summaryrefslogtreecommitdiff
path: root/legacy (follow)
AgeCommit message (Collapse)Author
2012-12-20edbus: avoid cyclic unrefLucas De Marchi
If we are freeing a EDBUS_Connection_Name its name_owner_changed signal handler may hold a pointer and try to unref it when deleting the signal handler. We can't simply make the signal handler hold a reference to the connection name, otherwise edbus_connection_name_gc will never be triggered because of cyclic references. Thus, just set the cn->name_owner_changed->bus to NULL before trying to delete the signal handler. Related log found by Lucas Jóia: ==20607== Invalid read of size 4 ==20607== at 0x6FE29EE: edbus_connection_name_gc.isra.3 (edbus_core.c:375) ==20607== by 0x6FE4287: edbus_connection_unref (edbus_core.c:1028) ==20607== by 0x4C8D94: e_msgbus_shutdown (e_msgbus.c:167) ==20607== by 0x436194: _e_main_shutdown (e_main.c:1136) ==20607== by 0x434F25: main (e_main.c:1074) ==20607== Address 0x1461ba68 is 24 bytes inside a block of size 64 free'd ==20607== at 0x4C2A739: free (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) ==20607== by 0x6FF0E78: edbus_signal_handler_unref (edbus_signal_handler.c:269) ==20607== by 0x6FE2A48: edbus_connection_name_gc.isra.3 (edbus_core.c:384) ==20607== by 0x6FE4287: edbus_connection_unref (edbus_core.c:1028) ==20607== by 0x4C8D94: e_msgbus_shutdown (e_msgbus.c:167) ==20607== by 0x436194: _e_main_shutdown (e_main.c:1136) ==20607== by 0x434F25: main (e_main.c:1074) SVN revision: 81463
2012-12-20edbus: Do not modify cn->names hash while walking itLucas De Marchi
Bug triggered by Lucas Jóia: ==10042== Invalid read of size 8 ==10042==    at 0x6B86626: _eina_rbtree_iterator_next (eina_rbtree.c:165) ==10042==    by 0x6B7228D: _eina_hash_iterator_next (eina_hash.c:622) ==10042==    by 0x6FE41DC: edbus_connection_unref (edbus_core.c:1015) ==10042==    by 0x4C8D94: e_msgbus_shutdown (e_msgbus.c:167) ==10042==    by 0x436194: _e_main_shutdown (e_main.c:1136) ==10042==    by 0x434F25: main (e_main.c:1074) ==10042==  Address 0x15c1b958 is 40 bytes inside a block of size 96 free'd ==10042==    at 0x4C2A739: free (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) ==10042==    by 0x6B71CB7: _eina_hash_del_by_hash_el (eina_hash.c:441) ==10042==    by 0x6FE2A1E: edbus_connection_name_gc.isra.2 (edbus_core.c:385) ==10042==    by 0x6FE4217: edbus_connection_unref (edbus_core.c:1026) ==10042==    by 0x4C8D94: e_msgbus_shutdown (e_msgbus.c:167) ==10042==    by 0x436194: _e_main_shutdown (e_main.c:1136) ==10042==    by 0x434F25: main (e_main.c:1074) SVN revision: 81462
2012-12-20edbus: no need for this auxiliary variables, and check refcount firstLucas De Marchi
SVN revision: 81461
2012-12-20edje_entry: check NULL parameterJihoon Kim
SVN revision: 81418
2012-12-20edje_entry: NULL checking should be processed before accessingJihoon Kim
SVN revision: 81417
2012-12-20edbus: Remove object path of service if ObjectManager is the last optional ifaceJosé Roberto de Souza
Patch by: José Roberto de Souza <zehortigoza@profusion.mobi> SVN revision: 81416
2012-12-20edbus: Fix undefined ref. to 'EINA_LOG_DOMAIN_GLOBAL' in examplesJosé Roberto de Souza
Patch by: José Roberto de Souza <zehortigoza@profusion.mobi> SVN revision: 81415
2012-12-20edbus: Only create one ref of proxys when listen object managerJosé Roberto de Souza
Patch by: José Roberto de Souza <zehortigoza@profusion.mobi> SVN revision: 81414
2012-12-19Edje: delete real part object after swallows got handledLeandro Dorileo
Case the real part object(rp->object) is an smart object it start to delete the whole smart object hierarchy and a child object may be a swallow of this real part. So just delete the rp->object after swallows got handled. SVN revision: 81403
2012-12-19EPhysics: on velocity example unregister callback on restartLeandro Dorileo
Unregister callback on restart and avoid trying to use an to be deleted body. SVN revision: 81401
