I used some c++11 magic and renewed the signal and callback classes. They are now classes with variadic templates and accept as many arguments as needed. Also removed the Slot(-container) class, because it's useless in my opinion. Maybe some documentation could be made.

I know that this is a big cut, because I removed the Slot class, but really, you don't need it. I understand if you don't accept the new signal class, but at least the callback thing - commit 6e9a8e2, a8cdc7f, f5d02b3, d060716, 9ec1533 - aren't breaking any API (at least it shouldn't). Let me hear your doubt about it.

This patch removes the following ClanLib modules: GUI (replaced by uicore) CSSLayout (no longer needed when GUI is gone) GameIDE (no community interest) Physics2D (no maintainer) Physics3D (no community interest) Scene3D (no community interest) SWRender (no community interest; obsolete - we live in a GPU world now) Database (no community interest) Sqlite (no community interest) Compute (no community interest; semi-obsolete in gaming context now that OpenGL has compute shaders)

• Templatify NetGameEventDispatcher to allow arbitrary number of parameters.
• Remove redundant constructors for NetGameEvent.
• Moved NetGameEvent::to_string(const NetGameEventValue &) to a static function in NetGameEventValue.
• Rename NetGameEventValue::to_x() to ::get_x() because of new ::to_string().

I used some c++11 magic and renewed the signal and callback classes. They are now classes with variadic templates and accept as many arguments as possible. Also removed the Slot(-container) class, because it's useless in my opinion. Maybe some documentation could be made.

I know that this is a big cut, because I removed the Slot class, but really, you don't need it. I understand if you don't accept the new signal class, but at least the callback thing - commit 6e9a8e2, a8cdc7f, f5d02b3, d060716, 9ec1533 - aren't breaking any API (at least it shouldn't). Let me hear your doubt about it.

In the clanlib's architecture is widely practiced that behind the class there is an implementation. This speeds up the compilation of the library itself at the expense of the optimality of the final code. Is it generally justified for libraries?

IODevice is written in С 32bits style with int to use for indicate the amounts instead size_t, to return -1 as error. It still works but not safety on a 64-bit architecture.

I would like to hear the developers' opinion - will accepted the commit that change int to size_t and -1 to SIZE_MAX?

Replace assigning with member initialization. Add virtual destructors to basic classes. Replace c-style cast with cpp-style cast. Some minor style improvements.

I'm not a friend of this class, because it has nearly 3k lines of code (unmaintable) and is not used currently. There are good libraries out there and they are with good performance. Also rombust I'd like to ask you to test your code before commiting it, e.g. build one/two examples ;) How about automatic testing (e.g. all examples, as soon as a new commit has been made?) A jenkins server should do the job.

Hi all,

I was tinkering around with the OSX project and made some changes. I thought I would upload them so you could review and comment.

I was mainly trying to get the signals and slots to work for the mouse and keyboard. Currently, the events and callbacks almost work correctly. However, I ran into a problem where I had to comment out part of the "if" check in signal_v4.h:

void invoke(Param1 param1, Param2 param2, Param3 param3, Param4 param4) const
{
    std::vector< std::shared_ptr<SlotCallback> > callbacks = impl->connected_slots;
    std::vector< std::shared_ptr<SlotCallback> >::size_type i, size;
    size = callbacks.size();
    for (i = 0; i < size; i++)
        if (/*callbacks[i]->valid &&*/ callbacks[i]->enabled)
            ((SlotCallback_v4<Param1, Param2, Param3, Param4> *) callbacks[i].get())->invoke(param1, param2, param3, param4);
}

I am not sure why the connected_slots are flagged as invalid. There's probably an error somewhere in my OSX implementation. They seem to work if you ignore the "valid" check. Hopefully, someone knows what would cause that.

The event data itself also needs some work.

The keyboard text data should be correct. However, the other values like ID and repeat count are wrong. The mouse data has similar issues. The mouse button ID should be correct. However, values like position will not be correct at the moment.

The application window as a whole was unresponsive at first. I threaded the main call in applicationDidFinishLaunching so that function could return and give control back to the framework. Now, I don't know if that's the ideal solution. However, it fixed the unresponsiveness and I didn't notice and crazy concurrency problems (maybe they just haven't surfaced yet?). You can now resize the window and minimize it etc.

