summaryrefslogtreecommitdiff
path: root/glib/glibmm
Commit message (Collapse)AuthorAgeFilesLines
* Add an unwrap_copy overload for references.Krzesimir Nowak2012-07-221-0/+9
| | | | | It is provided mainly for consistency - it simply calls gobj_copy() method.
* Do not use misstyped enum name.Krzesimir Nowak2012-07-221-1/+1
| | | | | | It is not provided in gir anyway. If we want to provide it then we will have to add a way to inject enum values into _WRAP_ENUM or _WRAP_GERROR.
* Don't include deprecated thread.h.Krzesimir Nowak2012-07-221-12/+0
| | | | | | We are not wrapping it anyway. But if we want to do it, then we will have to provide some kind of override or addition for some API wrapped there - some of it is not available in gir files.
* Merge branch 'master' into gmmproc-refactorKrzesimir Nowak2012-07-1314-116/+343
|\ | | | | | | Also, removed lots of files.
| * Glib::add_exception_handler(): Fix misplaced endif.Kjell Ahlstedt2012-06-041-2/+1
| | | | | | | | | | * glib/glibmm/exceptionhandler.h: Fix a misplaced endif, which has made Doxygen exclude add_exception_handler() from the documentation.
| * Fix typos in miscutils.Krzesimir Nowak2012-04-171-2/+2
| | | | | | | | | | * glib/glibmm/miscutils.cc: Return early from get_system_data_dirs() and get_system_config_dirs() when C function returns NULL.
| * Glib::Dispatcher: Don't send messages to a deleted Dispatcher.Kjell Ahlstedt2012-04-042-12/+82
| | | | | | | | | | | | | | * glib/glibmm/dispatcher.h: Add missing usage rules. * glib/glibmm/dispatcher.cc: Avoid delivering messages to deleted Dispatchers. Don't block message delivery while a second main loop is running. Bug #651942.
| * Make SignalTimeout,SignalIdle::connect_once() more thread safe.Kjell Ahlstedt2012-04-041-7/+1
| | | | | | | | | | * glib/glibmm/main.cc: Call SourceConnectionNode::install() before g_source_attach(). Suggested by Chris Vine. Bug #396963 comments 15-18.
| * Some improvements to the documentation in the previous commit.Murray Cumming2012-04-031-51/+41
| | | | | | | | * glib/glibmm/property.h:
| * Document Glib::Property and Glib::PropertyBaseMark Vender2012-04-031-2/+86
| | | | | | | | | | * glib/glibmm/property.h: Bug #673291
| * Make SignalTimeout,SignalIdle::connect_once() more thread safe.Kjell Ahlstedt2012-04-031-6/+52
| | | | | | | | | | * glib/glibmm/main.cc: Don't create a sigc::connection in the connect_once() methods. Bug #396963.
| * Fix some warnings found by clang++.Murray Cumming2012-03-191-1/+1
| | | | | | | | | | * glib/glibmm/class.cc: registered_derived_type(): Use (void*)0 instead of NULL to avoid a missing sentintel warning.
| * Glib::Threads::Private, Glib::Dispatcher: Delete Glib::DispatchNotifier.Kjell Ahlstedt2012-03-121-2/+2
| | | | | | | | | | | | | | * glib/src/thread.hg: Add missing @deprecated. * glib/src/threads.hg: Add Private::replace(). Add documentation to Private. * glib/glibmm/dispatcher.cc: Use Private::replace() instead of set(). Bug #671587.
| * Signal*: Better description of the connect*() methods.Kjell Ahlstedt2012-03-021-14/+35
| | | | | | | | | | | | * glib/glibmm/main.h: Mention that returning false from a bool-returning handler disconnects the handler. Other minor changes of the descriptions. Bug #669906.
| * Fix the build with --disable-deprecated-api.Kjell Ahlstedt2012-02-281-0/+22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | * glib/glibmm/main.cc: Call get_time() instead of the deprecated get_current_time(), when GLIBMM_DISABLE_DEPRECATED is defined. * glib/src/thread.[hg|ccg]: Use _IS_DEPRECATED instead of _DEPRECATE_IFDEF_{START|END}. _DEPRECATE_IFDEF_{START|END} does not include code generated by gmmproc within #ifndef/#endif. * glib/src/threads.ccg: Change catch(Glib::Thread::Exit&) to catch(Glib::Threads::Thread::Exit&). * glib/src/threads.hg: Add _GMMPROC_EXTRA_NAMESPACE(Threads). * glib/src/valuearray.hg: * gio/src/application.hg: #undef G_DISABLE_DEPRECATED in the .cc file. * tools/m4/class_shared.m4: Define _IS_DEPRECATED. Bug #640029.
