gtkmm merge requestshttps://gitlab.gnome.org/GNOME/gtkmm/-/merge_requests2021-04-06T13:02:52Zhttps://gitlab.gnome.org/GNOME/gtkmm/-/merge_requests/63treeiter—Delete TreeValueProxy copy-assign/Avoid warning2021-04-06T13:02:52ZDaniel Bolestreeiter—Delete TreeValueProxy copy-assign/Avoid warning>>>
The user-provided, unimplemented operator=(TreeValueProxy const&)
existed to prevent the class being copied, but since a long time C++
provides a better way to do that: declaring the operator as deleted.
Doing this also avoids warnin...>>>
The user-provided, unimplemented operator=(TreeValueProxy const&)
existed to prevent the class being copied, but since a long time C++
provides a better way to do that: declaring the operator as deleted.
Doing this also avoids warnings about the implicitly declared default
copy constructor, which is deprecated due to the previously user-given
operator=, and since the latter is now deleted the warnings are avoided.
>>>
resolves https://gitlab.gnome.org/GNOME/gtkmm/-/issues/94https://gitlab.gnome.org/GNOME/gtkmm/-/merge_requests/23tests/builder: Test derived props declared in C++2019-12-01T12:52:31ZDaniel Bolestests/builder: Test derived props declared in C++```
tests/builder: Test derived props declared in C++,
by giving DerivedButton one and then setting it in the Builder .ui file.
```
I'm not just pushing this immediately because
* I welcome feedback on how to do it...```
tests/builder: Test derived props declared in C++,
by giving DerivedButton one and then setting it in the Builder .ui file.
```
I'm not just pushing this immediately because
* I welcome feedback on how to do it more nicely
* It should probably also include a 'real' demo, and/or equivalent example in `gtkmm-documentation`
* I don't have a (more C++17-ish) version for `master` as that doesn't build at all right now, especially `Builder`; see #61
* I'm sure I couldn't get this (setting derived properties declared in C++ in a `GtkBuilder` `.ui` file) to work in another case, so am I perhaps missing a key detail that would break more complex situations?
However, it does seem to work and test useful things even here in `gtkmm-3-24`.https://gitlab.gnome.org/GNOME/gtkmm/-/merge_requests/22add Gtk::Button::unset_image() and remove mentions of nullptr in set_image()2019-11-24T13:34:24ZDaniel Bolesadd Gtk::Button::unset_image() and remove mentions of nullptr in set_image()Fix #58
The doc fix can go into `gtkmm-3-24`. The new method presumably wants to wait until `gtkmm-3-26` with the rest, unless we decide to change that policy.
`master` has no `set_image()`, so neither are applicable there.Fix #58
The doc fix can go into `gtkmm-3-24`. The new method presumably wants to wait until `gtkmm-3-26` with the rest, unless we decide to change that policy.
`master` has no `set_image()`, so neither are applicable there.https://gitlab.gnome.org/GNOME/gtkmm/-/merge_requests/19Gdk::Cursor|Window: Fix leaks/docs. Replace/deprecate methods using Cairo::Su...2023-06-23T08:54:30ZDaniel BolesGdk::Cursor|Window: Fix leaks/docs. Replace/deprecate methods using Cairo::Surface that should use ImageSurfaceThis would close https://gitlab.gnome.org/GNOME/gtkmm/issues/30
Note that
* this is currently based on `gtkmm-3-24` but of course should wait until we branch to `gtkmm-3-26` - and indeed assumes that in its `newin` annotations
* other...This would close https://gitlab.gnome.org/GNOME/gtkmm/issues/30
Note that
* this is currently based on `gtkmm-3-24` but of course should wait until we branch to `gtkmm-3-26` - and indeed assumes that in its `newin` annotations
* others may want to make alternative proposals for the names of the new functions or how I wrote the deprecation notices
***
```
commit 3ec71b5eed8fd016c02b139533adb46e58da36a9 (HEAD -> wip/dboles/Gdk_ImageSurface, origin/wip/dboles/Gdk_ImageSurface)
Author: Daniel Boles <dboles.src@gmail.com>
Date: Thu Jun 20 18:37:14 2019 +0100
Cursor: Add replacements with correct ImageSurface
create() and get_surface() should've taken and returned
RefPtr<ImageSurface>, but they used RefPtr<Surface> instead.
Deprecate the old versions, and add new ones that replace them, by
overloading create() and renaming get_surface() to get_image_surface().
https://gitlab.gnome.org/GNOME/gtkmm/issues/30
```
***
```
commit 9e100a52b133b635a7197608418e7a4726cc9d72
Author: Daniel Boles <dboles.src@gmail.com>
Date: Thu Jun 20 18:37:10 2019 +0100
GdkWindow: Fix create_similar_image_surface return
This should return a Cairo::ImageSurface, but it returned just a Surface
We can't change the return type of the existing function in 3 because
that would break ABI, but we can add a new function suffixed _derived
that returns the correct RefPtr type and deprecate the older symbol.
https://gitlab.gnome.org/GNOME/gtkmm/issues/30
```
***
```
commit 265be45408d2b5a8ace4e091b8a6c36851eb635c (wip/dboles/Gdk_ImageSurface_refs)
Author: Daniel Boles <dboles.src@gmail.com>
Date: Thu Jun 20 18:49:05 2019 +0100
Window: Do not add 2nd ref to new Cairo::Surface
The C functions already return a newly allocated objected that the
caller owns, so if we pass `false` as Cairo's `has_reference`, then we
add another on the gtkmm side, so we will end up leaking that object.
https://gitlab.gnome.org/GNOME/gtkmm/merge_requests/19#note_535429
```
***
```
commit 769933b202813b35b7f0448e850d4823f3fbd8aa
Author: Daniel Boles <dboles.src@gmail.com>
Date: Thu Jun 20 18:38:44 2019 +0100
Cursor: Fix meaning of comment about Cairo ref arg
Cairo::Object takes a has_reference parameter, not take_copy with
inverted meaning as in glibmm/gtkmm.
Say "do not take ref" like e.g. window.hg already did, which is a bit
clearer than has_reference (when does it get that? before or after us?)
https://gitlab.gnome.org/GNOME/gtkmm/merge_requests/19#note_535429
https://gitlab.gnome.org/GNOME/gtkmm/merge_requests/19#note_536603
```Daniel BolesDaniel Boleshttps://gitlab.gnome.org/GNOME/gtkmm/-/merge_requests/14Gtk::Grid: Default parameter value for set_column_homogeneous() (gtkmm-3-24)2019-05-20T09:53:03ZPatrick StorzGtk::Grid: Default parameter value for set_column_homogeneous() (gtkmm-3-24)This achieves consistency with `set_row_homogeneous()` which already defaults to `true`.
MR for master is !15This achieves consistency with `set_row_homogeneous()` which already defaults to `true`.
MR for master is !15