and
2575 <includedir> (at least mostly)
2577 * dbus/dbus-sysdeps.c (_dbus_change_identity): set the group ID
2578 first, then the user ID
2580 2003-04-01 Havoc Pennington <hp@pobox.com>
2582 * dbus/dbus-server.c (dbus_server_set_auth_mechanisms): new
2585 * dbus/dbus-auth.c (_dbus_auth_set_mechanisms): new
2587 * dbus/dbus-internals.c (_dbus_dup_string_array): new function
2589 * dbus/dbus-sysdeps.c (_dbus_listen_unix_socket): chmod the
2590 socket 0777, and unlink any existing socket.
2592 * bus/bus.c (bus_context_new): change our UID/GID and fork if
2593 the configuration file so specifies; set up auth mechanism
2596 * bus/config-parser.c (bus_config_parser_content): add support
2597 for <fork> option and fill in code for <auth>
2599 * bus/system.conf.in: add <fork/> to default configuration,
2600 and limit auth mechanisms to EXTERNAL
2602 * doc/config-file.txt (Elements): add <fork>
2604 * dbus/dbus-sysdeps.c (_dbus_become_daemon): new function
2605 (_dbus_change_identity): new function
2607 2003-03-31 Havoc Pennington <hp@redhat.com>
2609 * dbus/dbus-sysdeps.c (_dbus_connect_unix_socket)
2610 (_dbus_listen_unix_socket): fix off-by-one error in null
2611 termination spotted by Nalin
2613 2003-03-31 Havoc Pennington <hp@redhat.com>
2615 * dbus/dbus-keyring.c (_dbus_keyring_new_homedir): allow setting
2616 DBUS_TEST_HOMEDIR when tests are enabled, so we can test without
2617 having a real home directory available.
2619 2003-03-31 Havoc Pennington <hp@redhat.com>
2621 * bus/Makefile.am (install-data-hook): create /var/run/dbus
2623 * bus/messagebus.in: add init script for Red Hat /etc/init.d
2625 * configure.in: add support for specifying a style of init script
2628 2003-03-31 Havoc Pennington <hp@redhat.com>
2630 Fix some annoying DBusString API and fix all affected code.
2632 * dbus/dbus-string.c (_dbus_string_init): get rid of annoying
2634 (_dbus_string_get_data): change to return string instead of using
2636 (_dbus_string_get_const_data): ditto
2637 (_dbus_string_get_data_len): ditto
2638 (_dbus_string_get_const_data_len): ditto
2640 2003-03-31 Havoc Pennington <hp@redhat.com>
2642 * bus/main.c (main): fix up the command line arguments to be nicer
2644 2003-03-31 Havoc Pennington <hp@redhat.com>
2646 * dbus/Makefile.am (INCLUDES): use EXPANDED_LOCALSTATEDIR to
2647 define DBUS_SYSTEM_BUS_PATH as we want to compile in the same
2648 final location that lands in the config file
2650 * bus/config-loader-expat.c (bus_config_load): fix type of
2653 * doc/TODO: remove TODO item for dbus_bus_get()
2655 * dbus/dbus-bus.c (bus_data_free): add missing lock/unlock
2657 2003-03-31 Havoc Pennington <hp@pobox.com>
2659 * dbus/dbus-transport-unix.c (_dbus_transport_new_for_domain_socket)
2660 (_dbus_transport_new_for_tcp_socket): these didn't need the "server"
2661 argument since they are always client side
2663 * dbus/dbus-server.c (dbus_server_get_address): new function
2665 * bus/main.c (main): take the configuration file as an argument.
2667 * test/data/valid-config-files/debug-allow-all.conf: new file to
2668 use with dispatch.c tests for example
2670 * bus/test-main.c (main): require test data dir
2672 * bus/bus.c (bus_context_new): change this to take a
2673 configuration file name as argument
2675 * doc/config-file.txt (Elements): add <servicedir>
2677 * bus/system.conf, bus/session.conf: new files
2679 * dbus/dbus-bus.c (dbus_bus_get): look for system bus on
2680 well-known socket if none set
2682 * configure.in: create system.conf and session.conf
2684 2003-03-30 Havoc Pennington <hp@pobox.com>
2686 * bus/config-parser.c: hacking
2688 * dbus/dbus-memory.c: don't use DBusList for the list of stuff
2689 to shut down, since it could cause weirdness with the DBusList
2692 * dbus/dbus-list.c (_dbus_list_test): add tests for the
2693 link-oriented stack routines
2694 (alloc_link): free the mempool if the first alloc from it fails
2696 * dbus/dbus-mempool.c (struct DBusMemBlock): fix alignment issue
2698 * dbus/dbus-string.c (UNICODE_VALID): sync new version of this
2700 (_dbus_string_skip_white): new
2702 * doc/config-file.txt (Elements): add <includedir>
2704 2003-03-28 Havoc Pennington <hp@pobox.com>
2706 * dbus/dbus-string.c (_dbus_string_copy_data_len)
2707 (_dbus_string_copy_data): new functions
2709 2003-03-28 Anders Carlsson <andersca@codefactory.se>
2711 * dbus/dbus-bus.c: (bus_data_free), (dbus_bus_get):
2715 * dbus/dbus-memory.c:
2718 2003-03-28 Havoc Pennington <hp@pobox.com>
2720 * bus/test.c (bus_test_flush_bus): remove the sleep from here,
2721 I think it may have just been superstition. Not sure.
2723 * dbus/dbus-string.c (_dbus_string_base64_decode): catch some OOM
2724 failures that were not being handled.
2726 * dbus/dbus-auth.c (process_auth): fix a memleak in OOM handling
2728 * dbus/dbus-memory.c: add ability to set number of mallocs in a
2729 row that will fail on out-of-memory.
2731 * dbus/dbus-internals.c (_dbus_test_oom_handling): convenience
2732 function for testing out-of-memory handling.
2734 * bus/config-loader-expat.c (memsuite): don't wrap the dbus
2735 allocation functions, they do map exactly to the expat ones.
2737 2003-03-27 Havoc Pennington <hp@redhat.com>
2739 * bus/config-loader-libxml.c (bus_config_load): add another error
2742 2003-03-26 Anders Carlsson <andersca@codefactory.se>
2745 Add note about automatic service activation.
2747 * doc/dbus-specification.sgml:
2748 Rename the specification and clarify a few things.
2750 2003-03-26 Anders Carlsson <andersca@codefactory.se>
2753 * dbus/dbus-address.c:
2755 * dbus/dbus-marshal.c:
2756 * dbus/dbus-server-debug-pipe.c:
2757 * dbus/dbus-transport-unix.c:
2758 Fix documentation warnings.
2760 2003-03-26 Havoc Pennington <hp@pobox.com>
2762 * bus/test-main.c, dbus/dbus-test.c (main): check memleaks
2763 after every test so it's quick and easy to see which leaked, and
2764 so we test multiple dbus_shutdown() calls
2766 * configure.in: change configure.in XML stuff to also support
2769 * config-loader-libxml.c: some hacking
2771 * config-loader-expat.c: some hacking
2773 * config-parser.c: some hacking, plus tests
2775 2003-03-25 Havoc Pennington <hp@redhat.com>
2777 * throughout - add more _DBUS_ASSERT_ERROR_IS_CLEAR
2779 * configure.in: add --with-xml option to specify XML library,
2780 right now only libxml is supported.
2782 * bus/config-loader-libxml.c, config-parser.c: sync some minor
2783 nonworking code between home and work, still just stubs
2785 2003-03-24 Havoc Pennington <hp@redhat.com>
2787 * dbus/dbus-sysdeps.c (_dbus_set_fd_nonblocking): move to this
2790 * dbus/dbus-errors.c (dbus_set_error, dbus_set_error_const): allow
2791 NULL argument for "message" if the error is a well-known one,
2792 fill in a generic message in this case.
2794 * dbus/dbus-errors.h (DBusResultCode): Kill DBusResultCode in
2797 * bus/test.c (bus_test_flush_bus): add
2799 * bus/policy.c (bus_policy_test): test code stub
2801 2003-03-24 Havoc Pennington <hp@pobox.com>
2803 * bus/connection.c (bus_connections_setup_connection): set up
2804 the "can this user connect" function, but it always returns
2805 TRUE until we have a config file parser so we can have a config
2806 file that allows connections.
2808 2003-03-23 Havoc Pennington <hp@pobox.com>
2810 * dbus/dbus-threads.c (dbus_mutex_new, dbus_condvar_new): with
2811 DBUS_BUILD_TESTS, actually alloc/free a block of memory for
2812 the mutex, so we can check for proper memory management
2815 * dbus/dbus-dataslot.c: remove the mutex from
2816 DBusDataSlotAllocator and lock it manually when using it,
2817 to simplify fitting it into the global slots framework.
2819 * dbus/dbus-threads.c (init_static_locks): rework how we're
2820 handling global locks so they are easily shut down.
2822 * bus/policy.c (bus_policy_append_rule): fix
2824 * bus/test-main.c (main): check for memleaks
2826 * dbus/dbus-test.c (dbus_internal_do_not_use_run_tests): make
2827 test suite check for memleaks
2829 * dbus/dbus-memory.c: add support in test mode for tracking
2830 number of outstanding blocks
2832 2003-03-23 Havoc Pennington <hp@pobox.com>
2834 * bus/policy.c, bus/bus.c, bus/connection.c: implement allow/deny
2837 * dbus/dbus-hash.h: add ULONG hash keys
2839 * dbus/dbus-sysdeps.c (_dbus_get_groups): new
2840 (_dbus_get_group_id): new function
2842 2003-03-20 Havoc Pennington <hp@redhat.com>
2844 * dbus/dbus-connection.c (dbus_connection_set_unix_user_function):
2846 (dbus_connection_get_unix_user): new function
2848 2003-03-20 Havoc Pennington <hp@pobox.com>
2850 * bus/connection.c (bus_connection_send_oom_error): assert that
2851 message has a sender
2852 (connection_execute_transaction): ditto
2853 (bus_connection_preallocate_oom_error): fix to set the sender, and
2854 set recipient to the destination service, not the bus driver
2856 * bus/policy.c: hacking
2858 * dbus/dbus-message.c (dbus_message_service_is): new function
2859 (dbus_message_sender_is): new
2861 2003-03-19 Havoc Pennington <hp@redhat.com>
2863 * bus/policy.c: start sketching code for policy restrictions on
2864 what connections can do.
2866 2003-03-18 Havoc Pennington <hp@redhat.com>
2868 * doc/TODO: some notes on high-level todo items. Little nitpick
2869 stuff is all in @todo, so no need to add it here.
2871 * doc/config-file.txt: some notes on how config file might look
2873 2003-03-18 Anders Carlsson <andersca@codefactory.se>
2879 2003-03-17 Havoc Pennington <hp@redhat.com>
2881 * dbus/dbus-internals.h: add gcc attributes so that
2882 our printf-style functions warn on bad arguments to
2885 * dbus/dbus-sysdeps.c (_dbus_connect_tcp_socket): fix printf
2888 * dbus/dbus-message.c (_dbus_message_loader_queue_messages): fix
2891 2003-03-17 Havoc Pennington <hp@redhat.com>
2893 * bus/test-main.c (main): make it print something as it runs
2894 so make check doesn't look stuck
2896 * doc/negotiation.txt, doc/dbus-sasl-profile.txt: remove
2897 from CVS, now obsolete
2899 2003-03-17 Anders Carlsson <andersca@codefactory.se>
2901 * bus/dispatch.c: (bus_dispatch):
2902 Refetch the service name since it may have been reallocated
2903 when dbus_message_set_sender was called.
2905 * dbus/dbus-sysdeps.c: (_dbus_accept):
2906 Add address and address length variables and use them to stop
2907 valgrind from complaining.
2909 2003-03-17 Havoc Pennington <hp@pobox.com>
2911 All tests pass, no memleaks, no valgrind complaints.
2913 * bus/test.c: refcount handler_slot
2915 * bus/connection.c (bus_connections_new): refcount
2916 connection_data_slot
2918 * dbus/dbus-auth-script.c (_dbus_auth_script_run): delete unused
2919 bytes so that auth scripts pass.
2921 * bus/dispatch.c: init message_handler_slot so it gets allocated
2924 * bus/dispatch.c (message_handler_slot_ref): fix memleak
2926 * dbus/dbus-server-debug-pipe.c (_dbus_server_debug_pipe_new):
2927 dealloc server_pipe_hash when no longer used for benefit of
2930 * dbus/dbus-auth.c (process_command): memleak fix
2932 * bus/dispatch.c (check_hello_message): memleak fix
2934 2003-03-16 Havoc Pennington <hp@pobox.com>
2936 * dbus/dbus-bus.c (ensure_bus_data): fix double-unref of the data slot
2938 2003-03-17 Anders Carlsson <andersca@codefactory.se>
2940 * bus/activation.c (bus_activation_activate_service): Append
2941 the pending activation entry to the list of pending activations.
2943 2003-03-16 Havoc Pennington <hp@pobox.com>
2945 * bus/dispatch.c (bus_dispatch_test): remove double-unrefs of
2948 * dbus/dbus-address.c (create_entry): fix OOM handling when
2949 failing to alloc entry->method
2951 2003-03-16 Havoc Pennington <hp@pobox.com>
2953 * dbus/dbus-watch.c (_dbus_watch_new): handle failure to malloc
2956 * dbus/dbus-server-debug-pipe.c (_dbus_transport_debug_pipe_new):
2957 add some missing dbus_set_result
2959 * bus/dispatch.c (bus_dispatch_add_connection): handle failure to
2960 alloc the DBusMessageHandler
2962 * dbus/dbus-transport.c (_dbus_transport_disconnect): don't ref
2963 the transport here, since we call this from the finalizer; it
2964 resulted in a double-finalize.
2966 * dbus/dbus-transport.c (_dbus_transport_disconnect): fix a bug
2967 where we tried to use transport->connection that was NULL,
2968 happened when transport was disconnected early on due to OOM
2970 * bus/*.c: adapt to handle OOM for watches/timeouts
2972 * dbus/dbus-transport-unix.c: port to handle OOM during
2975 * dbus/dbus-auth.c (_dbus_auth_get_unused_bytes): return a
2976 reference to unused bytes instead of a copy
2978 * dbus/dbus-server.c (dbus_server_handle_watch): return FALSE for
2981 * dbus/dbus-connection.c (dbus_connection_handle_watch): return
2984 * dbus/dbus-timeout.c (dbus_timeout_handle): return FALSE for out
2987 2003-03-16 Anders Carlsson <andersca@codefactory.se>
2989 * doc/dbus-specification.sgml:
2990 Document reply message for ActivateService.
2992 2003-03-16 Anders Carlsson <andersca@codefactory.se>
2994 * bus/activation.c: (bus_pending_activation_entry_free),
2995 (bus_pending_activation_free), (bus_activation_new),
2996 (bus_activation_unref), (bus_activation_service_created),
2997 (bus_activation_activate_service):
2999 * bus/bus.c: (bus_context_new):
3000 * bus/desktop-file.c: (new_section):
3001 * bus/driver.c: (bus_driver_send_service_deleted),
3002 (bus_driver_handle_activate_service):
3003 * bus/services.c: (bus_registry_new), (bus_registry_ensure):
3005 * dbus/dbus-connection.c:
3006 (dbus_connection_send_with_reply_and_block):
3007 * dbus/dbus-message.c: (dbus_message_append_args_valist):
3008 * dbus/dbus-protocol.h:
3009 Make activation work better. Now pending activations will be queued
3010 and the daemon won't try to activate services that are already registered.
3012 2003-03-16 Havoc Pennington <hp@pobox.com>
3014 * dbus/dbus-bus.c (ensure_bus_data): handle failure to set
3017 * dbus/dbus-memory.c (_dbus_initialize_malloc_debug): support
3018 DBUS_MALLOC_BACKTRACES to print trace when failing an alloc
3020 2003-03-16 Havoc Pennington <hp@pobox.com>
3022 * dbus/dbus-string.c (_dbus_string_validate_utf8): oops, unbreak
3023 this. always run the test suite before commit...
