Creating many native GUI frontends for a cross-platform application - c++

I've been away from GUI programming for quite some time so please pardon my ignorance.
I would like to attempt the following:
Write a Mac OSX app but still be able to port to Win/Linux (i.e. C++ core with Obj-C GUI)
Avoid Qt/other toolkits on OSX (i.e. talk to Cocoa directly - I feel that many Qt apps I use stick out like sore thumbs compared to the rest of my system)
Not as important, but it would be nice to avoid Visual Studio if it means I can have the freedom to use newer C++ features even on Windows if they help create better code.
I believe this configuration might get me what I'm looking for:
Core C++ Static Library
OSX GUI (Cocoa)
Windows GUI (Qt+MinGW?) OR (no new C++ features, Visual Studio + ManagedC++/C#/????)
Linux GUI (Qt)
Once again, sorry for my ignorance but is this possible? Is this sane? Are there any real-world open source examples accomplish something like this?

There is quite a few OS X applications that have completely custom-designed looks that don't use very many stock controls. iStat Menus comes to mind, but there are many other examples. They still look good, but it's done by manually designing them to look good and to "mesh" with the overall look of OS X applications. Even their preferences pane doesn't use stock buttons.
Thus, you can go quite far using Qt, you just have to pay close attention to what you're doing - similarly to the way other developers are paying close attention even when using Cocoa. You'll find that Qt's controls offer functionality often above and beyond what's offered in Cocoa.
That said, on OS X sometimes you may need to run some native code that expects a CFRunLoop to be present. It's good to know that Qt's event loop already spins a runloop for you, so as long as you have an event loop spinning in a given thread, you can use runloop-based code - the default runloop is provided by Qt's implementation of QEventDispatcher (somewhere in its guts). For non-gui threads, the unmodified QThread does it for you. This is useful for using asynchronous IOKit functionality, for example. Another answer of mine presents some Cocoa mouse event grabbing code. A previous version that used Carbon can be found in the edit history of that answer.
Same goes for Windows: Qt runs a message sink for all top-level windows it owns, and you can integrate native controls/windows using qtwinmigrate. You can also integrate ActiveX controls using the Active Qt framework.

Well I think you should try Qt even on OSX. Qt allows native/custom look of applications (those cases you mentioned are probably bad examples - you probably haven't noticed that lots of other applications also use Qt).
Tools I usually use for multi-platform development:
C++ (now C++11 since all major compilers more or less support it)
Boost
Qt
CMake as build system generator
If you use this tool-set you can choose whichever platform you like for development and still be multi-platform without extensive work on the other platforms.

Related

QML running with Qt for Embedded System on uclinux?

I only write QML app with Qt 5.1.1 running on Mac & Windows.
Anyone know what's the version of Qt for Embedded System integrated in the uclinux?
Do the Qt on uclinux support QML? Do I need to re-implement the UI with Qt widgets in C++ to run it on the uclinux?
First and foremost, please do not ask several subquestions in a question.
Anyone know what's the version of Qt for Embedded System integrated in the uclinux?
There is no such a thing. Qt 5 is meant to be well supported for embedded, too, without external third-party projects.
That being said, you may ask Digia about their embedded plans, like Boot2Qt and so on.
Do the Qt on uclinux support QML?
That is the wrong question around. The correct question is whether Qt/QML suppports uclinux. I do not think this will work off-hand, no. You will probably start writing a custom mkspecs file for your scenario and fix lots of issues.
Do I need to re-implement the UI with Qt widgets in C++ to run it on the uclinux?
That would probably be even worse as they are having a lot of overhead in terms additional layer. What I would suggest is to forget about QML for now, and try to bring up a simple screen with QScreen.
As stated, I do not think even that will work off-hand, but if you are enthusiastic, you could get it work.
Since Qt 5 has a hard dependency on the standard library these days, your first task is to get that building againt your mmu-free uclibc. There was some discussion about that here.
But honestly, you may be better off with some lightweight gui framework that supports framebuffer for rendering. Qt is big and not properly tested on minimal systems.

Which, if any, achieves Windows native look: GTK+, wxWidgets, Qt, FLTK?

