See also top-level FAQ page.
List of questions in this category
wxWidgets is a class library that allows you to compile graphical C++ programs on a range of
different platforms. wxWidgets defines a common API across platforms, but uses the native graphical user interface (GUI) on each platform,
so your program will take on the native 'look and feel' that users are familiar with.
Although GUI applications are mostly built programmatically, there are several dialog editors to help
build attractive dialogs and panels. Robert Roebling's wxDesigner
and Anthemion Software's DialogBlocks
are two commercial examples, but there are others: see the Useful Tools page.
You don't have to use C++ to use wxWidgets: there is a Python interface for wxWidgets,
and also a Perl interface.
Yes. Please see the licence for details, but basically
you can distribute proprietary binaries without distributing any source code, and neither will wxWidgets
conflict with GPL code you may be using or developing with it.
The conditions for using wxWidgets are the same whether you are a personal, academic
or commercial developer.
No official support, but the mailing list is very helpful and some people say that
wxWidgets support is better than for much commercial software. The developers are
keen to fix bugs as soon as possible, though obviously there are no guarantees.
Many organisations - commercial, government, and academic - across the
world. It's impossible to estimate the true number of users, since
wxWidgets is obtained by many different means, and we cannot monitor
distribution. The mailing list contains around 300-400 entries which is
quite large for a list of this type.
See Users for a list of some users and their applications, and
also Feedback for comments.
Our highest-profile user yet is industry veteran and Lotus Corp. founder Mitch Kapor
and his Open Source Applications Foundation.
- Windows 3.1, Windows 95/98, Windows NT, Windows 2000, Windows ME.
- Linux and other Unix platforms with GTK+.
- Unix with Motif or the free Motif clone Lesstif.
- Mac OS.
- Embedded platforms are being investigated. See the wxUniversal project.
- An OS/2 port is in progress, and you can also compile wxWidgets for GTK+ or Motif
This is a hotly-debated topic amongst the developers. My own philosophy
is to make wxWidgets as platform-independent as possible, but allow in a
few classes (functions, window styles) that are platform-specific.
For example, Windows metafiles and Windows 95 taskbar icons have
their own classes on Windows, but nowhere else. Because these classes
are provided and are wxWidgets-compatible, it doesn't take much
coding effort for an application programmer to add support for
some functionality that the user on a particular platform might otherwise
miss. Also, some classes that started off as platform-specific, such
as the MDI classes, have been emulated on other platforms. I can imagine
that even wxTaskBarIcon may be implemented for Unix desktops one day.
In other words, wxWidgets is not a 'lowest common denominator' approach,
but it will still be possible to write portable programs using the
core API. Forbidding some platform-specific classes would be a stupid
approach that would alienate many potential users, and encourage
the perception that toolkits such as wxWidgets are not up to the demands
of today's sophisticated applications.
Currently resources such as bitmaps and icons are handled in a platform-specific
way, but it is hoped to reduce this dependence in due course.
Another reason why wxWidgets is not a 'lowest common denominator' toolkit is that
some functionality missing on some platform has been provided using generic,
platform-independent code, such as the wxTreeCtrl and wxListCtrl classes.
No. This is a much-discussed topic that has (many times) ended with the conclusion that it is in
wxWidgets' best interests to avoid use of templates. Not all compilers can handle
templates adequately so it would dramatically reduce the number of compilers
and platforms that could be supported. It would also be undesirable to make
wxWidgets dependent on another large library that may have to be downloaded and installed.
In addition, use of templates can lead to executable bloat, which is something
wxWidgets is strenuously trying to avoid.
The standard C++ string class is not used, again because it is not available to all compilers,
and it is not necessarily a very efficient implementation. Also, we retain more flexibility
by being able to modify our own string class. Some compatibility with the string class
has been built into wxString.
There is nothing to stop an application using templates or the string class for its own
purposes. With wxWidgets debugging options on, you may find you get errors when including
STL headers. You can work around it either by switching off memory checking,
or by adding this to a header before you include any STL files:
These are the possibilities so far:
- See www.scintilla.org for
a very nice syntax-highlighting editor widget. Robin Dunn has written a wxWidgets wrapper
for this widget, available in the wxWidgets distribution under contrib/src/stc.
- If you only need to display marked-up information, rather than edit it,
then wxHTML will suit your needs. wxHTML is built into wxWidgets - please see the reference
manual for details, and samples/html.
- There are rich edit widgets in both WIN32 and GTK+, but there is currently
no wxWidgets wrapper for these (but text attribute functions are being added in the wxWidgets 2.3.x series).
wxWidgets library itself is unfortunately not exception-safe (as its
initial version predates, by far, the addition of the exceptions to the C++
language). However you can still use the exceptions in your own code and use
the other libraries using the exceptions for the error reporting together with
There are a few issues to keep in mind, though:
- You shouldn't let the exceptions propagate through wxWidgets code,
in particular you should always catch the exceptions thrown by the
functions called from an event handler in the handler itself and not
let them propagate upwards to wxWidgets.
- You may need to ensure that the compiler support for the exceptions is
enabled as, considering that wxWidgets itself doesn't use the
exceptions and turning their support on results in the library size
augmentation of 10% to 20%, it is turned off by default for a few
compilers. Moreover, for gcc (or at least its mingw version) you must
also turn on the RTTI support to be able to use the exceptions, so you
should use --disable-no_rtti --disable-no_exceptions options
when configuring the library (attention to the double negation).