| * Dispatcher: Fix the build on win32.TS2012-01-301-3/+3
| | | | | | | | | | * glib/glibmm/dispatcher.cc: Mention the new Threads:: namespace in the ifdefed win32 code.
| * Restore unintended change.Murray Cumming2011-11-231-1/+1
| |
| * Add Glib::Threads::* in threads.h, deprecating everything in thread.hMurray Cumming2011-11-237-16/+49
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * glib/src/filelist.am: * glib/src/thread.[hg|ccg]: Deprecate the whole file, adding deprecation doxygen comments to all API. * glib/src/threads.[hg|ccg]: A new Threads namespace containing equivalents for everything in thread.h, implemented using only non-deprecated glib API. This was necessary because we had to break the ABI to do this. * glib/glibmm.h: Include threads.h * glib/glibmm/main.[h|cc]: Added a wait() method overload that takes the new types, deprecating the existing wait() method. * examples/network/resolver.cc: * examples/network/socket-client.cc: * examples/network/socket-server.cc: * examples/thread/dispatcher.cc: * examples/thread/dispatcher2.cc: * examples/thread/thread.cc: * examples/thread/threadpool.cc: * glib/glibmm/dispatcher.cc: * glib/glibmm/exceptionhandler.cc: * glib/glibmm/threadpool.[h|cc]: Use the new Glib::Threads::* types instead of thread.h.
| * Avoid a deprecation warning in implelementation of deprecated API.Murray Cumming2011-11-231-0/+1
| | | | | | | | | | | | * glib/glibmm/main.cc: Define GLIB_DISABLE_DEPRECATION_WARNINGS to avoid a warning about g_source_get_current_time(), which we must use in our own deprecated code.
| * class.cc: Avoid compiler warnings when using -std=c++0x with g++.Murray Cumming2011-11-231-4/+20
| | | | | | | | | | | | * glib/glibmm/class.cc: Add casts because GTypeQuery::class_size and instance_size are guint but GTypeInfo::class_size and instance_size are guint16.
| * thread.h: Allow use of this without causing deprecation compiler warnings.Murray Cumming2011-10-263-3/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * glib/src/thread.hg: undef and then redefine G_DISABLE_DEPRECATED, and define GLIB_DISABLE_DEPRECATION_WARNINGS, if G_DISABLE_DEPRECATED is defined, so we can use the deprecated API in this header (though only in our own deprecated API). * tools/generate_wrap_init.pl.in: Include glibmm.h instead of glib.h so we can include thread.h first (via glibmm.h) instead of only later. This seems more logical than explicitly including thread.h in all (even non-glibmm) generated wrap_init.cc files. * tools/m4/base.m4: * glib/glibmm.h: * glib/glibmm/dispatcher.cc: * glib/glibmm/exceptionhandler.cc: * glib/glibmm/main.cc: * examples/thread/thread.cc: * examples/thread/threadpool.cc: Adjust the includes to include thread.h first.
| * Move IOCondition from main.h to iochannel.h and autogenerate it.Murray Cumming2011-10-262-53/+2
| | | | | | | | | | | | | | | | | | | | | | * glib/src/iochannel.hg: Use _WRAP_ENUM() for IOCondition instead of hand-coding it in * glib/glibmm/main.h: though this requires some manual editing of * glib/src/glib_enums.defs: due to the weird defines that GioChannel uses for its values. See the comment. * glib/glibmm/streamiochannel.cc: Adjust the includes. This avoids the need to include glib.h from main.h.
| * Fix the make check build by always linking to gthread-2.0.Murray Cumming2011-10-251-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | * configure.ac: Always use gthread-2.0 because there is now no advantage to not doing that, because threading is now always enabled in glib. Remove the separate GTHREAD* variables. * glib/glibmm/Makefile.am: * gio/giomm/Makefile.am: * tests/Makefile.am: * tools/Makefile.am: * examples/Makefile.am: Remove use of the GTHREAD* variables.. * examples/network/resolver.cc: Remove the non-threading option and code that uses it.