I need to write an application that will be visually indistinguishable from something written natively for Windows XP/Vista/7 using whatever comes by default with the most modern Visual Studio. But I'm developing using MinGW and Vim (in C++).
In particular, I want the following controls to be native on the above three versions of Windows: form chrome, buttons, check boxes, menus, combo boxes, progress bars, scrollbars, rich text boxes. This will be enough for me.
I know that if you load GdiPlus and other things like riched32.dll as needed, and use Windows API to instantiate controls, then the OS will substitute its version of GdiPlus or other library, so it will look like XP style controls on XP, Vista on Vista, etc.
But I don't want to use plain Windows API, because even retrieving the default font takes half a page of code, and similar stories whatever I want to do. So I'd like to use a toolkit.
wxWidgets, Qt, GTK+, FLTK seem like the most widely used. But they are all cross-platform. I've used cross-platform applications, and many of them have foreign GUI controls (I call them widgets). So my question is: which of these toolkits can be made to produce true native-looking UI controls listed above, appearing correctly on the three versions of MSWin listed above?
I've typed each of them +" windows" into Google Images, but it's hard to tell, except that FLTK probably can't do it. Many of you must know the answer off the top of your head...
I won't talk about FLTK as I don't know it.
wxWidgets uses the native toolkit of the platform, (GTK on Linux, Win32 GUI API
on Windows, Cocoa on MacOS X).
GTK uses a theming API to fake the look and feel of the platform (custom theming engine on GTK2, CSS-based engine on GTK3).
Qt uses styles to fake the look and feel of the platform.
wxWidgets API is quite ugly from my own experience, because it had too many method just available on one or the other platform making stuff non-portable unless you'd workaround it. Unlike GTK+ and Qt, it also adds its own layer of bugs above the toolkit it uses as a backend. However, it tries hard to have the platform's native look as it uses the native toolkit.
GTK+ 3 still has some rough edges on Windows, which it officially supports since GTK+ 3.6. The GTK+ project delegates to the MSYS2 project the distribution of Windows binaries. As you're already using MinGW, that's pretty much the same kind of environment. They have good C++ bindings with GTKmm. However, you may have some work to get the theming right for your version of Windows.
Qt is a good choice for cross-platform C++ development with the main target being Windows, tries to mimic the native look and feel of the platform but has its own theming limitations too.
To sum up, there are only 2 approches:
toolkits that provide their own widgets and try to look like the native platform by providing theming (GTK+ and Qt)
toolkits that use the native widgets but hide their API behind a layer of abstraction (wxWidgets)
Both have their pros and cons.
Implementation details aside, wxWidgets philosophy is, and has always been, to look as natively as possible. We, wxWidgets developers, don't always achieve the goal of looking indistinguishably from the native applications but we always strive to do it and. AFAIK this is not such an important goal for Qt and definitely not for GTK+, so in my (obviously biased) opinion, wxWidgets is your best choice if you are serious about providing the best experience for your users, especially under OS X.
To answer your question more precisely, everything you list above is implemented using native controls in wxWidgets for Windows (rich text control is not available natively under the other platforms though).
IUP - Portable User Interface library uses native widgets, C API and Lua bindings.
i used java for native cross-platform without changing the code, used c/c++ wxwidgets for exclusively cross-platform if you want go to little up performance and standalone executable, used c/c++ winapi for windows and x11 for gnu linux native platform and terminal console, used python for scripting console and platform if you want your software up to date fast, and used assembly for a little simple purely console. And sometimes i combined them all with shared library .dll on windows and .so on gnu linux. And i liked doing for do comparative performance on programming studies with small hardware requirements.

OS X from Carbon to Cocoa window creation?

Edit: This question has been answered indirectly in the comments (using GLUT to avoid the need for either) but because its not posted as an answer, I cannot "accept" it although my issue is solved.
I have a cross platform OpenGL application, written in C++, that (on Mac OS X) uses Carbon for the window creation and message pump (Think...NeHe Lesson 1). Recent changes to my code utilizing C++11 features has made it incompatible with the GCC that comes bundled with XCode.
All of my attempts to use systems like MacPorts, HomeBrew, etc... have failed for different reasons (Mostly due to my need to compile 32bit for Carbon support). I was originally going to ask if anyone had a better solution for installing an updated G++ in Mac OS X 10.8.
That seems like a short sighted solution (at some point, I will need to upgrade to Cocoa and 64bit) so my question has evolved to asking what would be the most simple way to have my existing (very large) application utilize Cocoa for this task rather than Carbon?
I have all of the "Carbon Code" in its own separate .cpp/.h file, which has only two functions, one to create an OpenGL window, and one that is a message pump that uses ReceiveNextEvent.
Edit:
To be clearer, I am looking for the best way to design (more or less) a drop in replacement for my existing Carbon window creation, and message pump functions, but in Cocoa. I understand that being a different API, using different language concepts, that a simple one to one command replacement will not suffice. My existing code has a source code file for Mac that has two functions which perform core tasks (with Carbon) that I call, InitWindow, and DoEvents. These create the window, and cycle through the pending events respectively. It is my goal to replace this source code file with one that uses Cocoa, exposing the same functions so there is little to no modification of my core code.
What would be the best way to achieve my goal? I am not familiar with Cocoa and Obj-C is alien to me.
Going to Cocoa 64-bit definitely sounds like a great idea. And I recommend using some of the native UI if you want to make your users happy, too.
The easiest thing to do would be to create an NSWindow with an NSOpenGLView in it. You can either do that in Interface Builder, or in code. It's probably best to do it by creating a new Cocoa app using Xcode's template (File > New Project > Cocoa App). That will give you a window by default. You can open it in Interface Builder and add an NSOpenGLView to it. You can make the view a custom subclass of NSOpenGLView, and add your event handling to it. See the docs for NSResponder for more information on how Cocoa dispatches events.