We are using the CVS system to develop and maintain wxWidgets. This allows
us to make alterations and upload them instantly to the server, from
which others can update their source.
To build source from CVS, see the file BuildCVS.txt in the top-level wxWidgets distribution
By ftp, and via the wxWidgets CD-ROM.
If you are feeling adventurous, you may also check out the sources directly
wxBase is a subset of wxWidgets comprised by the non-GUI classes. It includes
wxWidgets container and primitive data type classes (including wxString,
wxDateTime and so on) and also useful wrappers for the operating system objects
such as files, processes, threads, sockets and so on. With very minor
exceptions wxBase may be used in exactly the same way as wxWidgets but it
doesn't require a GUI to run and so is ideal for creating console mode
utilities or server programs. It is also possible to create a program which can
be compiled either as a console application (using wxBase) or a GUI one (using
a full featured wxWidgets port).
The main difference between wxUniversal-based ports (such as wxX11, wxMGL) and other ports (such as wxMSW, wxGTK+, wxMac)
is that wxUniversal implements all controls (or widgets) in
wxWidgets itself thus allowing to have much more flexibility (for example, support for
themes even under MS Windows). It also means that it is now much easier to
port wxWidgets to a new platform as only the low-level classes must be ported
which make for a small part of the library.
You may find more about wxUniversal here.
The Java honeymoon period is over :-) and people are realising that it cannot
meet all their cross-platform development needs. We don't anticipate a major threat
from Java, and the level of interest in wxWidgets is as high as ever.
Microsoft is spending a lot on promoting the .NET initiative, which
is a set of languages, APIs and web service components for Windows.
Ximian has started an open source version of .NET, mostly for Linux.
C# is Microsoft's alternative to Java, supporting 'managed code',
garbage collection and various other Java-like language features.
Although this may be attractive to some developers, there
is a variety of reasons why the .NET/Mono combination is unlikely
to make wxWidgets redundant. Please note that the following comments
are Julian Smart's opinions.
There is nothing to stop folk from developing a C# version of the wxWidgets API;
Update: a wx.NET project is now in progress.
- Not everyone wants or needs net services.
- C++ will be used for a long time to come; compared with C++, C# is a recent development and its future is not certain.
- Mono Forms may only target Winelib (at least to begin with), so the end result is not as native as
wxWidgets (I'm aware there is GTK# for use with the C# language).
- C# is usually byte-compiled and therefore slower. Plus, .NET adds a layer of overhead to the client computer
that wxWidgets does not require.
- Mono hasn't proven its long-term viability yet (it's a complex system of components); wxWidgets is ready now.
- You may not wish to buy into Microsoft marketing spin and APIs.
- Microsoft may at some point sue developers of non-Microsoft .NET implementations. After all,
platform-independence is not in Microsoft's interest.
- .NET might never be implemented on some platforms, especially Mac and embedded variants of Linux.
- wxPython and other language variants provide further reasons for wxWidgets to continue.
- The same issue exists for Qt: if Qt sales remain strong, it's a good indication that
the market for a C++-based approach is still there. (Either that, or everyone's turning to wxWidgets!)
Please check out the Community pages,
in particular the suggested projects, and
mail the developers' mailing list with your own suggestions.
Please subscribe to the wx-dev developers' mailing list and
ask if anyone else is interested in helping with the port, or
has specific suggestions. Also please read the coding standards.
Each port consists of a platform-specific part (e.g. src/msw, include/wx/msw),
a generic set of widgets and dialogs for when the port doesn't support
them natively (src/generic, include/wx/generic) and the common code
that all ports use (src/common, include/wx). By browsing the source
you should get a good idea of the general pattern.
Take a port that most closely matches your port, and strip out
the implementation so you have a skeleton port that compiles. Ask on wx-dev
first for the wxStubs port - however, any such predefined skeleton
port may be out of date, so make a judgement on whether to use it.
Perhaps it will still save you time to clean up wxStubs, and
others may benefit from this too.
You will need to define a symbol for the new port, e.g. __WXXBOX__.
Look at files such as wx/defs.h, wx/wxchar.h for areas where you'll
need to add to existing conditionals to set up wide character
support and other issues. If the GUI runs on a Unix variant,
define the __UNIX__ variable in your makefile.
Then you can start implementing the port, starting with
wxWindow, wxTopLevelWindow, wxFrame, wxDialog so you
can get the minimal sample running as soon as possible.
If GDI objects (wxPen, wxBrush, etc.) are not concepts in your
native GUI, you may wish to use very generic versions of
some of these - see the wxX11 port.
Consider using the wxUniversal widget set as a quick way
to implement wxWidgets on your platform. You only need
to define some basic classes such as device contexts,
wxWindow, wxTopLevelWindow, GDI objects etc. and
the actual widgets will be drawn for you. See wxX11,
wxMGL, and wxMSW/Univ for sample wxUniversal ports.
To begin with, you can use whatever makefiles or project
files work for you. Look at existing makefiles to see what
generic/common/Unix files need to be included. Later, you'll want to integrate support
for your port into configure (Unix-like systems and gcc under Windows),
and bakefile (for other makefiles on Windows).
Submit your port as patches via SourceForge; you might
wish to separate it into one patch that touches common headers
and source files, and another containing the port-specific code, to make
it much easier for us to review and apply the patches.