3025 * bus/*: adapt to DBusConnection API changes
3027 * glib/dbus-gmain.c: adapt to DBusConnection API changes,
3028 requires renaming stuff to avoid dbus_connection_dispatch name
3031 * dbus/dbus-transport.c (_dbus_transport_queue_messages): new
3034 * dbus/dbus-message.c (_dbus_message_loader_queue_messages):
3035 separate from _dbus_message_loader_return_buffer()
3037 * dbus/dbus-connection.c (dbus_connection_get_n_messages): remove
3038 this, because it's now always broken to use; the number of
3039 messages in queue vs. the number still buffered by the message
3040 loader is undefined/meaningless. Should use
3041 dbus_connection_get_dispatch_state().
3042 (dbus_connection_dispatch): rename from
3043 dbus_connection_dispatch_message
3045 2003-03-16 Havoc Pennington <hp@pobox.com>
3047 * dbus/dbus-string.c (_dbus_string_validate_utf8): copy in a real
3050 2003-03-16 Anders Carlsson <andersca@codefactory.se>
3052 * dbus/dbus-connection.c:
3053 (dbus_connection_send_with_reply_and_block):
3054 Decrease connection->n_incoming when removing an entry
3056 * dbus/dbus-dict.c: (dbus_dict_entry_free),
3057 (dbus_dict_set_boolean_array), (dbus_dict_set_int32_array),
3058 (dbus_dict_set_uint32_array), (dbus_dict_set_double_array),
3059 (dbus_dict_set_byte_array), (dbus_dict_set_string_array),
3060 (dbus_dict_get_boolean_array), (dbus_dict_get_double_array),
3061 (dbus_dict_get_byte_array):
3062 Handle NULL arrays and strings. Also add support for byte arrays.
3064 * dbus/dbus-marshal.c: (_dbus_marshal_byte_array),
3065 (_dbus_marshal_dict), (_dbus_demarshal_byte_array),
3066 (_dbus_demarshal_int32_array), (_dbus_demarshal_uint32_array),
3067 (_dbus_demarshal_double_array), (_dbus_demarshal_string_array),
3068 (_dbus_demarshal_dict), (demarshal_and_validate_len),
3069 (_dbus_marshal_validate_arg), (_dbus_marshal_test):
3070 * dbus/dbus-marshal.h:
3071 Add support for marshalling and demarshalling empty arrays and strings.
3073 * dbus/dbus-message.c: (dbus_message_append_args_valist),
3074 (dbus_message_append_string_array),
3075 (dbus_message_iter_get_boolean),
3076 (dbus_message_iter_get_boolean_array),
3077 (dbus_message_iter_get_int32_array),
3078 (dbus_message_iter_get_uint32_array),
3079 (dbus_message_iter_get_double_array),
3080 (dbus_message_iter_get_byte_array),
3081 (dbus_message_iter_get_string_array), (dbus_message_iter_get_dict),
3082 (check_message_handling):
3083 Add support for getting empty arrays and dicts.
3085 * dbus/dbus-string.c: (_dbus_string_validate_utf8):
3086 Don't do any validation at all for now, that's better than just checking
3089 * test/data/valid-messages/emptiness.message:
3090 New test message with lots of empty arrays.
3092 2003-03-16 Havoc Pennington <hp@pobox.com>
3094 * dbus/dbus-connection.c
3095 (_dbus_connection_queue_received_message_link): new function that
3096 can't fail due to OOM
3098 * dbus/dbus-message.c (_dbus_message_loader_pop_message_link):
3099 new function pops a message together with a list link
3102 * dbus/dbus-transport-unix.c (queue_messages): use new link-based
3103 message queuing functions to avoid needing to alloc memory
3105 2003-03-16 Havoc Pennington <hp@pobox.com>
3107 Oops - test code was only testing failure of around 30 of the
3108 mallocs in the test path, but it turns out there are 500+
3109 mallocs. I believe this was due to misguided linking setup such
3110 that there was one copy of dbus_malloc etc. in the daemon and one
3111 in the shared lib, and only daemon mallocs were tested. In any
3112 case, the test case now tests all 500+ mallocs, and doesn't pass
3113 yet, though there are lots of fixes in this patch.
3115 * dbus/dbus-connection.c (dbus_connection_dispatch_message): fix
3116 this so that it doesn't need to allocate memory, since it
3117 has no way of indicating failure due to OOM (and would be
3118 annoying if it did).
3120 * dbus/dbus-list.c (_dbus_list_pop_first_link): new function
3122 * bus/Makefile.am: rearrange to create two self-contained
3123 libraries, to avoid having libraries with overlapping symbols.
3124 that was resulting in weirdness, e.g. I'm pretty sure there
3125 were two copies of global static variables.
3127 * dbus/dbus-internals.c: move the malloc debug stuff to
3130 * dbus/dbus-list.c (free_link): free list mempool if it becomes
3133 * dbus/dbus-memory.c (_dbus_disable_mem_pools): new function
3135 * dbus/dbus-address.c (dbus_parse_address): free list nodes
3138 * bus/dispatch.c (bus_dispatch_add_connection): free
3139 message_handler_slot when no longer using it, so
3140 memory leak checkers are happy for the test suite.
3142 * dbus/dbus-server-debug-pipe.c (debug_finalize): free server name
3144 * bus/bus.c (new_connection_callback): disconnect in here if
3145 bus_connections_setup_connection fails.
3147 * bus/connection.c (bus_connections_unref): fix to free the
3149 (bus_connections_setup_connection): if this fails, don't
3150 disconnect the connection, just be sure there are no side
3153 * dbus/dbus-string.c (undo_alignment): unbreak this
3155 * dbus/dbus-auth.c (_dbus_auth_unref): free some stuff we were
3157 (_dbus_auth_new): fix the order in which we free strings
3160 * bus/connection.c (bus_connection_disconnected): fix to
3161 not send ServiceDeleted multiple times in case of memory
3164 * dbus/dbus-bus.c (dbus_bus_get_base_service): new function to
3165 get the base service name
3166 (dbus_bus_register_client): don't return base service name,
3167 instead store it on the DBusConnection and have an accessor
3169 (dbus_bus_register_client): rename dbus_bus_register()
3171 * bus/dispatch.c (check_hello_message): verify that other
3172 connections on the bus also got the correct results, not
3173 just the one sending hello
3175 2003-03-15 Havoc Pennington <hp@pobox.com>
3177 Make it pass the Hello handling test including all OOM codepaths.
3178 Now to do other messages...
3180 * bus/services.c (bus_service_remove_owner): fix crash when
3181 removing owner from an empty list of owners
3182 (bus_registry_ensure): don't leave service in the list of
3183 a connection's owned services if we fail to put the service
3186 * bus/connection.c (bus_connection_preallocate_oom_error): set
3187 error flag on the OOM error.
3189 * dbus/dbus-connection.c (_dbus_connection_new_for_transport):
3190 handle _dbus_transport_set_connection failure
3192 * dbus/dbus-transport-unix.c (_dbus_transport_new_for_fd): modify
3193 to create watches up front and simply enable/disable them as
3195 (unix_connection_set): this can now fail on OOM
3197 * dbus/dbus-timeout.c, dbus/dbus-watch.c: add concept
3198 of enabling/disabling a watch or timeout.
3200 * bus/loop.c (bus_loop_iterate): don't touch disabled
3203 * glib/dbus-gmain.c: adapt to enable/disable watches and timeouts
3205 2003-03-15 Havoc Pennington <hp@pobox.com>
3207 * bus/dispatch.c (bus_dispatch_test): OK, now finally actually
3208 write useful test code, after all that futzing around ;-)
3210 Test does not yet pass because we can't handle OOM in
3211 _dbus_transport_messages_pending (basically,
3212 dbus_connection_preallocate_send() does not prealloc the write
3213 watch). To fix this, I think we need to add new stuff to
3214 set_watch_functions, namely a SetEnabled function so we can alloc
3215 the watch earlier, then enable it later.
3217 * dbus/Makefile.am (libdbus_convenience_la_SOURCES): move
3218 dbus-memory.c to the convenience lib
3220 * bus/test.c: rename some static functions to keep them clearly
3221 distinct from stuff in connection.c. Handle client disconnection.
3223 2003-03-14 Havoc Pennington <hp@pobox.com>
3225 * bus/dispatch.c (bus_dispatch_test): do test using debug-pipe
3226 transport, tests more of the real codepath. Set up clients
3227 with bus_setup_debug_client.
3229 * bus/test.c (bus_setup_debug_client): function to set up debug
3230 "clients" on the main loop
3232 * dbus/dbus-transport.c (_dbus_transport_open): add debug-pipe
3235 * dbus/dbus-server.c (dbus_server_listen): add debug-pipe
3238 * dbus/dbus-server-debug.c: support a debug server based on pipes
3240 * dbus/dbus-sysdeps.c (_dbus_full_duplex_pipe): new function
3241 (_dbus_close): new function
3243 * configure.in: check for socketpair
3245 2003-03-14 Havoc Pennington <hp@redhat.com>
3247 * dbus/dbus-memory.c: add a "detect buffer overwrites on free"
3250 * dbus/dbus-transport-debug.c: rework this a good bit to be
3251 less complicated. hopefully still works.
3253 * dbus/dbus-server-debug.c (handle_new_client): remove timeout
3256 * glib/dbus-gmain.c (timeout_handler): don't remove timeout
3259 * dbus/dbus-message.c (dbus_message_copy): rename from
3260 dbus_message_new_from_message, fix it up to copy
3261 all the message fields, add test case
3263 * bus/dispatch.c (bus_dispatch_test): add some more test code,
3264 not quite passing yet
3266 2003-03-14 Havoc Pennington <hp@pobox.com>
3268 * bus/loop.c (bus_loop_iterate): add this so we can "run loop
3269 until no work remains" in test code. (the large diff here
3270 is just code movement, no actual changes)
3272 * dbus/dbus-server-debug.c (DEFAULT_INTERVAL): change interval to
3273 1, no point waiting around for test code.
3274 (_dbus_server_debug_accept_transport): unref the timeout
3275 after adding it (right?)
3277 * dbus/dbus-transport-debug.c (DEFAULT_INTERVAL): ditto
3279 2003-03-13 Havoc Pennington <hp@redhat.com>
3281 * dbus/dbus-timeout.c (_dbus_timeout_list_set_functions): handle
3284 * dbus/dbus-watch.c (_dbus_watch_list_set_functions): handle out
3287 * dbus/dbus-connection.h: Make AddWatchFunction and
3288 AddTimeoutFunction return a bool so they can fail on out-of-memory
3290 * bus/bus.c (bus_context_new): set up timeout handlers
3292 * bus/connection.c (bus_connections_setup_connection): set up
3295 * glib/dbus-gmain.c: adapt to the fact that set_functions stuff
3298 * bus/bus.c (bus_context_new): adapt to changes
3300 * bus/connection.c: adapt to changes
3302 * test/watch.c: adapt to DBusWatch changes
3304 * bus/dispatch.c (bus_dispatch_test): started adding this but
3307 2003-03-14 Anders Carlsson <andersca@codefactory.se>
3309 * bus/dispatch.c (send_service_nonexistent_error): Fix typo.
3311 2003-03-13 Havoc Pennington <hp@pobox.com>
3313 * bus/test.c, bus/test.h, bus/Makefile.am, bus/test-main.c:
3314 set up a test framework as for the library
3316 2003-03-12 Havoc Pennington <hp@pobox.com>
3318 Throughout: purge global variables, introduce BusActivation,
3319 BusConnections, BusRegistry, etc. objects instead.
3321 * bus/bus.h, bus/bus.c: introduce BusContext as a global
3324 * test/Makefile.am (TEST_BINARIES): disable bus-test for now,
3325 going to redo this a bit differently I think
3327 2003-03-12 Havoc Pennington <hp@redhat.com>
3329 Mega-patch that gets the message bus daemon initially handling
3330 out-of-memory. Work still needed. Also lots of random
3331 moving stuff to DBusError instead of ResultCode.
3333 * dbus/dbus-list.c (_dbus_list_length_is_one): new function
3335 * dbus/dbus-connection.c
3336 (dbus_connection_send_with_reply_and_block): use DBusError
3338 * dbus/dbus-bus.c: adapt to API changes, make it use DBusError not
3341 * dbus/dbus-connection.c (dbus_connection_send): drop the result
3342 code here, as the only failure possible is OOM.
3344 * bus/connection.c (bus_connection_disconnect):
3345 rename bus_connection_disconnected as it's a notification only
3347 * bus/driver.c (bus_driver_handle_acquire_service): don't free
3348 "name" on get_args failure, should be done by get_args;
3349 don't disconnect client for bad args, just return an error.
3350 (bus_driver_handle_service_exists): ditto
3352 * bus/services.c (bus_services_list): NULL-terminate returned array
3354 * bus/driver.c (bus_driver_send_service_lost)
3355 (bus_driver_send_service_acquired): send messages from driver to a
3356 specific client to the client's unique name, not to the broadcast
3359 * dbus/dbus-message.c (decode_header_data): reject messages that
3360 contain no name field
3361 (_dbus_message_get_client_serial): rename to
3362 dbus_message_get_serial and make public
3363 (_dbus_message_set_serial): rename from set_client_serial
3364 (_dbus_message_set_reply_serial): make public
3365 (_dbus_message_get_reply_serial): make public
3367 * bus/connection.c (bus_connection_foreach): allow stopping
3368 iteration by returning FALSE from foreach function.
3370 * dbus/dbus-connection.c (dbus_connection_send_preallocated)
3371 (dbus_connection_free_preallocated_send)
3372 (dbus_connection_preallocate_send): new API for sending a message
3373 without possibility of malloc failure.
3374 (dbus_connection_send_message): rename to just
3375 dbus_connection_send (and same for whole function family)
3377 * dbus/dbus-errors.c (dbus_error_free): make this reinit the error
3379 * dbus/dbus-sysdeps.c (_dbus_exit): new function
3381 * bus/activation.c: handle/return errors
3383 * dbus/dbus-errors.h: add more DBUS_ERROR #define
3385 * dbus/dbus-sysdeps.c (_dbus_directory_open) (_dbus_file_get_contents)
3386 (_dbus_directory_get_next_file): use DBusError instead of DBusResultCode
3387 (_dbus_result_from_errno): move to this file
3389 2003-03-10 Anders Carlsson <andersca@codefactory.se>
3391 * dbus/dbus-marshal.c:
3392 (_dbus_marshal_set_string):
3393 Take a length argument so we can marshal the correct string
3396 (_dbus_marshal_dict), (_dbus_demarshal_dict),
3397 (_dbus_marshal_get_arg_end_pos), (_dbus_marshal_validate_arg),
3398 (_dbus_marshal_test):
3399 * dbus/dbus-marshal.h:
3400 Add support for marshalling and demarshalling dicts.
3402 * dbus/dbus-message-builder.c: (_dbus_message_data_load):
3403 Add support for TYPE DICT.
3405 * dbus/dbus-message.c: (set_string_field):
3406 Adjust header padding.
3408 (dbus_message_append_args_valist), (dbus_message_append_dict),
3409 (dbus_message_get_args_valist), (dbus_message_iter_get_arg_type),
3410 (dbus_message_iter_get_dict), (_dbus_message_loader_return_buffer),
3411 (check_message_handling), (check_have_valid_message):
3412 * dbus/dbus-message.h:
3413 Add functions for setting and getting dicts.
3415 * dbus/dbus-protocol.h:
3421 * doc/dbus-specification.sgml:
3422 Add information about how dicts are marshalled.
3424 * test/data/invalid-messages/dict-with-nil-value.message:
3425 * test/data/invalid-messages/too-short-dict.message:
3426 * test/data/valid-messages/dict-simple.message:
3427 * test/data/valid-messages/dict.message:
3428 Add sample messages containing dicts.
3430 2003-03-08 Anders Carlsson <andersca@codefactory.se>
3432 * dbus/dbus-dict.h: Add DBUS_END_DECLS.
3434 2003-03-07 Anders Carlsson <andersca@codefactory.se>
3437 * dbus/dbus-dict.c: (dbus_dict_entry_free), (dbus_dict_new),
3438 (dbus_dict_get_keys), (insert_entry), (dbus_dict_set_boolean),
3439 (dbus_dict_set_int32), (dbus_dict_set_uint32),
3440 (dbus_dict_set_double), (dbus_dict_set_string),
3441 (dbus_dict_set_boolean_array), (dbus_dict_set_int32_array),
3442 (dbus_dict_set_uint32_array), (dbus_dict_set_double_array),
3443 (dbus_dict_set_string_array), (_dbus_dict_test):
3445 Fix according to comments from Havoc.
3447 2003-03-06 Michael Meeks <michael@server.home>
3449 * configure.in: if we don't have kde-config, disable have_qt.