* | Do not use file global AM_ variables.Krzesimir Nowak2012-07-091-3/+2
| |
* | Add Glib::unwrap() for references.Krzesimir Nowak2012-06-061-0/+16
|/ | | | | | | | glib/glibmm/wrap.h: While Glib::unwrap() for pointers (be them smart or plain C) makes sense, because it does checking for NULL pointers, then Glib::unwrap() for references is just a convenience addition, so I do not have to differentiate much when generating code basing on whether passed value is a reference or a pointer.
* Update a TODOMurray Cumming2011-04-041-1/+2
|
* ArrayHandle: Make it compilable with MSVC 2005, 2008, 2010.Krzesimir Nowak2011-03-291-43/+52
| | | | | | * glib/glibmm/arrayhandle.h: Put operator std::container methods definitions inside class template definition. Also added decrementation operators for ArrayHandleIterator for MSVC 2005.
* Correct some range checks to use >= instead of just >.Murray Cumming2011-03-211-1/+1
| | | | | | | * glib/glibmm/stringutils.cc: Glib::Ascii::strtod(): * glib/src/variant.[hg|ccg]: get(): Correct the range check. This would previously not have caught all wrong ranges. Thanks to Michael Edwards.
* Vectorutils, ArrayHandle: Check for NULL pointers.Krzesimir Nowak2011-03-214-28/+39
| | | | | | | | | | | | | | | * glib/glibmm/vectorutils.[h|cc]: Don't crash in compute_array_size2, when NULL is passed - in that case just return 0. Also, array_to_vector method return empty vector in such case. * glib/glibmm/arrayhandle.[h|cc]: Set array_size to zero if passed array is NULL. * tests/glibmm_null_vectorutils/main.cc: New test for NULL arrays and lists for vectorutils. * tests/glibmm_null_containerhandle/main.cc: New test for NULL arrays and lists for containerhandles. * tests/Makefile.am: Added new tests to build. Spotted by Kalev Lember. Bug #645245.
* Added missing files.Krzesimir Nowak2011-02-211-0/+38
|
* Mark some new API.Murray Cumming2011-02-211-0/+12
| | | | | * glib/glibmm/miscutils.h: Add @newin{2,28} to the new build_filename() overloads.
* Extended build_filename() utility functionFabrício Godoy2011-02-212-1/+173
| | | | | Extended build_filename() to accept up to nine parameters. Added a new testcase in tests/glibmm_buildfilename.
* Fix the build with --enable-warnigns=fatal.Murray Cumming2011-02-211-2/+2
| | | | | | * glib/glibmm/arrayhandle.h: Remove some typename keywords to avoid this compiler warning: "error: using ‘typename’ outside of template"
* Added bool specialization for Glib::ArrayHandle.Krzesimir Nowak2011-02-212-0/+215
| | | | | | | | | | | * glib/glibmm/arrayhandle.h: Added specialization for bool ArrayHandle. This is needed because std::vector<bool> is a specialization for which iterators does not return a reference to actual value it holds. * glib/glibmm/arrayhandle.cc: New file implementing destructor of bool ArrayHandle - the only method that is not inlined. * tests/glibmm_bool_arrayhandle/main.cc: New file implementing test checking if bool ArrayHandle actually works. * tests/Makefile.am: Added new test to build.
* Vector utils: Corrected documentation and extended it bit.Krzesimir Nowak2011-02-011-6/+167
| | | | | | * glib/glibmm/vectorutils.h: Documented important parts of *Keeper classes and corrected examples in *Handler classes. The documentation mostly tries to explain memory management.
* Vector utils: Added simple documentation.Murray Cumming2011-01-301-15/+41
| | | | | | * glib/glibmm/vectorutils.h: Some simple documentation about the *Handler utility classes, though I need to correct the example code and explain the memory mangement when that is clear to me.
* ArrayHandle, ListHandle, SListHandle: Document these as almost-deprecated.Murray Cumming2011-01-283-9/+28
| | | | | | | | | | * glib/glibmm/arrayhandle.h: * glib/glibmm/listhandle.h: * glib/glibmm/slisthandle.h: Mention that we try not to use these now, though we cannot actually deprecate them yet without forcing other libraries such as pangomm to deprecate the methods that use them, which they cannot really do because they cannot just add a method overload without introducing ambiguity.