wxwidgets vs gtkmm with my requirements

I'm implementing a GUI program (a cards game) in gtkmm (at the moment). I want that my program runs well at least under Windows and Linux (GTK). Mac OS X would be a plus, although not required. Although I'm using gtkmm, I have some concerns with it.
More concretely:
Lack of support for connecting signals at runtime.
The API for treeview is horrible in the sense that you need to make your own classes
in c++ code every tie you have a new Treemodel.
These two things get a lot in my way when implementing GUI code. My questions are, if I switch to wxwidgets:
1.- Is it possible to connect signals at runtime?
2.- Is the ListCtrl API from Wxwidgets comfortable to use in your opinion?
3.- Is there a decent gui builder (preferably free, but not required) in which I don't have to connect signals by hand in c++ code? This is quite important to keep controller and view code separated.
4.- Is there an alternative to wxwidgets which is better and looks good under gtkmm, besides working on Windows. I'd rather not use qt with its preprocessing step
Thanks in advance.
It is possible.
ListCtrl if just fine to use, and you will maybe find some nicer alternative (UltimateListCtrl, wxGrid...).
wxFormBuilder.
Not that I know of.

C++ UI framework from scratch?

I want to create a C++ UI framework (something like QT or like ubuntu unity Desktop)
How is programmed , is it using OpenGL or lets take plasma ui of QT (how is this programmed )?
Direct answers , reference links anything will be helpful.
Some interesting opengl based UI I founf on the web
LiquidEngine
http://www.youtube.com/watch?v=k0saaAIjIEY
Libnui
en.wikipedia.org/wiki/Libnui
Some UI frameworks render everything themselves, and work based on some kind of clipping-window-within-the-host-systems-screen. Non-display aspects (such as input event handling) have to be translated to/from the host systems underlying APIs.
Some UI frameworks translate as much as possible to some underlying framework.
wxWidgets can do both. You can choose a native version (e.g. wxMSW if you're on Windows) and most wxWidgets controls will be implemented using native Windows controls. Equally, you can choose the wxUniversal version, where all controls are implemented by the wxWidgets library itself.
The trouble is that typical GUI frameworks are huge. If you want a more manageable example to imitate, you might look at FLTK. I haven't got around to studying it myself, but it has a reputation for being consise.
There are also some GUI toolkits that are specifically aimed at games programming, such as Crazy Eddies GUI. My guess - these are probably as idependent of the underlying API as possible, so that particular applications can implement the mapping to whichever underlying API they happen to target (OpenGL, DirectX, SDL, whatever) and can be the boss of the GUI rather than visa versa.
http://www.wxwidgets.org/
http://www.fltk.org/
http://www.cegui.org.uk/wiki/index.php/Main_Page
"no really, don't write your own wm or toolkit"
The #Xorg-devel guys on irc.freenode.org
doing one anyway means that you have to test against a wide range of more or less buggy WMs and X implementations, and that you have to frequently update to be compatible with the latest Xorg server and X protocol features (like Xinput 2.1)
understandably, the Xorg people are tired to support old, unmaintained toolkits and applications. They already have enough bugs.
The GUI frameworks are very dependant on a windows system, which dictates what is allowed and how windows are created and rendered. For example, pass a specific option to create a borderless or full-screen window.
Since you mentioned opengl and ubuntu, I guess you want to start on a linux platform. You should study xlib, for which you can find reference here.
Since the qt library is open source, you can download it and peek into it's sources.
A UI library isn't developed from scratch. It relies on the OS' windowing system, which relies on the driver from your graphics adapter, which relies on the OS kernel, which relies on... and so on.
To develop any software "from scratch", you can start by writing your own BIOS. Once you're done with that, move on to writing an OS, and then you should be just about ready to write the software you wanted. Good luck.
And this is assuming you're willing to cheat, of course, and use a compiler you didn't write from scratch.
Before you do that, it's worth that you spend one week on thinking:
1, Do you really know how to do it? I doubt that.
2, Do you really need to do it? I doubt that too.