3451 2003-03-07 Anders Carlsson <andersca@codefactory.se>
3456 * dbus/dbus-dict.c: (dbus_dict_entry_free), (dbus_dict_new),
3457 (dbus_dict_ref), (dbus_dict_unref), (dbus_dict_contains),
3458 (dbus_dict_remove), (dbus_dict_get_value_type),
3459 (dbus_dict_get_keys), (dbus_dict_put_boolean),
3460 (dbus_dict_put_int32), (dbus_dict_put_uint32),
3461 (dbus_dict_put_double), (dbus_dict_put_string),
3462 (dbus_dict_put_boolean_array), (dbus_dict_put_int32_array),
3463 (dbus_dict_put_uint32_array), (dbus_dict_put_double_array),
3464 (dbus_dict_put_string_array), (dbus_dict_get_boolean),
3465 (dbus_dict_get_int32), (dbus_dict_get_uint32),
3466 (dbus_dict_get_double), (dbus_dict_get_string),
3467 (dbus_dict_get_boolean_array), (dbus_dict_get_int32_array),
3468 (dbus_dict_get_uint32_array), (dbus_dict_get_double_array),
3469 (dbus_dict_get_string_array), (_dbus_dict_test):
3471 Add DBusDict implementation
3473 * dbus/dbus-test.c: (dbus_internal_do_not_use_run_tests):
3477 2003-03-04 Havoc Pennington <hp@pobox.com>
3479 * test/data/auth/*: adapt to changes
3481 * dbus/dbus-auth-script.c (_dbus_auth_script_run): add
3482 USERID_BASE64 and change USERNAME_BASE64 to put in username not
3485 * dbus/dbus-keyring.c (_dbus_keyring_validate_context): prevent
3486 more stuff from being in a context name, to make the protocol
3487 simpler to deal with
3489 * dbus/dbus-errors.c (dbus_error_has_name): new function
3490 (dbus_error_is_set): new function
3492 * dbus/dbus-auth.c: replace DBUS_STUPID_TEST_MECH auth
3493 with DBUS_COOKIE_SHA1, implement DBUS_COOKIE_SHA1
3495 * dbus/dbus-connection.c (dbus_connection_flush): also read
3496 messages during a flush operation
3498 * dbus/Makefile.am: remove dbus-md5 since it isn't currently used.
3500 2003-03-05 Anders Carlsson <andersca@codefactory.se>
3502 * configure.in: Check for gethostbyname on Solaris.
3504 * dbus/dbus-transport.c: (_dbus_transport_open):
3505 Remove duplicate "tcp" entry.
3507 * doc/dbus-specification.sgml:
3508 Clarify some things.
3510 2003-03-05 Anders Carlsson <andersca@codefactory.se>
3512 * dbus/dbus-auth.c: (send_rejected), (process_test_subdir):
3513 * dbus/dbus-keyring.c: (_dbus_keyring_new_homedir),
3514 (_dbus_keyring_test):
3515 * dbus/dbus-md5.c: (_dbus_md5_compute):
3516 * dbus/dbus-sha.c: (_dbus_sha_compute):
3519 2003-03-05 Anders Carlsson <andersca@codefactory.se>
3521 * README: Add some things.
3523 2003-03-04 Anders Carlsson <andersca@codefactory.se>
3525 * dbus/dbus-message.c (dbus_message_append_args_valist): Add a break;
3526 after case DBUS_TYPE_BOOELAN.
3528 2003-03-02 Havoc Pennington <hp@pobox.com>
3530 * test/break-loader.c (randomly_set_extreme_ints): add test that
3531 sets really huge and small integers
3533 * dbus/dbus-marshal.c (_dbus_marshal_validate_arg): add check
3534 that length of boolean array fits in the string, and that
3535 string has room for boolean value in single-bool case.
3537 * dbus/dbus-message-builder.c (_dbus_message_data_load): add
3538 optional value to "ALIGN" command which is what to fill the
3541 * test/data/valid-messages/no-padding.message: add regression
3542 test for the message padding problem
3544 2003-03-02 Havoc Pennington <hp@pobox.com>
3546 * dbus/dbus-message.c (decode_header_data): fix to always init
3547 message_padding, from Benjamin Dauvergne
3549 2003-03-02 Havoc Pennington <hp@pobox.com>
3555 2003-03-01 Joe Shaw <joe@assbarn.com>
3557 * configure.in: Check for "struct cmsgcred" and try to access its
3558 members for BSD-like unices.
3560 * dbus/dbus-sysdeps.c (read_credentials_byte): Fold this back into
3561 _dbus_read_credentials_unix_socket().
3562 (_dbus_read_credentials_unix_socket): Use recvmsg() instead of
3563 read() for reading the credential byte off the unix socket. Use
3564 struct cmsgcred on systems that support it.
3566 2003-02-27 Alexander Larsson <alexl@redhat.com>
3570 Make gthreads-2.0 dependency optional. Don't build thread test if
3573 2003-02-27 Havoc Pennington <hp@pobox.com>
3575 * dbus/dbus-connection.c
3576 (dbus_connection_send_message_with_reply_and_block): fix doh!
3577 doh! doh! bug that resulted in never removing a reply from the
3578 queue, no wonder we called get_reply_serial so much ;-)
3580 * dbus/dbus-message.c (struct DBusMessage): cache reply serial
3581 and client serial instead of demarshaling them every time
3583 2003-02-27 Havoc Pennington <hp@pobox.com>
3585 * dbus/dbus-marshal.c (_dbus_demarshal_int32): rewrite to be much
3586 more inlined, using dbus-string-private.h, speeds things up
3589 * dbus/dbus-string.c (_dbus_string_free): apply align offset
3590 when freeing the string
3591 (_dbus_string_steal_data): fix for align offset
3592 (undo_alignment): new function
3594 2003-02-26 Havoc Pennington <hp@redhat.com>
3596 All kinds of audit fixes from Owen, plus initial attempt to
3597 handle unaligned memory returned from malloc.
3599 * dbus/dbus-string.c (_dbus_string_init): clamp max length to
3600 leave room for align_offset and nul byte
3601 (fixup_alignment): function to track an align_offset and
3602 ensure real->str is aligned
3603 (DBUS_GENERIC_STRING_PREAMBLE): len must be less than allocated,
3604 to allow a nul byte plus align offset
3605 (_dbus_string_lock): fix overflow issue
3606 (_dbus_string_init_const_len): add assertions on sanity of len,
3607 assign allocated to be ALLOCATION_PADDING larger than len
3608 (set_length): fixup the overflow handling
3609 (_dbus_string_get_data_len): fix overflow in assertion
3610 (open_gap): detect overflow in size of gap to be opened
3611 (_dbus_string_lengthen): add overflow check
3612 (_dbus_string_align_length): fix overflow with _DBUS_ALIGN_VALUE
3613 (_dbus_string_append): add overflow check
3614 (_dbus_string_append_unichar): overflow
3615 (_dbus_string_delete): fix overflow in assertion
3616 (_dbus_string_copy_len): overflow in assertion
3617 (_dbus_string_replace_len): overflows in assertions
3618 (_dbus_string_find): change to implement in terms of
3619 _dbus_string_find_to
3620 (_dbus_string_find_to): assorted fixage
3621 (_dbus_string_equal_c_str): assert c_str != NULL,
3622 fix logic so the function works
3623 (_dbus_string_ends_with_c_str): fix overflow thingy
3624 (_dbus_string_base64_encode): overflow fix
3625 (_dbus_string_validate_ascii): overflow
3626 (_dbus_string_validate_nul): overflow
3628 2003-02-26 Havoc Pennington <hp@redhat.com>
3630 * dbus/dbus-marshal.c (_dbus_marshal_test): fix to work with DISABLE_ASSERTS
3632 2003-02-26 Alexander Larsson <alexl@redhat.com>
3635 Set DBUS_GLIB_THREADS_LIBS for apps using gthread-2.0
3637 * dbus/dbus-connection.c:
3638 * dbus/dbus-connection.h:
3639 Fix _dbus_connection_acquire_io_path and _dbus_connection_acquire_dispatch.
3640 Add dbus_connection_set_wakeup_main_function and use it when queueing
3641 incoming and outgoing messages.
3644 * dbus/dbus-dataslot.c:
3645 Threadsafe usage of DBusDataSlotAllocator
3647 * dbus/dbus-message.c: (dbus_message_get_args_iter):
3650 * dbus/dbus-server-unix.c:
3653 * glib/dbus-gmain.c:
3654 Implement the new wakeup functions for glib.
3657 * glib/test-thread-client.c:
3658 * glib/test-thread-server.c:
3659 * glib/test-thread.h:
3660 Initial cut at some thread test code. Not really done yet.
3662 2003-02-26 Havoc Pennington <hp@pobox.com>
3664 * dbus/dbus-connection.c
3665 (dbus_connection_send_message_with_reply_and_block): fix crash
3666 where we ref'd the outgoing message instead of the returned reply
3668 * dbus/dbus-transport-unix.c (do_authentication): check read watch
3669 at the end of this function, so if we didn't need to read for
3670 authentication, we reinstall it for receiving messages
3672 * dbus/dbus-message.c (dbus_message_new_reply): allow replies to
3673 a NULL sender for peer-to-peer case
3675 * dbus/dbus-transport-unix.c (check_read_watch): handle
3676 !authenticated case correctly
3678 * glib/dbus-gmain.c: add support for DBusServer
3680 * dbus/dbus-server.c: add data slot support
3682 * glib/dbus-gmain.c (dbus_connection_setup_with_g_main): check
3683 return values and handle errors
3685 * dbus/dbus-dataslot.c: factor out the data slot stuff from
3688 * Doxyfile.in (INPUT): add glib subdir
3690 * glib/dbus-gmain.c (dbus_connection_setup_with_g_main): rename
3691 setup_with_g_main instead of hookup_with_g_main; write docs
3693 2003-02-24 Anders Carlsson <andersca@codefactory.se>
3695 * dbus/dbus-marshal.c: (_dbus_marshal_validate_arg):
3696 * dbus/dbus-message-builder.c: (_dbus_message_data_load):
3697 * dbus/dbus-message.c: (dbus_message_append_boolean),
3698 (dbus_message_append_boolean_array),
3699 (dbus_message_get_args_valist), (_dbus_message_test):
3700 * dbus/dbus-message.h:
3701 * doc/dbus-specification.sgml:
3702 Various fixes as pointed out by Havoc.
3704 * test/data/invalid-messages/bad-boolean-array.message:
3705 * test/data/invalid-messages/bad-boolean.message:
3706 Add invalid boolean value test cases.
3708 2003-02-24 Anders Carlsson <andersca@codefactory.se>
3710 * dbus/dbus-internals.c: (_dbus_type_to_string):
3711 * dbus/dbus-marshal.c: (_dbus_marshal_get_arg_end_pos),
3712 (_dbus_marshal_validate_arg):
3713 * dbus/dbus-message-builder.c: (_dbus_message_data_load):
3714 * dbus/dbus-message.c: (dbus_message_append_args_valist),
3715 (dbus_message_append_boolean), (dbus_message_append_boolean_array),
3716 (dbus_message_get_args_valist), (dbus_message_iter_get_boolean),
3717 (dbus_message_iter_get_int32), (dbus_message_iter_get_uint32),
3718 (dbus_message_iter_get_double),
3719 (dbus_message_iter_get_boolean_array), (message_iter_test):
3720 * dbus/dbus-message.h:
3721 * dbus/dbus-protocol.h:
3722 * doc/dbus-specification.sgml:
3723 * test/data/valid-messages/lots-of-arguments.message:
3724 Add support for boolean and boolean array types.
3726 2003-02-23 Havoc Pennington <hp@pobox.com>
3728 * dbus/dbus-keyring.c: finish most of this implementation and
3731 * dbus/dbus-errors.c (dbus_set_error_const, dbus_set_error): make
3732 these barf if the error isn't cleared to NULL
3734 * dbus/dbus-sysdeps.c (_dbus_delete_file): set error on failure
3735 (_dbus_create_directory): new function
3737 * dbus/dbus-errors.c (dbus_set_error): fix warning
3739 * dbus/dbus-string.c (_dbus_string_hex_encode): new function
3740 (_dbus_string_hex_decode): new function
3741 (test_hex_roundtrip): test code
3743 * dbus/dbus-sha.c (_dbus_sha_compute): use dbus_string_hex_encode
3745 * dbus/dbus-md5.c (_dbus_md5_compute): use dbus_string_hex_encode
3747 * dbus/dbus-sysdeps.c (_dbus_string_save_to_file): make this use
3748 the save-to-temp/rename trick to atomically write the new file
3749 (_dbus_string_parse_uint): new function
3751 2003-02-22 Havoc Pennington <hp@pobox.com>
3753 * test/Makefile.am (dist-hook): fix dist for test/data/sha-1
3755 2003-02-22 Havoc Pennington <hp@pobox.com>
3757 * dbus/dbus-message.c (dbus_message_iter_get_string_array):
3758 (dbus_message_iter_get_byte_array): Fix up doxygen warnings
3760 * dbus/dbus-sha.c: add implementation of SHA-1 algorithm
3762 * dbus/test/data/sha-1: add US government test suite for SHA-1
3764 2003-02-21 Anders Carlsson <andersca@codefactory.se>
3766 * dbus/dbus-marshal.c: (_dbus_demarshal_string_array):
3767 Make string arrays NULL-terminated.
3769 * dbus/dbus-memory.c: (dbus_free_string_array):
3770 * dbus/dbus-memory.h:
3771 New function for freeing NULL-terminated string arrays.
3773 * dbus/dbus-message-builder.c: (append_quoted_string),
3774 (_dbus_message_data_load):
3775 Add support for array types.
3777 * dbus/dbus-message.c: (check_message_handling):
3778 Add more types as test cases.
3780 * dbus/dbus-sysdeps.c: (_dbus_string_parse_int),
3781 (_dbus_string_parse_double):
3782 Add the start offset to the end offset.
3784 * test/data/valid-messages/lots-of-arguments.message:
3785 New test message with lots of arguments.
3787 2003-02-21 Anders Carlsson <andersca@codefactory.se>
3789 * dbus/dbus-message.c: (dbus_message_append_nil),
3790 (dbus_message_append_int32), (dbus_message_append_uint32),
3791 (dbus_message_append_double), (dbus_message_append_string),
3792 (dbus_message_append_int32_array),
3793 (dbus_message_append_uint32_array),
3794 (dbus_message_append_double_array),
3795 (dbus_message_append_byte_array),
3796 (dbus_message_append_string_array):
3797 Fix all out-of-memory handling in these functions.
3799 2003-02-21 Anders Carlsson <andersca@codefactory.se>
3801 * dbus/dbus-message.c: (dbus_message_append_nil):
3804 2003-02-21 Anders Carlsson <andersca@codefactory.se>
3806 * dbus/dbus-message.c: (dbus_message_append_args_valist),
3807 (dbus_message_append_nil), (dbus_message_append_int32_array),
3808 (dbus_message_append_uint32_array),
3809 (dbus_message_append_double_array),
3810 (dbus_message_append_byte_array),
3811 (dbus_message_append_string_array), (dbus_message_get_args_valist),
3812 (dbus_message_iter_get_int32_array),
3813 (dbus_message_iter_get_uint32_array),
3814 (dbus_message_iter_get_double_array),
3815 (dbus_message_iter_get_byte_array),
3816 (dbus_message_iter_get_string_array):
3818 * dbus/dbus-message.h:
3819 Add functions for appending and getting arrays.
3821 2003-02-21 Anders Carlsson <andersca@codefactory.se>
3823 * dbus/dbus-mempool.c (_dbus_mem_pool_new): Make the
3824 element size at least 8 bytes, fixes mempool tests on
3827 2003-02-20 Alexander Larsson <alexl@redhat.com>
3829 * dbus/dbus-transport-unix.c (unix_do_iteration):
3830 Unlock the connection mutex during a blocking select call.
3831 Add todo about how we need a way to wake up the select.
3833 * dbus/dbus-connection-internal.h:
3834 * dbus/dbus-connection.c:
3835 Add _dbus_connection_lock and _dbus_connection_unlock.
3837 2003-02-19 Havoc Pennington <hp@pobox.com>
3839 * Doxyfile.in (PREDEFINED): put DOXYGEN_SHOULD_SKIP_THIS in
3840 Doxyfile.in, not Doxyfile
3842 * dbus/dbus-keyring.c: do some hacking on this
3844 * dbus/dbus-sysdeps.c (_dbus_delete_file): new
3846 * dbus/dbus-errors.c (dbus_set_error_const): do not call
3848 (dbus_set_error): remove dbus_error_init, check for message ==
3849 NULL *before* we sprintf into it, and add @todo about including
3850 system headers in this file
3852 * dbus/dbus-sysdeps.c (_dbus_create_file_exclusively): new
3854 * dbus/dbus-errors.h (DBUS_ERROR_FAILED): add
3856 * dbus/dbus-sysdeps.c (get_user_info): break this function out to
3857 get various bits of user information based on either username
3859 (_dbus_homedir_from_username): new function
3861 2003-02-19 Anders Carlsson <andersca@codefactory.se>
3864 Add check for nonposix getpwnam_r
3866 * dbus/dbus-mempool.c: (_dbus_mem_pool_new):
3867 Align the pool element size to a sizeof (void *) boundary.
3869 * dbus/dbus-sysdeps.c: (_dbus_setenv), (_dbus_connect_unix_socket),
3870 (_dbus_listen_unix_socket), (_dbus_credentials_from_username):
3871 General Solaris fixes.
3873 * test/data/valid-messages/simplest-manual.message:
3874 Explicitly state that we want little-endian packing.
3876 2003-02-19 Mikael Hallendal <micke@codefactory.se>
3878 * dbus/dbus-server.c (dbus_server_listen): Support tcp: addresses.