Let me know what you think. Honestly, a lot of this code probably needs reviewed and sanitized. I don't fully understand your code base very well. So, I was just looking at how the Windows version was implemented and pushing forward.

Thanks. Have a good one.

The payment for the dedicated server hosting the ClanLib web pages will not be renewed in February 2018 (due to inactivity).

Links to "clanlib.org" in the source should be modified to point to : https://github.com/sphair/ClanLib

The unit used for mouse positioning was switched from Device Dependent Pixel (hardware pixel unit) to Device Independent Pixel (DPI-scaled pixel unit) in the DPR patch on January. This commit brings back DDP positioning for pointing devices and adds DIP positioning into clan::InputDevice as separate functions.

• InputDevice::get_position() and InputEvent::mouse_pos are in integer hardware pixel units like in ClanLib 3.x.
• InputDevice::get_dip_position() and InputEvent::mouse_dip_pos are in the new floating-point DIP pixel unit.

Note that the UI API has not changed at all; the module still uses DIP unit on everything. PointerEvent's pos() and set_pos() uses the DIP unit.

P/S: Implemented for Linux and Windows; only tested on Linux. I also removed Doxygen comments and added override keyword on the member functions of InputDeviceProvider implementors

i can't compile the MapMode example. it fails with

g++ -I Sources -I ./ pkg-config --cflags clanApp-4.1 clanDisplay-4.1 clanCore-4.1 clanGL-4.1 clanUI-4.1 -pthread -c Sources/options.cpp -o Sources/options.o g++ -I Sources -I ./ pkg-config --cflags clanApp-4.1 clanDisplay-4.1 clanCore-4.1 clanGL-4.1 clanUI-4.1 -pthread -c Sources/app.cpp -o Sources/app.o g++ -I Sources -I ./ pkg-config --cflags clanApp-4.1 clanDisplay-4.1 clanCore-4.1 clanGL-4.1 clanUI-4.1 -pthread -c ../../ThemeAero/Sources/theme.cpp -o ../../ThemeAero/Sources/theme.o ../../ThemeAero/Sources/theme.cpp: In static member function ‘static std::shared_ptrclan::ListBoxView Theme::create_listbox()’: ../../ThemeAero/Sources/theme.cpp:237:55: error: cannot bind non-const lvalue reference of type ‘std::shared_ptrclan::ScrollBarView&’ to an rvalue of type ‘std::shared_ptrclan::ScrollBarView’ 237 | initialize_scrollbar(listbox->scrollbar_x_view(), true); | ~~~~~~~~~~~~~~~~~~~~~~~~~^~ ../../ThemeAero/Sources/theme.cpp:52:72: note: initializing argument 1 of ‘static void Theme::initialize_scrollbar(std::shared_ptrclan::ScrollBarView&, bool)’ 52 | void Theme::initialize_scrollbar(std::shared_ptrclan::ScrollBarView &scrollbar, bool isHorizontal) | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~ ../../ThemeAero/Sources/theme.cpp:238:55: error: cannot bind non-const lvalue reference of type ‘std::shared_ptrclan::ScrollBarView&’ to an rvalue of type ‘std::shared_ptrclan::ScrollBarView’ 238 | initialize_scrollbar(listbox->scrollbar_y_view(), false); | ~~~~~~~~~~~~~~~~~~~~~~~~~^~ ../../ThemeAero/Sources/theme.cpp:52:72: note: initializing argument 1 of ‘static void Theme::initialize_scrollbar(std::shared_ptrclan::ScrollBarView&, bool)’ 52 | void Theme::initialize_scrollbar(std::shared_ptrclan::ScrollBarView &scrollbar, bool isHorizontal) | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~ make: *** [../../Makefile.conf:23: ../../ThemeAero/Sources/theme.o] Error 1`

g++ 11.2.0 is used

For my compile the Reference/html/Core.html isn't well formated. All Text on this site is in a < pre > box. my doxygen version was 1.8.17. I can't explain why.