2012-12-19edbus: (de)yodafy free cb functionsLucas De Marchi
edbus_connection_cb_free_add -> edbus_connection_free_cb_add edbus_connection_cb_free_del -> edbus_connection_free_cb_del edbus_object_cb_free_add -> edbus_object_free_cb_add edbus_object_cb_free_del -> edbus_object_free_cb_del edbus_pending_cb_free_add -> edbus_pending_free_cb_add edbus_pending_cb_free_del -> edbus_pending_free_cb_del edbus_proxy_cb_free_add -> edbus_proxy_free_cb_add edbus_proxy_cb_free_del -> edbus_proxy_free_cb_del edbus_signal_handler_cb_free_add -> edbus_signal_handler_free_cb_add edbus_signal_handler_cb_free_del -> edbus_signal_handler_free_cb_del SVN revision: 81389
2012-12-19edbus: Fix leftover unref after edbus_service_signal_send changeLucas De Marchi
SVN revision: 81388
2012-12-19edje: Fix memleak in codegen.Stefan Schmidt
Free apiname before we return. SVN revision: 81335
2012-12-18edbus: unexport edbus_message_signal_new()Lucas De Marchi
Signals need to be sent with edbus_service_signal_emit() -- for basic messages -- or edbus_service_signal_new() + edbus_service_signal_send -- for complex messages. Otherwise it's possible to send signals that are not in the service introspection or that have different signatures by mistake/typo. SVN revision: 81311
2012-12-18edbus: unref sent signal, also for complex typesLucas De Marchi
SVN revision: 81310
2012-12-18ethumb: better naming for signal argumentsLucas De Marchi
having the same argument names in a D-Bus signal/method is the equivalent in C to have a function with this signature: int my_func(int a, int a, int a, int a); Don't. SVN revision: 81309
2012-12-18ethumb: don't use edbus_service_signal_new()Lucas De Marchi
this function is being removed from edbus. SVN revision: 81308
2012-12-18EPhysics: add soft body light implementationLeandro Dorileo
Do not use the evas map 3d light util function for every slice and avoid the unexpected "mosaic" effect. SVN revision: 81306
2012-12-18ethumb: rename edbus functionsLucas De Marchi
SVN revision: 81298
2012-12-18efreet: rename edbus functionsLucas De Marchi
SVN revision: 81296
2012-12-18edbus: rename functions edbus_message{_iter,}_arguments_setLucas De Marchi
Rename these functions since they do not set the data in the iterator/message but rather they append the data. Also improve the documentation of edbus_message_iter_arguments_append() to clarify its usage. SVN revision: 81295
2012-12-18efl: Adding Evas doc images.Jonas M. Gastal
SVN revision: 81276
2012-12-18efl: Adding Ecore doc images.Jonas M. Gastal
SVN revision: 81275
2012-12-18Fix for edje_cc that didn't like insertion of element before the first elementDaniel Zaoui
Signed-off-by: Daniel Zaoui <daniel.zaoui@samsung.com> SVN revision: 81252
2012-12-18[edje/edje_entry] Before cursor is moved by mouse_down, imf_context_reset ↵WooHyun Jung
should be called with current cursor position. SVN revision: 81209
2012-12-18edje: ecore_imf cursor infomation update is not necessary in focus-out eventJihoon Kim
SVN revision: 81201
2012-12-17updating esperanto translationsMassimo Maiurana
SVN revision: 81195
2012-12-17Fix documentationEduardo de Barros Lima
SVN revision: 81181
2012-12-17edbus: Fix invalid access memory in edbus_service_interface_unregister()José Roberto de Souza
_object_unregister is called synchronized by libdbus, so when _interface_free() ran your object its already freed. ==30579== Invalid read of size 4 ==30579== at 0x4775190: _find_object_manager_parent (edbus_service.c:803) ==30579== by 0x4775292: _interface_free (edbus_service.c:1011) ==30579== by 0x4777F1D: edbus_service_interface_unregister (edbus_service.c:1101) ==30579== by 0x40CBD28: elm_dbus_menu_delete (elm_dbus_menu.c:128) ==30579== by 0x414552F: _elm_menu_smart_del (elm_menu.c:562) ==30579== by 0x4810F39: _eo_op_internal (eo.c:363) ==30579== by 0x4812E1B: eo_do_internal (eo.c:403) ==30579== by 0x4279D02: evas_object_smart_del (evas_object_smart.c:1080) Patch by: José Roberto de Souza <zehortigoza@profusion.mobi> SVN revision: 81180