3880 * dbus/dbus-transport-unix.c (_dbus_transport_new_for_tcp_socket):
3881 Added to create a transport connecting using a tcp/ip socket.
3883 * dbus/dbus-sysdeps.c (_dbus_connect_tcp_socket): Added to connect
3884 to a tcp socket at given host and port.
3885 (_dbus_listen_tcp_socket): added to listen on tcp socket for given
3888 * dbus/dbus-server.c (dbus_server_listen): Support tcp: addresses.
3890 * dbus/dbus-server-unix.c (_dbus_server_new_for_tcp_socket):
3891 Added to create a server listening on a TCP/IP socket.
3893 2003-02-19 Havoc Pennington <hp@pobox.com>
3895 Throughout: mop up all the Doxygen warnings and undocumented
3898 * dbus/dbus-sysdeps.c (do_exec): do not use execvp, we don't want
3899 to search any paths.
3901 * dbus/dbus-threads.c: move global mutex initializers to
3902 dbus-internals.h, multiple prototypes was confusing doxygen
3903 besides being kind of ugly
3905 * Doxyfile (PREDEFINED): have Doxygen define
3906 DOXYGEN_SHOULD_SKIP_THIS so we can exclude things from
3907 docs with #ifndef DOXYGEN_SHOULD_SKIP_THIS
3908 (do not abuse the feature! it's for stuff like the autogenerated
3909 macros in dbus-md5.c, not just for things you don't feel like
3912 2003-02-18 Havoc Pennington <hp@pobox.com>
3914 * dbus/dbus-string.c (_dbus_string_zero): new function
3916 * dbus/dbus-md5.c: include MD5 implementation by L. Peter Deutsch,
3917 wrap it in some dbus-friendly API
3919 * dbus/dbus-types.h: add 16-bit types
3921 2003-02-18 Joe Shaw <joe@assbarn.com>
3923 * dbus/dbus-auth.c (handle_server_data_stupid_test_mech): Just get
3924 credentials from our currently running process.
3925 (get_word): Fix a buglet where we were copying the entire length
3926 instead of relative to our position.
3928 * dbus/dbus-hash.c (_dbus_hash_test): Don't try to allocate the
3929 keys on the stack... it's 640k of data.
3931 * dbus/dbus-sysdeps.c (_dbus_read_credentials_unix_socket): Always
3932 read the credentials byte off the socket, even if we don't have
3934 (_dbus_poll): Implement poll() using select() for systems which
3937 * glib/test-dbus-glib.c (main): Print out an error if no
3938 parameters are given.
3940 * test/data/auth/fallback.auth-script: Added. Tests that a client
3941 can fallback to a secondary auth mechanism if the first fails.
3943 2003-02-18 Havoc Pennington <hp@pobox.com>
3947 2003-02-17 Havoc Pennington <hp@pobox.com>
3949 * doc/dbus-specification.sgml: lots of cosmetic
3950 cleanups/rearrangement, add assorted FIXME, change DBUS_ADDRESS
3951 env variable to DBUS_BUS_ADDRESS, s/client/application/,
3952 s/server/bus/ (except in authentication section). Add a section
3953 "Message Bus Message Routing"
3955 2003-02-17 Anders Carlsson <andersca@codefactory.se.>
3961 2003-02-17 Anders Carlsson <andersca@codefactory.se>
3963 * doc/dbus-specification.sgml:
3964 Specification updates.
3966 2003-02-17 Anders Carlsson <andersca@codefactory.se>
3968 * bus/activation.c: (bus_activation_init), (child_setup),
3969 (bus_activation_activate_service):
3971 * bus/main.c: (main):
3972 Set DBUS_ADDRESS environment variable.
3974 * dbus/dbus-errors.c: (dbus_set_error):
3975 Don't use va_copy since that's a C99 feature.
3977 * dbus/dbus-sysdeps.c: (_dbus_setenv), (do_exec),
3978 (_dbus_spawn_async):
3979 * dbus/dbus-sysdeps.h:
3980 Add child_setup_func to _dbus_spawn_async.
3982 * doc/dbus-specification.sgml:
3983 Update specification.
3985 * test/spawn-test.c: (setup_func), (main):
3988 2003-02-17 Alexander Larsson <alexl@redhat.com>
3990 * dbus/dbus-connection.c (_dbus_connection_handler_destroyed_locked):
3993 2003-02-17 Anders Carlsson <andersca@codefactory.se>
3997 * doc/dbus-test-plan.sgml:
3998 Add test plan document.
4003 2003-02-17 Anders Carlsson <andersca@codefactory.se>
4005 * dbus/dbus-message.c: (decode_header_data),
4006 (_dbus_message_loader_return_buffer):
4007 Set the header padding amount when loading a message.
4009 2003-02-16 Anders Carlsson <andersca@codefactory.se>
4011 * bus/dispatch.c: (send_one_message):
4012 Only send broadcast messages to registered connections.
4014 * dbus/dbus-message.c: (dbus_message_name_is):
4015 * dbus/dbus-message.h:
4016 New convenience function.
4018 * dbus/dbus-transport-debug.c: (do_reading):
4019 Only dispatch one message per run.
4022 * test/bus-test.c: (new_connection_callback), (die),
4023 (test_hello_client1_handler), (test_hello_client2_handler),
4024 (test_hello_replies), (main):
4026 * test/bus-test-loop.[ch]:
4029 2003-02-16 Havoc Pennington <hp@pobox.com>
4031 * dbus/dbus-connection.c (dbus_connection_dispatch_message): fix
4032 backward conditional
4034 2003-02-16 Alexander Larsson <alexl@redhat.com>
4036 * dbus/dbus-connection.c:
4037 Implement sent_message_with_reply. (with_reply_and block is still
4039 Made dispatch_message not lose message if OOM.
4041 * dbus/dbus-errors.h:
4042 Add NoReply error (for reply timeouts).
4044 2003-02-16 Alexander Larsson <alexl@redhat.com>
4046 * dbus/dbus-hash.c (_dbus_hash_table_unref):
4047 Actually free keys and values when destroying hashtable.
4049 2003-02-16 Anders Carlsson <andersca@codefactory.se>
4051 * dbus/dbus-auth.c: (client_try_next_mechanism):
4054 * dbus/dbus-threads.c: (dbus_condvar_wait_timeout):
4055 Return TRUE if there's no thread implementation around.
4057 * glib/dbus-gmain.c: (free_source),
4058 (dbus_connection_hookup_with_g_main):
4059 Make sure to remove the GSource when the connection is finalized.
4061 2003-02-16 Anders Carlsson <andersca@codefactory.se>
4063 * bus/dispatch.c: (bus_dispatch_message_handler):
4064 * dbus/dbus-errors.h:
4065 Return an error if someone tries to send a message to a service
4068 2003-02-16 Anders Carlsson <andersca@codefactory.se>
4070 * bus/activation.c: (load_directory), (bus_activation_init),
4071 (bus_activation_activate_service):
4074 (bus_driver_handle_activate_service), (bus_driver_handle_message):
4075 More work on the activation handling.
4077 * dbus/dbus-errors.h:
4078 Add some error messages
4080 * dbus/dbus-message.c: (dbus_message_new_error_reply):
4081 * dbus/dbus-message.h:
4082 New function that creates an error message.
4084 * dbus/dbus-protocol.h:
4085 Add ACTIVATE_SERVER message.
4087 * dbus/dbus-server-unix.c: (unix_handle_watch),
4088 (_dbus_server_new_for_domain_socket):
4089 Call _dbus_fd_set_close_on_exec.
4091 * dbus/dbus-sysdeps.c: (make_pipe), (do_exec),
4092 (_dbus_spawn_async), (_dbus_disable_sigpipe),
4093 (_dbus_fd_set_close_on_exec):
4094 * dbus/dbus-sysdeps.h:
4095 Add _dbus_fd_set_close_on exec function. Also add function that checks
4096 that all open fds are set to close-on-exec and warns otherwise.
4098 * dbus/dbus-transport-unix.c:
4099 (_dbus_transport_new_for_domain_socket):
4100 Call _dbus_fd_set_close_on_exec.
4102 2003-02-16 Havoc Pennington <hp@pobox.com>
4104 * dbus/dbus-connection.c (dbus_connection_set_change_sigpipe):
4105 allow people to avoid setting SIGPIPE to SIG_IGN
4106 (_dbus_connection_new_for_transport): disable SIGPIPE unless
4107 we've been asked not to
4109 2003-02-15 Anders Carlsson <andersca@codefactory.se>
4111 * dbus/dbus-list.c: (_dbus_list_append_link),
4112 (_dbus_list_prepend_link):
4113 * dbus/dbus-memory.c: (dbus_malloc), (dbus_malloc0),
4117 2003-02-15 Anders Carlsson <andersca@codefactory.se>
4120 * bus/activation.c: (bus_activation_entry_free),
4121 (add_desktop_file_entry), (load_directory), (bus_activation_init):
4123 * bus/main.c: (main):
4124 Add simple activation support, doesn't work yet though.
4126 2003-02-15 Zack Rusin <zack@kde.org>
4128 * qt/dbus-qthread.cpp: small casting fix
4130 2003-02-15 Anders Carlsson <andersca@codefactory.se>
4132 * dbus/dbus-errors.c: (dbus_set_error):
4133 * dbus/dbus-errors.h:
4134 Add a few errors and make dbus_set_error void.
4136 * dbus/dbus-sysdeps.c:
4137 (_dbus_errno_to_string), (close_and_invalidate), (make_pipe),
4138 (write_err_and_exit), (read_ints), (do_exec), (_dbus_spawn_async):
4139 * dbus/dbus-sysdeps.h:
4140 Add _dbus_spawn_async.
4142 * test/spawn-test.c: (main):
4143 Test for _dbus_spawn_async.
4145 2003-02-15 Anders Carlsson <andersca@codefactory.se>
4147 * dbus/dbus-internals.h:
4148 Fix build without tests.
4150 * dbus/dbus-list.c: (alloc_link):
4151 Fix a segfault when a malloc fails.
4153 * dbus/dbus-memory.c: (initialize_malloc_debug), (dbus_malloc),
4154 (dbus_malloc0), (dbus_realloc):
4155 Add support for malloc debugging.
4157 2003-02-15 Alexander Larsson <alexl@redhat.com>
4159 * dbus/dbus-threads.c:
4160 * dbus/dbus-threads.h:
4161 Add condvars. Remove static mutext from API.
4162 Implement static mutexes by initializing them from threads_init.
4164 * glib/dbus-gthread.c:
4165 * qt/dbus-qthread.cpp:
4166 Update with the thread api changes.
4171 Turn StaticMutex into normal mutex + init function.
4172 Export new functions _dbus_list_alloc_link, _dbus_list_free_link,
4173 _dbus_list_append_link, _dbus_list_prepend_link
4176 * dbus/dbus-sysdeps.c:
4177 * dbus/dbus-sysdeps.h:
4178 New type dbus_atomic_t, and new functions _dbus_atomic_inc,
4179 _dbus_atomic_dec. Only slow fallback implementation at the moment.
4181 * dbus/dbus-protocol.h:
4182 Add DBUS_MESSAGE_LOCAL_DISCONNECT define
4184 * dbus/dbus-message.c:
4185 Make ref/unref atomic.
4188 * dbus/dbus-connection-internal.h:
4189 * dbus/dbus-connection.c:
4190 * dbus/dbus-connection.h:
4192 Change _peek to _borrow,_return & _steal_borrowed.
4193 Change disconnect callback to event.
4194 Make dbus_connection_dispatch_messages reentrant.
4196 * dbus/dbus-transport.c:
4197 Don't ref the connection on calls to the transport
4200 * dbus/dbus-message-handler.c:
4203 * glib/dbus-gmain.c:
4204 Don't use peek_message anymore
4207 * test/debug-thread.c:
4208 * test/debug-thread.h:
4209 Simple thread implementation that asserts() on deadlocks in
4210 single-threaded code.
4213 (main) Call debug_threads_init.
4216 Use disconnect message instead of disconnect callback.
4220 Don't call dbus_connection_set_disconnect_function. Instead export
4221 bus_connection_disconnect.
4224 Call bus_connection_disconnect when we get a disconnected message.
4226 2003-02-15 Havoc Pennington <hp@pobox.com>
4228 * dbus/dbus-message.c (dbus_message_new): fool around with the
4231 2003-02-14 Havoc Pennington <hp@pobox.com>
4233 * dbus/dbus-mempool.c: fail if the debug functions so indicate
4235 * dbus/dbus-memory.c: fail if the debug functions indicate we
4238 * dbus/dbus-internals.c (_dbus_set_fail_alloc_counter)
4239 (_dbus_decrement_fail_alloc_counter): debug functions to
4240 simulate memory allocation failures
4242 2003-02-14 Havoc Pennington <hp@pobox.com>
4244 * dbus/dbus-errors.h (struct DBusError): add a word of padding
4247 2003-02-13 Anders Carlsson <andersca@codefactory.se>
4249 * bus/driver.c: (bus_driver_handle_hello):
4251 * bus/services.c: (bus_service_lookup):
4252 Reorder message sending so we get a more sane order.
4254 * test/bus-test.c: (message_handler):
4257 2003-02-13 Anders Carlsson <andersca@codefactory.se>
4259 * bus/driver.c: (bus_driver_send_service_deleted),
4260 (bus_driver_send_service_created), (bus_driver_send_service_lost),
4261 (bus_driver_send_service_acquired), (bus_driver_handle_hello),
4262 (bus_driver_send_welcome_message),
4263 (bus_driver_handle_list_services),
4264 (bus_driver_handle_acquire_service),
4265 (bus_driver_handle_service_exists):
4266 * dbus/dbus-bus.c: (dbus_bus_register_client),
4267 (dbus_bus_acquire_service), (dbus_bus_service_exists):
4268 * dbus/dbus-errors.c: (dbus_result_to_string):
4269 * dbus/dbus-errors.h:
4270 * dbus/dbus-message.c: (dbus_message_append_args),
4271 (dbus_message_append_args_valist), (dbus_message_get_args),
4272 (dbus_message_get_args_valist), (dbus_message_get_args_iter),
4273 (dbus_message_iter_get_arg_type), (dbus_message_iter_get_string),
4274 (dbus_message_iter_get_byte_array),
4275 (dbus_message_iter_get_string_array), (message_iter_test),
4276 (check_message_handling), (_dbus_message_test):
4277 * dbus/dbus-message.h:
4278 * test/bus-test.c: (main):
4279 Change fields to arguments in messages, so that they won't be
4280 confused with header fields.