* Fixed code style.Murray Cumming2011-01-272-200/+200
|
* Added vectorutils.h.Krzesimir Nowak2011-01-273-0/+871
| | | | | | | | | | | * glib/glibmm/vectorutils.[h|cc]: New files implementing an array, list, slist <-> std::vector convertions. It implements also a specialization for bool types. * glib/glibmm.h: Include vectorutils.h. * glib/glibmm/filelist.am: Added vectorutils to build. * tests/glibmm_vector/main.cc: New test for general vectorutils check. * tests/glibmm_bool_vector/main.cc: New test checking bool specializations. * tests/Makefile.am: Added above tests to build system.
* Revert "Gio::Application: Initialize glibmm and giomm."Murray Cumming2011-01-191-19/+4
| | | | | This reverts commit 590000d08fa94a72d042ac1143fa042c90cf9705. Initializing glibmm in ObjectBase is not actually necessary.
* Gio::Application: Initialize glibmm and giomm.Murray Cumming2011-01-181-4/+19
| | | | | | | | | * glib/glibmm/objectbase.cc: Constructors: Call Glib::init(), making it unnecessary to call this explicitly before instantiating a Glib::Object. * gio/src/application.ccg: Make sure that giomm can be used as soon as a Gio::Application has been instantiated. This should let use just instantiate a Gtk::Application instead of using Gtk::Main.
* Added a TODOMurray Cumming2010-12-221-0/+3
|
* RefPtr: Make it work with sorted containers (e.g. std::set<RefPtr<T> >).David Kozub2010-12-221-0/+44
| | | | | | * glib/glibmm/refptr.h: Add operators <, <=, >, >= that compare RefPtrs by their underlying pointer. This makes RefPtr work with sorted containers (e.g. std::set<RefPtr<T> >)
* ObjectBase: Fixed refcount debugging message.Krzesimir Nowak2010-11-081-2/+2
| | | | | | * glib/glibmm/objectbase.cc: A parameter passed to g_warning() in destroy_notify_callback() describing a name of GObject should be indeed a name, not a pointer to GObject.
* Added a TODOMurray Cumming2010-11-041-0/+2
|
* Fix the build with glib from git master.Murray Cumming2010-11-032-42/+37
| | | | | * glib/glibmm/main.[h|cc]: get_time(): Now returns a gint64 instead of a GTimeSpec output parameter.
* Fix the build with --enable-warnings=fatalMurray Cumming2010-11-012-1/+31
| | | | | | * glib/glibmm/main.[h|cc]: get_current_time(): undef G_DISABLE_DEPRECATED so this can build. Deprecate it to match the C API. Add get_time() which replaces it. However, we should replace our use of it, which seems difficult.
* Added get_system_data_dirs() and get_system_config_dirs().Murray Cumming2010-10-212-8/+55
| | | | | | | | * glib/glibmm/miscutils.[h|cc]: Wrappping g_get_system_data_dirs() and g_get_system_config_dirs(), which were actually added in glib 2.6 but not noticed by us until now. Notice that these return std::vector<> instead of ArrayHandle, because we have decided to do that for new APIs.
* glibmm: ustring: Restore last block removed by #ifdef removal script.José Alburquerque2010-10-181-1/+6
| | | | | | | | * glib/glibmm/ustring.cc: Restore the last block that was accidentally removed by the optional API #ifdef removal script. There were three. Two were restored (GNOME bug #622030), but one (final one) was not. This commit does that, restoring the file to its original condition without the optional API #ifdefs.
* Glib::Variant<>: Use gmmproc files so wrapping is not done manually.José Alburquerque2010-06-193-270/+0
| | | | | | | | | | | | | | | * glib/src/glib_functions.defs: Regenerate to get the GVariant functions. * glib/src/variant.ccg: * glib/src/variant.hg: * glib/src/filelist.am: Add the new .hg and .ccg files and mention them. Wrapping is now done using gmmproc instead of handwriting the methods. * glib/glibmm/variant.cc: * glib/glibmm/variant.h: * glib/glibmm/filelist.am: Remove these files and their mentions because now they are written as .hg and .ccg files. * glib/src/variant_basictypes.h.m4: Add docs.
View Lorry Controller Status