Clanlib 4.1.0 fails to build on aarch64. Error log:

[  183s] System/detect_cpu_ext.cpp: In function 'detect_cpu_extension':
[  183s] System/detect_cpu_ext.cpp:81:4: error: impossible constraint in 'asm'
[  183s]    81 |    __cpuid((int*)cpuinfo, 0x1);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:86:4: error: impossible constraint in 'asm'
[  183s]    86 |    __cpuid((int*)cpuinfo, 0x80000000);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:90:4: error: impossible constraint in 'asm'
[  183s]    90 |    __cpuid((int*)cpuinfo, 0x80000001);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:95:4: error: impossible constraint in 'asm'
[  183s]    95 |    __cpuid((int*)cpuinfo, 0x80000000);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:99:4: error: impossible constraint in 'asm'
[  183s]    99 |    __cpuid((int*)cpuinfo, 0x80000001);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:104:4: error: impossible constraint in 'asm'
[  183s]   104 |    __cpuid((int*)cpuinfo, 0x80000000);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:108:4: error: impossible constraint in 'asm'
[  183s]   108 |    __cpuid((int*)cpuinfo, 0x80000001);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:113:4: error: impossible constraint in 'asm'
[  183s]   113 |    __cpuid((int*)cpuinfo, 0x1);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:118:4: error: impossible constraint in 'asm'
[  183s]   118 |    __cpuid((int*)cpuinfo, 0x1);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:123:4: error: impossible constraint in 'asm'
[  183s]   123 |    __cpuid((int*)cpuinfo, 0x1);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:128:4: error: impossible constraint in 'asm'
[  183s]   128 |    __cpuid((int*)cpuinfo, 0x1);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:134:4: error: impossible constraint in 'asm'
[  183s]   134 |    __cpuid((int*)cpuinfo, 0x80000000);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:138:4: error: impossible constraint in 'asm'
[  183s]   138 |    __cpuid((int*)cpuinfo, 0x80000001);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:143:4: error: impossible constraint in 'asm'
[  183s]   143 |    __cpuid((int*)cpuinfo, 0x1);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:148:4: error: impossible constraint in 'asm'
[  183s]   148 |    __cpuid((int*)cpuinfo, 0x1);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:153:4: error: impossible constraint in 'asm'
[  183s]   153 |    __cpuid((int*)cpuinfo, 0x80000000);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:157:4: error: impossible constraint in 'asm'
[  183s]   157 |    __cpuid((int*)cpuinfo, 0x80000001);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:162:4: error: impossible constraint in 'asm'
[  183s]   162 |    __cpuid((int*)cpuinfo, 0x1);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:167:4: error: impossible constraint in 'asm'
[  183s]   167 |    __cpuid((int*)cpuinfo, 0x1);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:172:4: error: impossible constraint in 'asm'
[  183s]   172 |    __cpuid((int*)cpuinfo, 0x1);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:177:4: error: impossible constraint in 'asm'
[  183s]   177 |    __cpuid((int*)cpuinfo, 0x80000000);
[  183s]       |    ^
[  183s] System/detect_cpu_ext.cpp:181:4: error: impossible constraint in 'asm'
[  183s]   181 |    __cpuid((int*)cpuinfo, 0x80000001);
[  183s]       |    ^

Was trying to get super methane brothers to compile from the dead again.

https://sourceforge.net/projects/methane/

However it needs ClanLib 2.2 , as clanlib.org as down currently there probably no hope this game building from source again. Options are to patch methane brothers to use more recent version of clanlib.

This seems counter-intuitive auto result = clan::PathHelp::get_basepath("c:\ABC\DEF", PathHelp::path_type_file); result --> "\ABC". The c: is chopped off.

The code below creates window, which is not always on top

`

// Set the window description
clan::DisplayWindowDescription desc_window;
desc_window.set_title("Motion Learning");
desc_window.set_topmost();
desc_window.set_popup_window();
desc_window.set_allow_resize(false);
desc_window.set_layered(true);
desc_window.show_caption(false);
desc_window.set_size(clan::Size(1792, 828), false);

// Open the windows
window = clan::DisplayWindow(desc_window);

`