4282 * glib/test-dbus-glib.c: (main):
4283 Remove append_fields from hello message.
4285 2003-02-13 Anders Carlsson <andersca@codefactory.se>
4287 * dbus/dbus-errors.c:
4288 * dbus/dbus-message.c:
4289 * dbus/dbus-string.c:
4290 Documentation fixes.
4292 2003-02-13 Anders Carlsson <andersca@codefactory.se>
4294 * glib/dbus-gmain.c: (timeout_handler), (add_timeout),
4296 Implement support for timeouts in dbus-glib.
4298 2003-02-13 Anders Carlsson <andersca@codefactory.se>
4300 * dbus/dbus-message-builder.c: (_dbus_message_data_load):
4301 * dbus/dbus-message.c: (process_test_subdir):
4302 * test/break-loader.c: (find_breaks_based_on):
4303 Plug some memory leaks.
4305 2003-02-13 Richard Hult <rhult@codefactory.se>
4307 * bus/main.c: Fix build.
4309 * dbus/dbus-errors.h:
4310 * dbus/dbus-errors.c: Fix copyright for Anders.
4312 2003-02-13 Anders Carlsson <andersca@codefactory.se>
4317 * bus/connection.c: (bus_connection_foreach):
4320 * bus/desktop-file.c: (grow_lines_in_section), (grow_sections),
4321 (unescape_string), (new_section), (parse_section_start),
4322 (parse_key_value), (report_error), (bus_desktop_file_load),
4323 (bus_desktop_file_get_string):
4324 * bus/desktop-file.h:
4325 Use DBusError for error reporting.
4327 * bus/dispatch.c: (send_one_message),
4328 (bus_dispatch_message_handler):
4329 * bus/driver.c: (bus_driver_send_service_deleted),
4330 (bus_driver_send_service_created), (bus_driver_send_service_lost),
4331 (bus_driver_send_service_acquired), (bus_driver_handle_hello),
4332 (bus_driver_send_welcome_message),
4333 (bus_driver_handle_list_services),
4334 (bus_driver_handle_acquire_service),
4335 (bus_driver_handle_service_exists):
4336 * bus/loop.c: (bus_loop_run):
4338 Use BUS_HANDLE_OOM instead of _DBUS_HANDLE_OOM.
4340 * bus/utils.c: (bus_wait_for_memory):
4342 New files with general utility functions.
4344 * dbus/dbus-internals.h:
4345 Remove _DBUS_HANDLE_OOM.
4347 2003-02-13 Anders Carlsson <andersca@codefactory.se>
4349 * dbus/dbus-errors.c: (dbus_result_to_string), (dbus_error_init),
4350 (dbus_error_free), (dbus_set_error_const), (dbus_set_error):
4351 * dbus/dbus-errors.h:
4352 Add DBusError structure.
4354 2003-02-13 Anders Carlsson <andersca@codefactory.se>
4356 * test/data/valid-messages/standard-acquire-service.message:
4357 * test/data/valid-messages/standard-hello.message:
4358 * test/data/valid-messages/standard-list-services.message:
4359 * test/data/valid-messages/standard-service-exists.message:
4360 Add some standard messages.
4362 2003-02-13 Anders Carlsson <andersca@codefactory.se>
4364 * bus/driver.c: (bus_driver_send_welcome_message),
4365 (bus_driver_handle_list_services),
4366 (bus_driver_handle_acquire_service),
4367 (bus_driver_handle_service_exists), (bus_driver_handle_message):
4368 Update for API changes in libdbus.
4370 * dbus/dbus-message.c: (dbus_message_new_reply):
4371 * dbus/dbus-message.h:
4372 Remove the name argument. The spec states that replies shouldn't
4375 2003-02-13 Anders Carlsson <andersca@codefactory.se>
4377 * bus/desktop-file.c: (parse_section_start), (parse_key_value),
4378 (report_error), (bus_desktop_file_load), (lookup_section),
4379 (lookup_line), (bus_desktop_file_get_raw),
4380 (bus_desktop_file_get_string):
4381 * bus/desktop-file.h:
4382 Some fixes, and new functions for getting a key value from a section.
4384 2003-02-13 Havoc Pennington <hp@pobox.com>
4386 * test/data/auth/fail-after-n-attempts.auth-script: new test
4388 * dbus/dbus-auth.c (send_rejected): shutdown_mech() when we
4391 2003-02-13 Havoc Pennington <hp@pobox.com>
4393 * dbus/dbus-auth.c (handle_server_data_external_mech): args to
4394 dbus_credentials_match were backward
4396 * dbus/dbus-auth-script.c (_dbus_auth_script_run): support
4397 NO_CREDENTIALS and ROOT_CREDENTIALS
4399 * dbus/dbus-auth.c (_dbus_auth_do_work): move get_state() routine
4400 into here. Never process more commands after we've reached an
4401 end state; store further data as unused bytes.
4403 * test/data/auth/*: add more auth tests
4405 * dbus/dbus-auth-script.c (_dbus_auth_script_run): support EXPECT
4406 command to match exact string and EXPECT_UNUSED to match unused
4409 * test/Makefile.am (dist-hook): fix to dist all the test stuff
4411 2003-02-12 Havoc Pennington <hp@pobox.com>
4413 * dbus/dbus-string.c (_dbus_string_pop_line): fix to also strip
4414 \r off of popped lines
4416 * dbus/dbus-auth.c (_dbus_auth_test): write code to run auth
4419 * dbus/dbus-auth-script.c (_dbus_auth_script_run): when doing a
4422 2003-02-12 Havoc Pennington <hp@pobox.com>
4424 * dbus/Makefile.am: remove break-loader from the build, since it
4427 * configure.in: add --enable-gcov to turn on coverage profiling
4428 flags and disable optimization
4430 2003-02-10 Havoc Pennington <hp@pobox.com>
4432 * dbus/dbus-auth-script.c, dbus/dbus-auth-script.h: sync
4433 initial cut at test framework for DBusAuth from laptop.
4434 Doesn't quite work yet but it compiles and I need to get
4435 it off the 266mhz laptop. ;-)
4437 * dbus/dbus-server-debug.c (_dbus_server_debug_accept_transport):
4438 fix a memleak in error case
4440 2003-02-12 Anders Carlsson <andersca@codefactory.se>
4443 * bus/desktop-file.c:
4444 * bus/desktop-file.h:
4445 Add a desktop file parser.
4447 2003-02-11 Zack Rusin <zack@kde.org>
4449 * qt/message.[h|cpp]: sample implementation
4450 of the KDE wrapper for DBusMessage
4452 2003-02-09 Zack Rusin <zack@kde.org>
4454 * test/bus-test.c: make_it_compile
4455 * doc/dbus-specification.sgml: minimal semantic fix
4457 2003-02-06 Anders Carlsson <andersca@codefactory.se>
4463 2003-02-06 Anders Carlsson <andersca@codefactory.se>
4466 * dbus/dbus-break-loader.c:
4468 * test/break-loader.c:
4469 Move dbus-break-loader to test/ and rename it to break-loader.
4471 2003-02-02 Havoc Pennington <hp@pobox.com>
4473 * dbus/dbus-keyring.c, dbus/dbus-keyring.h: template files
4474 for code to manage cookies in your home directory
4476 * dbus/dbus-sysdeps.c (_dbus_generate_random_bytes): new function
4478 * dbus/dbus-auth.c (get_state): impose a maximum number of tries
4479 to authenticate, then disconnect the client.
4481 2003-02-03 Alexander Larsson <alexl@redhat.com>
4483 * dbus/dbus-message.c (dbus_message_append_fields):
4486 2003-02-02 Anders Carlsson <andersca@codefactory.se>
4488 * doc/dbus-specification.sgml:
4489 Update address format section.
4491 2003-02-02 Anders Carlsson <andersca@codefactory.se>
4494 * test/bus-test.c: (get_time), (add_timeout), (remove_timeout),
4495 (message_handler), (new_connection_callback), (loop_quit),
4499 2003-02-02 Anders Carlsson <andersca@codefactory.se>
4501 * bus/driver.c: (bus_driver_handle_service_exists):
4502 Simplify the code a bit.
4504 * dbus/dbus-bus.c: (dbus_bus_service_exists):
4507 2003-02-02 Anders Carlsson <andersca@codefactory.se>
4510 Add libdbus-daemon.la and link to it.
4512 2003-02-01 James Willcox <jwillcox@gnome.org>
4514 * bus/driver.c: (bus_driver_handle_own_service):
4515 Actually include the service reply code in the message.
4517 2003-02-02 Anders Carlsson <andersca@codefactory.se>
4519 * bus/driver.c: (bus_driver_handle_service_exists):
4520 Don't unref the incoming message.
4522 2003-02-02 Anders Carlsson <andersca@codefactory.se>
4524 * dbus/dbus.h: Add dbus-address.h and dbus-bus.h
4526 2003-02-02 Anders Carlsson <andersca@codefactory.se>
4528 * dbus/dbus-server.c: (dbus_server_listen):
4529 * dbus/dbus-transport.c: (_dbus_transport_open):
4530 ifdef out the calls to the debug transport and server.
4532 2003-02-02 Alexander Larsson <alexl@redhat.com>
4534 * dbus/dbus-watch.c (dbus_watch_get_flags):
4535 Add note in the docs that ERROR or HANGUP won't be returned
4536 and are assumed always on.
4538 * glib/dbus-gmain.c (add_watch):
4539 Always add IO_ERR | IO_HUP
4541 * dbus/dbus-message.h:
4542 Add semicolon after dbus_message_iter_get_string_array().
4543 Makes qt code build again
4545 2003-02-01 Anders Carlsson <andersca@codefactory.se>
4547 * bus/driver.c: (create_unique_client_name),
4548 (bus_driver_handle_hello):
4549 Don't take a name, just use a numeric id to identify
4553 * dbus/dbus-bus.c: (dbus_bus_register_client),
4554 (dbus_bus_acquire_service), (dbus_bus_service_exists):
4556 Add new convenience functions for communicating with the bus.
4558 * dbus/dbus-message.h:
4560 * dbus/dbus-protocol.h:
4563 2003-02-01 Alexander Larsson <alexl@redhat.com>
4565 * dbus/dbus-message.c (dbus_message_append_fields):
4566 Add some more doc comments.
4568 2003-02-01 Havoc Pennington <hp@pobox.com>
4570 * dbus/dbus-break-loader.c (randomly_modify_length): change
4571 a 4-byte value in the message as if it were a length
4573 * dbus/dbus-sysdeps.c (_dbus_string_save_to_file): don't set
4574 execute bit on saved files
4576 2003-02-01 Havoc Pennington <hp@pobox.com>
4578 * dbus/dbus-break-loader.c (main): new program to find messages
4579 that break the loader.
4581 * dbus/dbus-sysdeps.c (_dbus_string_append_uint): new function
4582 * dbus/dbus-sysdeps.c (_dbus_string_save_to_file): new function
4584 * dbus/dbus-string.c (_dbus_string_set_byte): new
4586 2003-01-31 Havoc Pennington <hp@pobox.com>
4588 * dbus/dbus-message.c: refactor the test code to be more general,
4589 in preparation for writing a "randomly permute test cases to
4590 try to break the loader" program.
4592 2003-01-31 Havoc Pennington <hp@pobox.com>
4594 * doc/dbus-specification.sgml: work on the specification
4596 * dbus/dbus-message.c (_dbus_message_loader_return_buffer): check
4597 the protocol version of the message.
4599 * dbus/dbus-protocol.h: drop special _REPLY names, the spec
4600 no longer specifies that.
4601 (DBUS_SERVICE_REPLY_SERVICE_EXISTS): fix flags (1/2/4/8 not
4604 * dbus/dbus-marshal.c (_dbus_marshal_get_arg_end_pos): add missing
4605 "break" for DBUS_TYPE_NIL, remove @todo
4607 2003-01-31 Havoc Pennington <hp@pobox.com>
4609 * dbus/dbus-message.c (dbus_message_set_is_error_reply): rename
4610 just set_is_error/get_is_error as this is a commonly-used
4611 function, and write docs.
4613 2003-01-31 Anders Carlsson <andersca@codefactory.se>
4615 * dbus/dbus-address.c: (dbus_address_entry_free):
4616 Free key and value lists.
4618 * dbus/dbus-internals.c: (_dbus_type_to_string):
4619 Add the types we didn't have.
4621 * dbus/dbus-marshal.c: (_dbus_marshal_get_arg_end_pos),
4622 (_dbus_marshal_validate_arg):
4625 * dbus/dbus-message.c: (dbus_message_set_sender):
4626 Remove todo about being able to set sender to NULL.
4628 (dbus_message_set_is_error_reply),
4629 (dbus_message_get_is_error_reply):
4630 * dbus/dbus-message.h:
4633 * dbus/dbus-protocol.h:
4634 Add error reply flag.
4636 * test/data/valid-messages/opposite-endian.message:
4637 Add NIL type to test.
4639 2003-01-31 Havoc Pennington <hp@pobox.com>
4641 * doc/dbus-specification.sgml: fully specify the header. Add
4642 flags and major protocol version, and change header/body len to
4645 * dbus/dbus-message-builder.c (append_saved_length): append length
4648 * dbus/dbus-message.c (dbus_message_create_header): change header
4649 length and body length to unsigned. Add the new fields from the
4651 (_dbus_message_loader_return_buffer): unsigned header/body len
4653 2003-01-30 Havoc Pennington <hp@pobox.com>
4655 * dbus/dbus-auth.c: rework to use only REJECTED, no
4658 * doc/dbus-sasl-profile.txt: drop MECHANISMS and just
4659 use REJECTED, suggested by Mark McLoughlin
4661 2003-01-30 Havoc Pennington <hp@pobox.com>
4663 * dbus/dbus-server.c (dbus_server_listen): @todo about how we need
4664 a better way to report errors here. e.g. "unix address lacks
4665 path" or something. also "no such file" when the path doesn't
4668 * dbus/dbus-address.c (dbus_address_entries_free): add @todo about
4670 (dbus_parse_address): add @todo about documenting address format,
4671 and allowing , and ; to be escaped
4673 2003-01-30 Anders Carlsson <andersca@codefactory.se>
4676 Add dbus-address.[ch]
4678 * dbus/dbus-address.c: (dbus_address_entry_free),
4679 (dbus_address_entries_free), (create_entry),
4680 (dbus_address_entry_get_method), (dbus_address_entry_get_value),
4681 (dbus_parse_address), (_dbus_address_test):
4682 * dbus/dbus-address.h:
4683 New files for dealing with address parsing.
4685 * dbus/dbus-connection.c:
4686 Document timeout functions.
4688 * dbus/dbus-message.c:
4689 Document dbus_message_new_from_message.
4691 * dbus/dbus-server-debug.c:
4694 * dbus/dbus-server.c: (dbus_server_listen):
4695 Parse address and use correct server implementation.
4697 * dbus/dbus-string.c: (_dbus_string_find_to), (_dbus_string_test):
4698 * dbus/dbus-string.h:
4699 New function with test.
4701 * dbus/dbus-test.c: (dbus_internal_symbol_do_not_use_run_tests):
4705 * dbus/dbus-transport-debug.c:
4708 * dbus/dbus-transport.c: (_dbus_transport_open):
4709 Parse address and use correct transport implementation.
4711 2003-01-30 Havoc Pennington <hp@pobox.com>
4713 * dbus/dbus-message.c: use message->byte_order instead of
4714 DBUS_COMPILER_BYTE_ORDER throughout.
4715 (dbus_message_create_header): pad header to align the
4716 start of the body of the message to 8-byte boundary
4718 * dbus/dbus-marshal.h: make all the demarshalers take const
4719 DBusString arguments.
4721 * dbus/dbus-message.c (_dbus_message_loader_return_buffer):
4722 validate message args here, so we don't have to do slow validation
4723 later, and so we catch bad messages as they are incoming. Also add
4724 better checks on header_len and body_len. Also fill in
4727 * dbus/dbus-string.c (_dbus_string_validate_utf8): new (not
4728 implemented properly)
4729 (_dbus_string_validate_nul): new function to check all-nul
4731 * dbus/dbus-marshal.c (_dbus_marshal_get_field_end_pos): rename
4732 get_arg_end_pos and remove all validation
4733 (_dbus_marshal_validate_arg): actually do validation here.
4735 2003-01-29 Havoc Pennington <hp@pobox.com>
4737 * dbus/dbus-message.c (check_message_handling): fix assertion
4738 failure on set_client_serial
4740 2003-01-28 Havoc Pennington <hp@pobox.com>
4742 * dbus/dbus-server-debug.c: Add doc section comments
4744 * dbus/dbus-transport-debug.c: add doc section comments
4746 2003-01-28 Havoc Pennington <hp@redhat.com>
4748 * dbus/dbus-string.c (_dbus_string_base64_decode): append bytes in
4749 the reverse order from how I had it
4750 (_dbus_string_base64_encode): reverse encoding order. I was
4751 basically byteswapping everything during encoding.