2012-12-17edbus: fix missing blank linesLucas De Marchi
Apparently the previous sed script didn't caught all missing blank lines. SVN revision: 81179
2012-12-17efreet: rename DISCONNECTED event after edbus changeLucas De Marchi
SVN revision: 81178
2012-12-17edbus: rename DISCONNECTED eventLucas De Marchi
SVN revision: 81177
2012-12-17edbus: Fix formattingLucas De Marchi
SVN revision: 81176
2012-12-17edbus: simplify end condition check in loopLucas De Marchi
SVN revision: 81175
2012-12-17edbus: avoid using dbus_signature_iter_get_signature when not neededLucas De Marchi
SVN revision: 81174
2012-12-17edbus: Use eina_log in client example and generalize expected valuesLucas De Marchi
Keep expected values in a struct. It would be ideal to have all values and compare functions in an array, so we would be able to set the same callback function for all methods. But it's already short enough so keep it as is. The usage of eina_log here allows us to easily catch which test failed. SVN revision: 81173
2012-12-17edbus: Fix typo in error messageLucas De Marchi
SVN revision: 81172
2012-12-17edbus: Don't use "bool" as a variable nameLucas De Marchi
Using bool as variable name is asking for trouble in C99. If we end up including stdbool.h, this would fail. SVN revision: 81171
2012-12-17edbus: Let arguments_get fail if getting less args than passedLucas De Marchi
If user passed a string with more elements, return EINA_FALSE on edbus_message_arguments_get() so he knows not all elements are initialized. Before this patch, we would notify user of its error if he did something like: i) edbus_message_arguments_get(msg, "uu", &a) ii) edbus_message_arguments_get(msg, "uu", &a, &b) And "msg" containing only 1 argument. This also fixes the case in which user is getting the elements of an array iterator and the array is empty. We were previously returning EINA_TRUE, even if we were not filling the variable. Last but not least, if the user was calling edbus_message_iter_arguments_get() in an empty array, we would return EINA_FALSE, even if we didn't actually get any element. SVN revision: 81170
2012-12-17edbus: do not use dbus_signature_iter_get_signature in ↵Lucas De Marchi
edbus_message_iter_arguments_vget we are only interested in the first char of the signature so we can use dbus_signature_iter_get_current_type and: a) avoid the allocation of the signature for each complete type b) simplify the check for struct and dict, since *_get_current_type() does TheRightThing (TM) This also rename some variables to clarify the new semantics: iter_type -> iter sig_type -> sig_iter SVN revision: 81169
2012-12-14EPhysics: suppress conditional jump warnings on bulletLeandro Dorileo
SVN revision: 80987
2012-12-14EPhysics: simplify the restacking data free routineLeandro Dorileo
SVN revision: 80986
2012-12-14EPhysics: free body and constraint list on testLeandro Dorileo
SVN revision: 80985
2012-12-14edbus: Improve doc of edbus_{private_,}connection_getLucas De Marchi
SVN revision: 80968
2012-12-14edbus: add edbus.spec to gitignoreLucas De Marchi
SVN revision: 80967
2012-12-14edbus: Remove useless definesLucas De Marchi
These values are used only using they directly are more meaninful than those defines. SVN revision: 80966
2012-12-14edbus: remove unneeded codeLucas De Marchi
SVN revision: 80965
2012-12-14edbus: if you goto end, you don't use elseLucas De Marchi
SVN revision: 80964
2012-12-14edbus: Add missing EINA_WARN_UNUSED_RESULT to service APILucas De Marchi
SVN revision: 80963
2012-12-14edbus: Add blank lines to separate function docsLucas De Marchi
find . -name '*.h' -exec \ sed -i '/^\s*$/b; N; /\n\/\*\*/{P;x;p;x;D}' {} \; This ensures to add a blank line above a doxygen comment if and only if there wasn't one. SVN revision: 80962