4753 2003-01-28 Anders Carlsson <andersca@codefactory.se>
4755 * dbus/dbus-connection-internal.h:
4756 * dbus/dbus-connection.c: (_dbus_connection_add_timeout),
4757 (_dbus_connection_remove_timeout):
4758 Add functions for adding and removing timeouts.
4760 * dbus/dbus-message.c: (dbus_message_new_from_message):
4761 Add new function that takes a message and creates an exact
4762 copy of it, but with the refcount set to 1.
4763 (check_message_handling):
4766 * dbus/dbus-server-protected.h:
4767 * dbus/dbus-server.c: (_dbus_server_init_base),
4768 (_dbus_server_finalize_base), (_dbus_server_add_timeout),
4769 (dbus_server_set_timeout_functions):
4770 (_dbus_server_remove_timeout):
4771 New functions so that a server can add and remove timeouts.
4773 (dbus_server_listen):
4774 Add commented out call to dbus_server_debug_new.
4776 * dbus/dbus-timeout.c: (_dbus_timeout_new):
4777 Actually set the handler, doh.
4779 * dbus/dbus-transport.c: (_dbus_transport_open):
4780 Add commented out call to dbus_transport_debug_client_new.
4783 Add dbus-transport-debug.[ch] and dbus-server-debug.[ch]
4785 2003-01-28 Havoc Pennington <hp@pobox.com>
4787 * dbus/dbus-message.c (check_message_handling): function to check
4788 on the loaded message, iterates over it etc.
4790 2003-01-28 Havoc Pennington <hp@pobox.com>
4792 * test/Makefile.am (dist-hook): fix make distdir
4794 * dbus/Makefile.am (TESTS_ENVIRONMENT): fix make check
4796 2003-01-27 Havoc Pennington <hp@pobox.com>
4798 * dbus/dbus-mempool.c (time_for_size): replace printf with
4801 * dbus/dbus-message-builder.c (_dbus_message_data_load): allow
4802 empty lines; fix the SAVE_LENGTH stuff to be
4803 START_LENGTH/END_LENGTH so it actually works; couple other
4806 * test/Makefile.am (dist-hook): add dist-hook for .message files
4808 * dbus/dbus-string.c (DBUS_STRING_COPY_PREAMBLE): source of a copy
4809 can be constant or locked.
4810 (_dbus_string_free): allow freeing a const string as
4813 * dbus/dbus-sysdeps.c (_dbus_concat_dir_and_file): utility
4815 * dbus/dbus-test-main.c (main): take an argument which is the
4816 directory containing test data
4818 * dbus/dbus-message.c (_dbus_message_test): pass a test_data_dir
4819 argument to this and load all the messages in test/data/
4820 checking that they can be loaded or not loaded as appropriate.
4822 2003-01-27 Anders Carlsson <andersca@codefactory.se>
4824 * bus/dispatch.c: (bus_dispatch_message_handler):
4825 Dispatch messages sent to services.
4827 * bus/driver.c: (bus_driver_send_service_deleted),
4828 (bus_driver_send_service_created), (bus_driver_send_service_lost),
4829 (bus_driver_send_service_acquired):
4830 Add helper functions for sending service related messages.
4832 (bus_driver_send_welcome_message):
4833 Send HELLO_REPLY instead of WELCOME.
4835 (bus_driver_handle_list_services):
4836 Send LIST_SERVICES_REPLY instead of SERVICES.
4838 (bus_driver_handle_own_service),
4839 (bus_driver_handle_service_exists):
4840 New message handlers.
4842 (bus_driver_handle_message):
4843 Invoke new message handlers.
4845 (bus_driver_remove_connection):
4846 Don't remove any services here since that's done automatically
4847 by bus_service_remove_owner now.
4850 New function signatures.
4852 * bus/services.c: (bus_service_add_owner):
4853 Send ServiceAcquired message if we're the only primary owner.
4855 (bus_service_remove_owner):
4856 Send ServiceAcquired/ServiceLost messages.
4858 (bus_service_set_prohibit_replacement),
4859 (bus_service_get_prohibit_replacement):
4860 Functions for setting prohibit replacement.
4862 (bus_service_has_owner):
4863 New function that checks if a connection is in the owner queue of
4867 Add new function signatures.
4869 * dbus/dbus-list.c: (_dbus_list_test):
4870 Add tests for _dbus_list_remove_last and traversing the list backwards.
4873 Fix a typo in _dbus_list_get_prev_link, if we're at the first element we can't
4874 go any further, so return NULL then.
4876 * dbus/dbus-protocol.h:
4877 Add new messages, service flags and service replies.
4879 2003-01-26 Havoc Pennington <hp@pobox.com>
4881 * dbus/dbus-message-builder.c: implement, completely untested.
4883 * test/data/*: add data to be used in testing.
4884 ".message" files are our simple loadable text format.
4885 ".message-raw" will be binary dumps of messages.
4887 * dbus/dbus-string.c (_dbus_string_starts_with_c_str): new
4889 2003-01-26 Havoc Pennington <hp@pobox.com>
4891 * dbus/dbus-sysdeps.c (_dbus_file_get_contents): new function
4893 * dbus/dbus-errors.c (dbus_result_to_string): add
4896 * dbus/dbus-message-builder.c: new file, will contain code to load
4897 up messages from files. Not implemented yet.
4899 2003-01-26 Havoc Pennington <hp@pobox.com>
4901 * dbus/dbus-message.c (dbus_message_set_sender): support deleting
4902 the sender by setting to NULL
4904 2003-01-26 Havoc Pennington <hp@pobox.com>
4906 The unit tests pass, but otherwise untested. If it breaks, the
4907 tests should have been better. ;-)
4909 * bus/driver.c (bus_driver_handle_hello): return if we disconnect
4912 * dbus/dbus-message.c: redo everything so we maintain
4913 message->header as the only copy of the various fields.
4914 This avoids the possibility of out-of-memory in some cases,
4915 for example dbus_message_lock() can't run out of memory anymore,
4916 and avoids extra copying. Figured I may as well go ahead and do
4917 this since it was busted for dbus_message_lock to not return
4918 failure on OOM, and dbus_message_write_header was totally
4919 unchecked for OOM. Also fixed some random other bugs.
4921 * dbus/dbus-marshal.c (_dbus_marshal_get_field_end_pos): verify
4922 that strings are nul-terminated. Also, end_pos can be equal
4923 to string length just not greater than, I think.
4924 (_dbus_marshal_set_int32): new function
4925 (_dbus_marshal_set_uint32): new function
4926 (_dbus_marshal_set_string): new function
4928 * dbus/dbus-connection.c (_dbus_connection_new_for_transport): fix
4929 a warning, init timeout_list to NULL
4930 (dbus_connection_send_message): don't use uninitialized variable
4933 * dbus/dbus-string.c (_dbus_string_replace_len): new function
4935 2003-01-26 Anders Carlsson <andersca@codefactory.se>
4937 * bus/driver.c: (bus_driver_handle_hello),
4938 (bus_driver_send_welcome_message):
4941 2003-01-26 Anders Carlsson <andersca@codefactory.se>
4943 * dbus/dbus-auth.c: (process_auth), (_dbus_auth_unref):
4944 * dbus/dbus-connection.c: (_dbus_connection_new_for_transport),
4945 (dbus_connection_unref):
4946 * dbus/dbus-marshal.c: (_dbus_marshal_test):
4947 * dbus/dbus-message.c: (dbus_message_unref),
4950 (dbus_message_get_fields):
4951 Remove debugging printout.
4953 (_dbus_message_loader_return_buffer):
4954 Don't store the header string.
4956 (_dbus_message_test):
4959 2003-01-26 Richard Hult <rhult@codefactory.se>
4961 * glib/dbus-gmain.c (dbus_connection_dispatch): Traverse a copy of
4962 the file descriptor list, since it can change under us.
4964 2003-01-25 Anders Carlsson <andersca@codefactory.se>
4966 * glib/dbus-gmain.c: (dbus_connection_prepare),
4967 (dbus_connection_check), (dbus_connection_dispatch), (add_watch),
4968 (remove_watch), (dbus_connection_hookup_with_g_main):
4969 Rewrite the glib handling to use its own GSource instead of a
4970 GIOChannel so we can catch messages put in the queue while waiting
4973 2003-01-25 Anders Carlsson <andersca@codefactory.se>
4976 * bus/connection.c: (connection_disconnect_handler),
4977 (connection_watch_callback), (bus_connection_setup):
4978 * bus/dispatch.c: (send_one_message),
4979 (bus_dispatch_broadcast_message), (bus_dispatch_message_handler),
4980 (bus_dispatch_add_connection), (bus_dispatch_remove_connection):
4982 * bus/driver.c: (bus_driver_send_service_deleted),
4983 (bus_driver_send_service_created), (bus_driver_handle_hello),
4984 (bus_driver_send_welcome_message),
4985 (bus_driver_handle_list_services), (bus_driver_remove_connection),
4986 (bus_driver_handle_message):
4988 Refactor code, put the message dispatching in its own file. Use
4989 _DBUS_HANDLE_OOM. Also send ServiceDeleted messages when a client
4992 2003-01-25 Anders Carlsson <andersca@codefactory.se>
4994 * dbus/dbus-internals.h:
4995 Add _DBUS_HANDLE_OOM macro, it doesn't do anything currently.
4997 * dbus/dbus-message.c: (dbus_message_get_sender):
4998 * dbus/dbus-message.h:
4999 Implement dbus_message_get_sender.
5001 * dbus/dbus-protocol.h:
5002 Add message and service defines.
5004 2003-01-25 Anders Carlsson <andersca@codefactory.se>
5006 * dbus/dbus-connection.c: (dbus_connection_send_message):
5007 * dbus/dbus-message-internal.h:
5008 * dbus/dbus-message.c: (_dbus_message_get_client_serial),
5009 (dbus_message_write_header):
5010 Remove _dbus_messag_unlock and don't set the client serial on a
5011 message if one already exists.
5013 2003-01-24 Havoc Pennington <hp@pobox.com>
5015 * dbus/dbus-list.c (alloc_link): put a thread lock on the global
5018 * bus/driver.c (bus_driver_handle_list_services): fix a leak
5021 2003-01-25 Anders Carlsson <andersca@codefactory.se>
5023 * dbus/dbus-list.c: (alloc_link), (free_link):
5024 Use a memory pool for the links.
5026 2003-01-25 Anders Carlsson <andersca@codefactory.se>
5028 * bus/connection.c: (bus_connection_foreach):
5030 Add new bus_connection_foreach function.
5032 * bus/driver.c: (send_one_message), (bus_driver_broadcast_message):
5033 Add function that broadcasts a message to all clients.
5035 (bus_driver_send_service_created), (bus_driver_handle_hello),
5036 (bus_driver_send_welcome_message),
5037 (bus_driver_handle_list_services), (bus_driver_message_handler):
5038 Implement functions that take care of listing services, and notifying
5039 clients when new services are created.
5041 * bus/services.c: (bus_services_list):
5043 Add new function that returns an array of strings with the currently
5044 registered services.
5047 * glib/dbus-gmain.c:
5048 Update copyright year.
5050 2003-01-25 Anders Carlsson <andersca@codefactory.se>
5052 * dbus/dbus-connection.c: (dbus_connection_send_message):
5053 Unlock the message in case it was sent earlier.
5055 (dbus_connection_send_message_with_reply_and_block):
5056 Remove the reply message from the list.
5058 * dbus/dbus-marshal.c: (_dbus_demarshal_string_array):
5059 Set array_len and new_pos correctly.
5061 (_dbus_marshal_test):
5062 Remove debug output.
5064 * dbus/dbus-message-internal.h:
5065 * dbus/dbus-message.c: (_dbus_message_get_reply_serial):
5066 New function that returns the reply serial.
5068 (_dbus_message_unlock):
5069 New function that unlocks a message and resets its header.
5071 (dbus_message_append_string_array),
5072 (dbus_message_get_fields_valist),
5073 (dbus_message_iter_get_field_type),
5074 (dbus_message_iter_get_string_array),
5075 (dbus_message_get_fields),
5076 (dbus_message_append_fields_valist):
5077 Handle string arrays.
5079 (dbus_message_set_sender):
5080 Make this function public since the bus daemon needs it.
5082 (decode_header_data):
5083 Set the reply serial to -1 initially.
5085 * dbus/dbus-message.h:
5086 Add dbus_message_set_sender.
5088 2003-01-24 Havoc Pennington <hp@pobox.com>
5090 * doc/dbus-specification.sgml: add some stuff
5092 2003-01-22 Havoc Pennington <hp@pobox.com>
5094 * doc/dbus-specification.sgml: Start to document the protocol.
5096 2003-01-22 Havoc Pennington <hp@pobox.com>
5098 * dbus/dbus-connection.c
5099 (dbus_connection_send_message_with_reply_and_block): add some @todo
5101 * bus/driver.c (bus_driver_add_connection): add a FIXME about memleak
5103 2003-01-21 Havoc Pennington <hp@pobox.com>
5105 (patch untested because can't compile)
5107 * bus/driver.c (create_unique_client_name): make this function
5108 never recycle client names. Also, caller should initialize
5111 * dbus/dbus-sysdeps.c (_dbus_get_current_time): new function
5113 2003-01-21 Anders Carlsson <andersca@codefactory.se>
5115 * dbus/dbus-marshal.c: (_dbus_marshal_double),
5116 (_dbus_marshal_int32), (_dbus_marshal_uint32),
5117 (_dbus_marshal_int32_array), (_dbus_marshal_uint32_array),
5118 (_dbus_marshal_double_array), (_dbus_marshal_string_array),
5119 (_dbus_demarshal_int32_array), (_dbus_demarshal_uint32_array),
5120 (_dbus_demarshal_double_array), (_dbus_demarshal_string_array),
5121 (_dbus_marshal_get_field_end_pos), (_dbus_marshal_test):
5122 * dbus/dbus-marshal.h:
5123 * dbus/dbus-protocol.h:
5124 Add support for marshalling and demarshalling integer, double
5127 2003-01-21 Anders Carlsson <andersca@codefactory.se>
5132 * bus/connection.c: (connection_disconnect_handler):
5133 Remove the connection from the bus driver's list.
5135 (connection_watch_callback): Dispatch messages.
5137 (free_connection_data): Free connection name.
5139 (bus_connection_setup): Add connection to the bus driver's list.
5140 (bus_connection_remove_owned_service):
5141 (bus_connection_set_name), (bus_connection_get_name):
5142 Add functions for setting and getting the connection's name.
5145 Add function headers.
5147 * bus/driver.c: (create_unique_client_name),
5148 (bus_driver_handle_hello_message),
5149 (bus_driver_send_welcome_message), (bus_driver_message_handler),
5150 (bus_driver_add_connection), (bus_driver_remove_connection):
5153 * bus/services.c: (bus_service_free):
5155 New file that handles communication and registreation with the bus
5158 2003-01-21 Anders Carlsson <andersca@codefactory.se>
5160 * dbus/dbus-connection.c: (dbus_connection_send_message):
5161 Add a new client_serial parameter.
5163 (dbus_connection_send_message_with_reply):
5164 Remove a @todo since we've implemented the blocking function.
5166 (dbus_connection_send_message_with_reply_and_block):
5167 New function that sends a message and waits for a reply and
5168 then returns the reply.
5170 * dbus/dbus-connection.h:
5173 * dbus/dbus-errors.c: (dbus_result_to_string):
5174 * dbus/dbus-errors.h:
5175 Add new DBUS_RESULT.
5177 * dbus/dbus-message-internal.h:
5178 * dbus/dbus-message.c: (_dbus_message_get_reply_serial),
5179 (_dbus_message_set_sender), (dbus_message_write_header),
5180 (dbus_message_new_reply), (decode_header_data),
5181 (_dbus_message_loader_return_buffer), (_dbus_message_test):
5182 * dbus/dbus-message.h:
5183 Add new functions that set the reply serial and sender.
5184 Also marshal and demarshal them correctly and add test.
5186 * dbus/dbus-protocol.h:
5187 Add new DBUS_MESSAGE_TYPE_SENDER.
5190 * glib/dbus-gmain.c: (watch_callback), (free_callback_data),
5191 (add_watch), (remove_watch), (add_timeout), (remove_timeout),
5192 (dbus_connection_hookup_with_g_main):
5193 * glib/test-dbus-glib.c: (main):
5194 Rewrite to use GIOChannel and remove the GSource crack.
5196 * test/echo-client.c: (main):
5197 * test/watch.c: (check_messages):
5198 Update for changed APIs
5200 2003-01-19 Anders Carlsson <andersca@codefactory.se>
5202 * dbus/Makefile.am: Add dbus-timeout.[cħ]
5204 * dbus/dbus-connection.c: (_dbus_connection_new_for_transport):
5205 Create a DBusTimeoutList.
5206 (dbus_connection_set_timeout_functions): Add new function to
5207 set timeout callbacks
5209 * dbus/dbus-connection.h: Add public DBusTimeout API.
5211 * dbus/dbus-message.c: (dbus_message_get_service):
5212 * dbus/dbus-message.h: New function.
5214 * dbus/dbus-server.c: Fix small doc typo.
5216 * dbus/dbus-timeout.[ch]: New files for mainloop timeouts.
5218 2003-01-19 Anders Carlsson <andersca@codefactory.se>
5220 * dbus/dbus-string.c (_dbus_string_move_len): Don't delete all
5221 of the string, just as long as specified.
5223 2003-01-19 Havoc Pennington <hp@pobox.com>
5225 * dbus/dbus-connection.c (dbus_connection_get_is_authenticated):
5228 * dbus/dbus-server.c (dbus_server_set_max_connections)
5229 (dbus_server_get_max_connections, dbus_server_get_n_connections):
5230 keep track of current number of connections, and add API for
5231 setting a max (but haven't implemented enforcing the max yet)
5233 2003-01-18 Havoc Pennington <hp@pobox.com>
5235 * dbus/dbus-transport-unix.c (unix_do_iteration): only do the
5236 reading/writing if read_watch != NULL or write_watch != NULL.
5238 * dbus/dbus-message.c (_dbus_message_loader_return_buffer): fix
5239 the message loader code to actually load message->header and
5240 message->body into the newly-created message.
5242 * dbus/dbus-transport-unix.c (check_write_watch): fix a mem leak
5245 * dbus/dbus-connection.c (dbus_connection_set_max_message_size)
5246 (dbus_connection_get_max_message_size)
5247 (dbus_connection_set_max_live_messages_size)
5248 (dbus_connection_get_max_live_messages_size): implement some
5249 resource limitation functions
5251 * dbus/dbus-resources.c: new file implementing some of the
5252 resource limits stuff
5254 * dbus/dbus-message.c (dbus_message_iter_get_byte_array): add
5255 missing docs, add @todo to handle OOM etc.
5257 * dbus/dbus-marshal.c (_dbus_demarshal_byte_array): add missing
5260 2003-01-18 Havoc Pennington <hp@pobox.com>
5262 * dbus/dbus-connection.c (dbus_connection_unref): disconnect the
5263 connection if it hasn't been already.
5265 * dbus/dbus-connection.h: kill off the idea of an ErrorFunction,
5266 replace with DisconnectFunction.
5268 2003-01-18 Havoc Pennington <hp@pobox.com>
5270 Building --disable-verbose-mode --disable-asserts --disable-tests
5271 cuts the library from 112K to 45K or so
5273 * configure.in: check for varargs macro support,
5274 add --enable-verbose-mode, --enable-asserts.
5276 * dbus/dbus-internals.h (_dbus_assert): support
5278 (_dbus_verbose): support DBUS_ENABLE_VERBOSE_MODE
5280 2003-01-18 Havoc Pennington <hp@pobox.com>
5282 * dbus/dbus-test.c: include config.h so that tests actually run
5284 * dbus/dbus-string.c: add assertions that stuff is 8-byte aligned,
5285 so the failure mode when that assumption fails will be plenty
5288 2003-01-18 Havoc Pennington <hp@pobox.com>
5290 * configure.in: default --enable-tests to $USE_MAINTAINER_MODE
5292 * dbus/Makefile.am: fix it up so dubs-test-main.c is included in
5295 * test/Makefile.am: don't use special variable "TESTS" for echo-*
5296 since we don't want to use those in make check
5298 2003-01-15 Havoc Pennington <hp@redhat.com>
5304 2003-01-15 Havoc Pennington <hp@redhat.com>
5306 * test/Makefile.am: fix so that test source code ends up in the
5307 distribution on make distcheck
5309 2003-01-15 Havoc Pennington <hp@redhat.com>
5315 2003-01-15 Havoc Pennington <hp@redhat.com>
5317 * dbus/dbus-test.c (dbus_internal_symbol_do_not_use_run_tests):
5318 fix build when --disable-tests
5320 * Makefile.am (EXTRA_DIST): put HACKING in here
5322 * HACKING: document procedure for making a tarball release.
5324 2003-01-14 Anders Carlsson <andersca@codefactory.se>
5326 * bus/connection.c: (connection_error_handler),
5327 (bus_connection_setup):
5328 * bus/main.c: (main):
5329 Make sure that the DBusConnectionData struct is NULLed
5330 out to prevent a segfault.
5332 * dbus/dbus-errors.c: (dbus_result_to_string):
5333 * dbus/dbus-errors.h:
5334 * dbus/dbus-message.c: (dbus_message_get_fields),
5335 (dbus_message_get_fields_valist), (_dbus_message_test):
5336 * dbus/dbus-message.h:
5337 Make dbus_message_get_fields return a result code so we can
5338 track invalid fields as well as oom.
5340 2003-01-11 Havoc Pennington <hp@pobox.com>
5342 * configure.in: change --enable-test/--enable-ansi action-if-given
5343 to enable_foo=$enableval instead of enable_foo=yes
5345 2003-01-08 Havoc Pennington <hp@pobox.com>
5347 * dbus/dbus-string.c (_dbus_string_align_length): new function
5349 * dbus/dbus-test-main.c: move main() for test app here
5351 (dbus_internal_symbol_do_not_use_run_tests): we have to export a
5352 symbol to run tests, because dbus-test isn't in the main
5355 Code review nitpicks.
5357 * dbus/dbus-message.c (dbus_message_write_header): add newlines
5358 for people with narrow emacs ;-). Assert client_serial was filled
5359 in. Assert message->name != NULL.
5360 (dbus_message_append_fields): have "first_field_type" arg separate
5361 from va list, needed for C++ binding that also uses varargs IIRC
5362 and helps with type safety
5363 (dbus_message_new): add @todo about using DBusString to store
5364 service/name internally
5365 (dbus_message_new): don't leak ->service and ->name on OOM later
5367 (dbus_message_unref): free the service name
5368 (dbus_message_get_fields): same change to varargs
5369 i.e. first_field_type
5370 (_dbus_message_loader_return_buffer): assert that the message data
5371 is aligned (if not it's a bug in our code). Put in verbose griping
5372 about why we set corrupted = TRUE.
5373 (decode_header_data): add FIXME that char* is evil. Was going to
5374 add FIXME about evil locale-specific string.h strncmp, but just
5375 switched to wacky string-as-uint32 optimization. Move check for
5376 "no room for field name" above get_const_data_len() to avoid
5377 assertion failure in get_const_data_len if we have trailing 2
5378 bytes or the like. Check for service and name fields being
5379 provided twice. Don't leak service/name on error. Require field
5380 names to be aligned to 4 bytes.
5382 * dbus/dbus-marshal.c: move byte swap stuff to header
5383 (_dbus_pack_int32): uscore-prefix
5384 (_dbus_unpack_int32): uscore-prefix
5385 (_dbus_unpack_uint32): export
5386 (_dbus_demarshal_string): add @todo complaining about use of
5388 (_dbus_marshal_get_field_end_pos): add @todo about bad error
5389 handling allowing corrupt data to go unchecked
5391 2003-01-08 Havoc Pennington <hp@redhat.com>
5393 * dbus/dbus-transport-unix.c (unix_do_iteration): add read/write
5394 to the select() as needed for authentication. (should be using
5395 _dbus_poll() not select, but for another day)
5397 * dbus/dbus.h: include dbus/dbus-protocol.h
5399 2003-01-08 Anders Carlsson <andersca@codefactory.se>
5401 * dbus/Makefile.am (dbusinclude_HEADERS): Install
5404 2003-01-08 Anders Carlsson <andersca@codefactory.se>
5406 * dbus/dbus-internals.c: (_dbus_type_to_string):
5407 New function that returns a string describing a type.
5409 * dbus/dbus-marshal.c: (_dbus_demarshal_byte_array):
5410 * dbus/dbus-marshal.h:
5411 * dbus/dbus-message.c: (dbus_message_get_fields_valist),
5412 (dbus_message_iter_get_field_type), (dbus_message_iter_get_double),
5413 (dbus_message_iter_get_byte_array):
5414 * dbus/dbus-message.h:
5415 Add new convenience functions for appending and getting message fields.
5416 Also add demarshalling routines for byte arrays.
5418 2003-01-07 Anders Carlsson <andersca@codefactory.se>
5420 * dbus/dbus-connection-internal.h:
5421 * dbus/dbus-connection.c: (_dbus_connection_new_for_transport),
5422 (_dbus_connection_get_next_client_serial),
5423 (dbus_connection_send_message):
5424 * dbus/dbus-internals.h:
5425 * dbus/dbus-marshal.c: (unpack_uint32), (dbus_unpack_int32),
5426 (dbus_pack_int32), (_dbus_marshal_double), (_dbus_marshal_int32),
5427 (_dbus_marshal_uint32), (_dbus_demarshal_double),
5428 (_dbus_demarshal_int32), (_dbus_demarshal_uint32),
5429 (_dbus_demarshal_string), (_dbus_marshal_get_field_end_pos),
5430 (_dbus_verbose_bytes), (_dbus_marshal_test):
5431 * dbus/dbus-marshal.h:
5432 * dbus/dbus-message-internal.h:
5433 * dbus/dbus-message.c: (_dbus_message_set_client_serial),
5434 (dbus_message_write_header), (_dbus_message_lock),
5435 (dbus_message_new), (dbus_message_ref), (dbus_message_unref),
5436 (dbus_message_get_name), (dbus_message_append_int32),
5437 (dbus_message_append_uint32), (dbus_message_append_double),
5438 (dbus_message_append_string), (dbus_message_append_byte_array),
5439 (dbus_message_get_fields_iter), (dbus_message_iter_ref),
5440 (dbus_message_iter_unref), (dbus_message_iter_has_next),
5441 (dbus_message_iter_next), (dbus_message_iter_get_field_type),
5442 (dbus_message_iter_get_string), (dbus_message_iter_get_int32),
5443 (dbus_message_iter_get_uint32), (dbus_message_iter_get_double),
5444 (decode_header_data), (_dbus_message_loader_return_buffer),
5445 (message_iter_test), (_dbus_message_test):
5446 * dbus/dbus-message.h:
5447 * dbus/dbus-protocol.h:
5448 * dbus/dbus-test.c: (main):
5450 * glib/test-dbus-glib.c: (message_handler), (main):
5451 * test/echo-client.c: (main):
5452 * test/watch.c: (check_messages):
5453 Make messages sendable and receivable for real.
5455 2003-01-07 Anders Carlsson <andersca@codefactory.se>
5457 * dbus/dbus-marshal.c: (_dbus_marshal_double),
5458 (_dbus_marshal_string), (_dbus_marshal_byte_array):
5459 * dbus/dbus-message.c: (dbus_message_append_int32),
5460 (dbus_message_append_uint32), (dbus_message_append_double),
5461 (dbus_message_append_string), (dbus_message_append_byte_array):
5462 Handle OOM restoration.
5464 2003-01-07 Anders Carlsson <andersca@codefactory.se>
5466 * dbus/dbus-marshal.c: (_dbus_marshal_string),
5467 (_dbus_demarshal_string), (_dbus_marshal_test):
5468 * dbus/dbus-marshal.h:
5469 * dbus/dbus-message.c: (dbus_message_get_name),
5470 Document these functions.
5472 (dbus_message_append_int32), (dbus_message_append_uint32),
5473 (dbus_message_append_double), (dbus_message_append_string),
5474 (dbus_message_append_byte_array):
5475 * dbus/dbus-message.h:
5476 Add functions for adding message fields of different types.
5478 * dbus/dbus-protocol.h:
5479 Add the different types.
5481 2003-01-05 Havoc Pennington <hp@pobox.com>
5483 * bus/connection.c: implement routines for handling connections,
5484 first thing is keeping a list of owned services on each connection
5485 and setting up watches etc.
5487 * bus/services.c: implement a mapping from service names to lists
5490 * dbus/dbus-hash.c: add DBUS_HASH_POINTER
5492 * dbus/dbus-threads.c (dbus_static_mutex_lock): add functions
5493 to use static mutexes for global data
5495 * dbus/dbus-connection.c (dbus_connection_set_data): add new
5496 collection of functions to set/get application-specific data
5497 on the DBusConnection.
5499 2003-01-04 Havoc Pennington <hp@pobox.com>
5501 * dbus/dbus-sysdeps.c (_dbus_sleep_milliseconds): new function
5502 (_dbus_poll): new function
5504 * dbus/dbus-internals.h (_DBUS_STRUCT_OFFSET): new macro
5507 * bus/loop.c: initial code for the daemon main loop
5509 2003-01-04 Havoc Pennington <hp@pobox.com>
5511 * test/watch.c (error_handler): make it safe if the error handler
5512 is called multiple times (if we s/error handler/disconnect
5513 handler/ we should just guarantee it's called only once)
5515 * dbus/dbus-transport.c (_dbus_transport_disconnect): call the
5516 error handler on disconnect (it's quite possible we should
5517 just change the error handler to a "disconnect handler," I'm
5518 not sure we have any other meaningful errors)
5520 * configure.in: check for getpwnam_r
5522 * dbus/dbus-transport.c, dbus/dbus-transport-unix.c,
5523 dbus/dbus-auth.c: add credentials support, add EXTERNAL auth
5524 mechanism as in SASL spec, using socket credentials
5526 * dbus/dbus-sysdeps.c (_dbus_read_credentials_unix_socket): new function
5527 (_dbus_send_credentials_unix_socket): new function
5529 * dbus/dbus-sysdeps.c (_dbus_accept_unix_socket): rename just
5531 (_dbus_write): only check errno if <0 returned
5532 (_dbus_write_two): ditto
5534 2003-01-02 Anders Carlsson <andersca@codefactory.se>
5536 * dbus/dbus-marshal.c: (_dbus_marshal_utf8_string),
5537 (_dbus_marshal_byte_array), (_dbus_demarshal_utf8_string),
5538 (_dbus_marshal_test):
5539 * dbus/dbus-marshal.h:
5540 Add _dbus_marshal_byte_array and rename _dbus_marshal_string
5541 to _dbus_marshal_utf8_string. Also fix some tests.
5543 2002-12-28 Harri Porten <porten@kde.org>
5545 * configure.in: added check for C++ compiler and a very cheesy
5546 check for the Qt integration
5548 * Makefile.am (SUBDIRS): compile qt subdir if support is enabled
5550 * qt/Makefile.am: added
5552 * qt/.cvsignore: added
5554 * qt/dbus-qthread.cc, qt/dbus-qthread.cpp: renamed former to
5555 latter, added #ifdef QT_THREAD_SUPPORT guard.
5557 * dbus/Makefile.am: added missing headers for make dist
5559 2002-12-28 Kristian Rietveld <kris@gtk.org>
5561 * dbus/Makefile.am: fixup export-symbols-regex.
5563 2002-12-27 Anders Carlsson <andersca@codefactory.se>
5565 * acinclude.m4: Add this file and put the
5566 PKG_CHECK_MODULE macro in it.
5568 2002-12-27 Anders Carlsson <andersca@codefactory.se>
5570 * dbus/dbus-marshal.c: (_dbus_marshal_string),
5571 (_dbus_demarshal_double), (_dbus_demarshal_int32),
5572 (_dbus_demarshal_uint32), (_dbus_demarshal_string),
5573 (_dbus_marshal_test):
5574 Make the demarshalling routines align the pos argument.
5575 Add string marshalling tests and fix the obvious bugs
5578 2002-12-26 Havoc Pennington <hp@pobox.com>
5580 * dbus/dbus-auth.c: fixes fixes fixes
5582 * dbus/dbus-transport-unix.c: wire up support for
5583 encoding/decoding data on the wire
5585 * dbus/dbus-auth.c (_dbus_auth_encode_data)
5586 (_dbus_auth_decode_data): append to target string
5587 instead of nuking it.
5589 2002-12-26 Havoc Pennington <hp@pobox.com>
5591 * dbus/dbus-marshal.h (DBUS_COMPILER_BYTE_ORDER): #ifdef
5592 WORDS_BIGENDIAN then compiler byte order is DBUS_BIG_ENDIAN,
5595 * dbus/dbus-marshal.c: Add macros to do int swapping in-place and
5596 avoid swap_bytes() overhead (ignoring possible assembly stuff for
5597 now). Main point is because I wanted unpack_uint32 to implement
5599 (_dbus_verbose_bytes): new function
5601 * dbus/dbus-string.c (_dbus_string_validate_ascii): new function
5603 * dbus/dbus-message.c (_dbus_message_loader_get_is_corrupted): add
5604 mechanism to handle a corrupt message stream
5605 (_dbus_message_loader_new): fix preallocation to only prealloc,
5608 * dbus/dbus-string.c (_dbus_string_skip_blank): fix this function
5609 (_dbus_string_test): enhance tests for copy/move and fix the
5612 * dbus/dbus-transport-unix.c: Hold references in more places to
5613 avoid reentrancy problems
5615 * dbus/dbus-transport.c: ditto
5617 * dbus/dbus-connection.c (dbus_connection_dispatch_message): don't
5618 leak reference count in no-message case
5620 * test/watch.c (do_mainloop): handle adding/removing watches
5621 during iteration over the watches. Also, ref the connection/server
5622 stored on a watch, so we don't try to mangle a destroyed one.
5624 * dbus/dbus-transport-unix.c (do_authentication): perform
5627 * dbus/dbus-auth.c (get_state): add a state
5628 AUTHENTICATED_WITH_UNUSED_BYTES and return it if required
5629 (_dbus_auth_get_unused_bytes): append the unused bytes
5630 to the passed in string, rather than prepend
5632 * dbus/dbus-transport.c (_dbus_transport_init_base): create
5633 the auth conversation DBusAuth
5635 * dbus/dbus-transport-unix.c (_dbus_transport_new_for_fd)
5636 (_dbus_transport_new_for_domain_socket): when creating a
5637 transport, pass in whether it's a client-side or server-side
5638 transport so we know which DBusAuth to create
5640 2002-12-03 Havoc Pennington <hp@pobox.com>
5642 * dbus/dbus-transport-unix.c (unix_finalize): finalize base
5643 _after_ finalizing the derived members
5644 (unix_connection_set): unref watch if we fail to add it
5646 * dbus/dbus-connection.c (dbus_connection_unref): delete the
5647 transport first, so that the connection owned by the
5648 transport will be valid as the transport finalizes.
5650 * dbus/dbus-transport-unix.c (unix_finalize): free the write_watch
5651 if necessary, and remove watches from the connection.
5653 * dbus/dbus-watch.c (_dbus_watch_list_free): improve a comment
5655 2002-12-26 Anders Carlsson <andersca@codefactory.se>
5657 * dbus/dbus-marshal.c: (_dbus_marshal_string),
5658 (_dbus_demarshal_double), (_dbus_demarshal_int32),
5659 (_dbus_demarshal_uint32), (_dbus_demarshal_string),
5660 (_dbus_marshal_test):
5661 * dbus/dbus-marshal.h:
5662 Add string marshal functions and have the demarshal functions
5663 return the new position.
5665 2002-12-25 Havoc Pennington <hp@pobox.com>
5667 * doc/dbus-sasl-profile.txt: docs on the authentication protocol,
5668 it is a simple protocol that just maps directly to SASL.
5670 * dbus/dbus-auth.h, dbus/dbus-auth.c: authentication protocol
5671 initial implementation, not actually used yet.
5673 * dbus/dbus-string.c (_dbus_string_find): new function
5674 (_dbus_string_equal): new function
5675 (_dbus_string_base64_encode): new function
5676 (_dbus_string_base64_decode): new function
5678 2002-12-25 Anders Carlsson <andersca@codefactory.se>
5681 * dbus/dbus-marshal.c: (swap_bytes), (_dbus_marshal_double),
5682 (_dbus_marshal_int32), (_dbus_marshal_uint32),
5683 (_dbus_demarshal_double), (_dbus_demarshal_int32),
5684 (_dbus_demarshal_uint32), (_dbus_marshal_test):
5685 * dbus/dbus-marshal.h:
5686 * dbus/dbus-protocol.h:
5687 * dbus/dbus-test.c: (main):
5689 Add un-optimized marshalling/demarshalling routines.
5691 2002-12-25 Harri Porten <porten@kde.org>
5693 * qt/dbus-qt.h: adjusted ctor and getter to KDE/Qt conventions
5695 2002-12-24 Zack Rusin <zack@kde.org>
5697 * qt/dbus-qthread.cc: adding - integrates QMutex into Dbus
5698 * qt/dbus-qt.h: skeleton with two sample implemenatation of the
5701 2002-12-24 Havoc Pennington <hp@pobox.com>
5703 * glib/dbus-gthread.c: fix include
5705 * glib/dbus-glib.h: rename DBusMessageHandler for now.
5706 I think glib API needs to change, though, as you don't
5707 want to use DBusMessageFunction, you want to use the
5708 DBusMessageHandler object. Probably
5709 dbus_connection_open_with_g_main_loop()
5710 and dbus_connection_setup_g_main_loop() or something like that
5711 (but think of better names...) that just create a connection
5712 that has watch/timeout functions etc. already set up.
5714 * dbus/dbus-connection.c
5715 (dbus_connection_send_message_with_reply): new function just to
5716 show how the message handler helps us deal with replies.
5718 * dbus/dbus-list.c (_dbus_list_remove_last): new function
5720 * dbus/dbus-string.c (_dbus_string_test): free a string that
5723 * dbus/dbus-hash.c: use memory pools for the hash entries
5724 (rebuild_table): be more paranoid about overflow, and
5725 shrink table when we can
5726 (_dbus_hash_test): reduce number of sprintfs and write
5727 valid C89. Add tests for case where we grow and then
5728 shrink the hash table.
5730 * dbus/dbus-mempool.h, dbus/dbus-mempool.c: memory pools
5732 * dbus/dbus-connection.c (dbus_connection_register_handler)
5733 (dbus_connection_unregister_handler): new functions
5735 * dbus/dbus-message.c (dbus_message_get_name): new
5737 * dbus/dbus-list.c: fix docs typo
5739 * dbus/dbus-message-handler.h, dbus/dbus-message-handler.c:
5740 an object representing a handler for messages.
5742 2002-12-16 Anders Carlsson <andersca@codefactory.se>
5745 * glib/dbus-gthread.c: (dbus_gthread_init):
5746 Don't use the gdbus prefix for public functions.
5748 2002-12-16 Anders Carlsson <andersca@codefactory.se>
5752 Add GLib checks and fixup .pc files
5756 * glib/dbus-gmain.c: (gdbus_connection_prepare),
5757 (gdbus_connection_check), (gdbus_connection_dispatch),
5758 (gdbus_add_connection_watch), (gdbus_remove_connection_watch),
5759 (dbus_connection_gsource_new):
5760 * glib/dbus-gthread.c: (dbus_gmutex_new), (dbus_gmutex_free),
5761 (dbus_gmutex_lock), (dbus_gmutex_unlock), (dbus_gthread_init):
5762 * glib/test-dbus-glib.c: (message_handler), (main):
5765 2002-12-15 Harri Porten <porten@kde.org>
5767 * autogen.sh: check for libtoolize before attempting to use it
5769 * dbus/dbus-transport-unix.c: include <sys/time.h> for timeval
5772 * .cvsignore: ignore more stamp files
5774 * dbus/dbus-watch.c (_dbus_watch_list_new): fixed doc error
5776 * test/Makefile.am: added -I$(top_srcdir) to be able to compile
5777 without make install.
5779 2002-12-15 Havoc Pennington <hp@pobox.com>
5781 * dbus/dbus-threads.c: add thread stubs that a higher library
5782 layer can fill in. e.g. the GLib wrapper might fill them in with
5783 GThread stuff. We still need to use this thread API to
5784 thread-safe-ize the library.
5786 2002-12-12 Havoc Pennington <hp@pobox.com>
5788 * dbus/dbus-transport-unix.c, dbus/dbus-server-unix.c: use the
5789 below new interfaces and include fewer system headers.
5791 * dbus/dbus-sysdeps.c (_dbus_read): new function
5792 (_dbus_write): new function
5793 (_dbus_write_two): new function
5794 (_dbus_connect_unix_socket): new function
5795 (_dbus_listen_unix_socket): new function
5797 * dbus/dbus-message-internal.h: change interfaces to use
5800 2002-12-11 Havoc Pennington <hp@pobox.com>
5802 * dbus/dbus-types.h: add dbus_unichar
5804 * dbus/dbus-internals.c (_dbus_verbose): use _dbus_getenv
5806 * dbus/dbus-connection.c (dbus_connection_send_message): return
5809 * dbus/dbus-transport.c: include dbus-watch.h
5811 * dbus/dbus-connection.c: include dbus-message-internal.h
5813 * HACKING: add file with coding guidelines stuff.
5815 * dbus/dbus-string.h, dbus/dbus-string.c: Encapsulate all string
5816 handling here, for security purposes (as in vsftpd). Not actually
5817 using this class yet.
5819 * dbus/dbus-sysdeps.h, dbus/dbus-sysdeps.c: Encapsulate all
5820 system/libc usage here, as in vsftpd, for ease of auditing (and
5821 should also simplify portability). Haven't actually moved all the
5822 system/libc usage into here yet.
5824 2002-11-25 Havoc Pennington <hp@pobox.com>
5826 * dbus/dbus-internals.c (_dbus_verbose): fix to not
5827 always print the first verbose message.
5829 2002-11-24 Havoc Pennington <hp@pobox.com>
5831 * test/echo-client.c, test/echo-server.c: cheesy test
5834 * configure.in (AC_CHECK_FUNCS): check for writev
5836 * dbus/dbus-message.c (_dbus_message_get_network_data): new
5839 * dbus/dbus-list.c (_dbus_list_foreach): new function
5841 * dbus/dbus-internals.c (_dbus_verbose): new function
5843 * dbus/dbus-server.c, dbus/dbus-server.h: public object
5844 representing a server that listens for connections.
5846 * dbus/.cvsignore: create
5848 * dbus/dbus-errors.h, dbus/dbus-errors.c:
5849 public API for reporting errors
5851 * dbus/dbus-connection.h, dbus/dbus-connection.c:
5852 public object representing a connection that
5853 sends/receives messages. (Same object used for
5854 both client and server.)
5856 * dbus/dbus-transport.h, dbus/dbus-transport.c:
5857 Basic abstraction for different kinds of stream
5858 that we might read/write messages from.
5860 2002-11-23 Havoc Pennington <hp@pobox.com>
5862 * dbus/dbus-internals.h (_DBUS_INT_MAX): add _DBUS_INT_MIN
5865 * dbus/dbus-test.c (main): add list test, and include
5866 dbus-test.h as intended
5868 * dbus/dbus-hash.c (_dbus_hash_table_remove_string)
5869 (_dbus_hash_table_remove_int): return value indicates
5870 whether the entry existed to remove
5872 * dbus/dbus-list.c: add linked list utility class,
5875 * dbus/dbus-hash.c: add TODO item about shrinking the hash bucket
5878 2002-11-23 Havoc Pennington <hp@pobox.com>
5880 * Doxyfile.in (INCLUDE_FILE_PATTERNS): expand DBUS_BEGIN_DECLS/
5881 DBUS_END_DECLS to nothing, that should fix this once and for all
5883 * Doxyfile.in (JAVADOC_AUTOBRIEF): set to YES
5885 * dbus/dbus-message.c, dbus/dbus-hash.c:
5886 add some missing @brief
5888 2002-11-23 Havoc Pennington <hp@pobox.com>
5890 * dbus/dbus-message.h: put semicolons after DEBUG_BEGIN_DECLS
5891 to avoid confusing Doxygen
5893 * dbus/dbus-hash.c: @} not }@
5895 * dbus/dbus-message.c (struct DBusMessage): split out
5898 2002-11-23 Havoc Pennington <hp@pobox.com>
5900 * configure.in: pile on more warning flags if using gcc
5902 * Doxyfile.in (EXTRACT_STATIC): set to NO, so we don't have
5903 to document static functions
5905 * configure.in: add summary to end of configure so it
5906 looks nice and attractive
5908 * dbus/dbus-hash.c: finish implementation and write unit
5911 * configure.in: add --enable-tests to enable unit tests
5913 * dbus/dbus-test.c: test program to run unit tests
5914 for all files in dbus/*, initially runs a test for
5917 * dbus/dbus-internals.h: file to hold some internal utility stuff
5919 2002-11-22 Havoc Pennington <hp@redhat.com>
5921 * dbus/dbus-hash.c: copy in Tcl hash table, not yet
5922 "ported" away from Tcl
5924 * dbus/dbus-types.h: header for types such as dbus_bool_t
5926 2002-11-22 Havoc Pennington <hp@redhat.com>
5928 * dbus/dbus.h: fixups for doc warnings
5930 * Doxyfile.in (FILE_PATTERNS): we need to scan .h to pick up
5932 (QUIET): make it quiet so we can see warnings
5934 * dbus/dbus-memory.c: teach D-BUS to allocate and free memory
5936 2002-11-22 Havoc Pennington <hp@redhat.com>
5938 * Makefile.am: include "Doxyfile" target in all-local
5940 * configure.in: generate the Doxyfile
5942 * Doxyfile.in: move Doxyfile here, so we can use
5943 configure to generate a Doxyfile with the right
5946 2002-11-22 Havoc Pennington <hp@redhat.com>
5948 * dbus/dbus-message.c: move inline docs into .c file
5950 * Doxyfile (OUTPUT_DIRECTORY): move output to doc/api
5951 so all docs are under doc/
5952 (MAN_EXTENSION): generate man pages. Use extension
5953 ".3dbus" which matches ".3qt" on my system,
5954 I guess this is OK, I don't know really.
5955 (FILE_PATTERNS): look for .c files not .h, makes sense
5958 2002-11-22 Havoc Pennington <hp@pobox.com>
5960 * Makefile.am (SUBDIRS): rename subdir "server" to "bus"
5961 because any app can be a server, and any app can be a client,
5962 the bus is a special kind of server.
5964 Thu Nov 21 23:35:31 2002 Zack Rusin <zack@kde.org>
5966 * Doxyfile : adding. Still needs Makefile rules to be generated
5967 automatically (just run "doxygen" in the toplevel dir for now to
5970 * dbus/dbus-message.h : Adding sample docs (javadoc since
5971 resembles gtk-doc a little more)
5973 * dbus/dbus.h : Adding sample docs
5975 2002-11-21 Havoc Pennington <hp@redhat.com>
5977 * dbus/Makefile.am (INCLUDES): define DBUS_COMPILATION
5978 so we can allow ourselves to include files directly,
5979 instead of having to use dbus.h
5981 * dbus/dbus.h: fill in
5983 * dbus/dbus-message.h: sketch out a sample header file.
5984 Include griping if you include it directly instead of
5987 * dbus/dbus-macros.h: new file with macros for extern "C",
5988 TRUE/FALSE, NULL, etc.
5990 * doc/file-boilerplate.c: put include guards in here
5992 2002-11-21 Havoc Pennington <hp@redhat.com>
5994 * doc/file-boilerplate.c: include both AFL and GPL boilerplate.
5996 * COPYING: include the GPL as well, and license code
5997 under both AFL and GPL.
5999 2002-11-21 Havoc Pennington <hp@redhat.com>
6001 * acconfig.h: get rid of this
6003 * autogen.sh (run_configure): add --no-configure option
6005 * configure.in: remove AC_ARG_PROGRAM to make
6006 autoconf complain less. add AC_PREREQ.
6007 add AC_DEFINE third arg.
6009 2002-11-21 Anders Carlsson <andersca@codefactory.se>
6012 Fix references so we can distcheck.
6014 2002-11-21 Havoc Pennington <hp@redhat.com>
6016 * Initial module creation