New line is not always marked by 2-character (\r\n) sequence. It can be
also 1-character (\r) for RichText controls or even no mark at all (for
wrapped line).
It is a synonym for wxIMPLEMENT_ABSTRACT_CLASS and not
wxIMPLEMENT_DYNAMIC_CLASS as it was wrongly stated.
Also avoid duplicating the documentation of the other macro and just
refer to it instead.
This reverts commit c4e1fb4ef9.
Last position is not necessarily the number of characters in the buffer,
rich edit controls still store the text using CR LF between lines, but
GetLastPosition() should return the position as if they used only LF to
be consistent with all the other positions in these controls and
changing this broke existing code passing GetLastPosition() to other
functions taking position.
There is a valid position after the last character of the text in
wxTextCtrl, e.g. position 0 in the empty control, so account for it and,
notably, don't return -1 from XYToPosition(0, 0) when the control is
empty.
This fixes a regression in a69ab2907c.
In addition to showing the position returned by HitTest(), also show its
return value which is not necessarily wxTE_HT_ON_TEXT if it is not
wxTE_HT_UNKNOWN under MSW.
Use wxScopedCharBuffer as we only need the buffer in the current scope.
This avoids a buffer copy done by wxCharBuffer which could be a
significant pessimization for large buffers.
Also don't call strlen() unnecessarily, we already know the buffer
length, so just use it.
This method seems completely unnecessary, the base
wxTextEntry::DoSetValue(), which delegates to Remove() and WriteText(),
seems to work just as well and avoids code duplication between this
method and wxTextCtrl::WriteText().
Notice that gtk_text_buffer_set_text() is just a trivial wrapper around
gtk_text_buffer_delete() and gtk_text_buffer_insert() anyhow, so there
is no efficiency loss in not using it neither.
This could happen if the code calling SetValue() was in a handler
invoked in response to a keyboard action as our insert-text signal
handler believed that it was called in response to this key in this
case, even when it wasn't the case.
This extends the changes of def8247c61
to WriteText() to DoSetValue(), thus fixing the same problem in
SetValue() and ChangeValue() too.
But the question of why do we have two so similar, yet different,
functions, resulting in having to fix the same bugs twice still remains
unanswered...
Remove the already described sizers from section Other Types of Sizers. Remove "See ..." sentences which are unnecessary when the class documentation is already linked right above. Also fix a typo in a sizer name that prevented doxygen linking to the class.
Remove references to concrete GUI-building programs. Such references are not necessary not to mention that most of the referred programs have not been maintained for a long time.
List of flags in the CreateButtonSizer sizer was unnecessary, some of the values were not valid flags for the methods and comment descriptions for most values did not make much sense in the given context. The section was therefore removed and the method is mentioned in the newly-added part about wxStdDialogButtonSizer.
Sizer types not mentioned in the overview before were added, the rest of changes was mostly adding @c and such.
Passing long argument to "%c" printf format specifier was correctly
flagged as invalid in 64 bit Unix builds where long != int.
Fix this by just making the "keycode" variable int in the first place,
there doesn't seem to be any reason whatsoever for it to be long and
this allows us to get rid of a couple of existing casts instead of
adding yet another one.
Fix compilation/link problems due to symbols redefinitions in user code
that could happen if it included some wx header first, then <windows.h>
(possibly indirectly, e.g. via another third party library) and then
another wx header -- in this case, the second wx header wasn't protected
from <windows.h> redefinitions resulting in all the usual problems.
Avoid this by always including winundef.h whenever any wx header is
included, not just when wx/defs.h is included for the first time.
This message now appears even in the (default) release builds as __WXDEBUG__
is always defined, so it needs to be at least somewhat understandable by
normal users, even if it remains primarily targeted at the developers.
Also remove __WXDEBUG__ checks in this function, this is a left-over from the
pre-3.0 debug mode.
Get rid of ugly wxDynamicCastThis()s in wxMSW wxControl code and add a
new virtual method overridden in wx{Tree,List}Ctrl instead.
Also stop comparing the font with wxSYS_DEFAULT_GUI_FONT, there doesn't
seem to be any good reason to set this particular font for these
controls neither.
In addition to simplifying and de-ugligying the code, this commit
incidentally fixes -Wnonnull-compare warnings in this code from gcc6
too.
This doesn't seem to make any sense as this function is always called
with non-null "*this" reference from wxPixelData dtor and results in
-Waddress warning from gcc 6.
In wxStdDialogButtonSizer, keyboard navigation order through the buttons
should be the same as the order they are positioned in the sizer regardless
of the order the were created or added to the sizer.
Closes#17940.
Under MSW calling UnblockAndRegisterWithEventLoop() for blocking sockets
is not only useless, but actually harmful when the socket is used from a
worker thread (which is the common case for blocking sockets), as it
means that the main thread will be receiving notifications for the
socket events and modifying the socket object while it's being used from
the other thread, resulting in data races and general brokenness.
This is similar to e18c8fd29a for Unix
sockets.
Closes#17937.
wxSocketManager::Install_Callback() and Uninstall_Callback() are only
called for successfully created sockets, so there should be no need for
these checks and there are none in the Unix version.
This doesn't work, as non-blocking sockets implementation requires
dispatching events generated by them, which is only possible from the
main thread event loop, and it's better to be upfront about it rather
than failing mysteriously later.
This is similar to the previous commit, but for active FTP connections.
wxSocketServer was also created directly in this case, without using
wxProtocol ctor, so wxSOCKET_BLOCK must be explicitly specified when
creating it from worker thread, just as it was already done in
d421373c2e for the other connections and
in the previous commit for passive FTP ones.
See #17937.
This extends the changes of d421373c2e
to the case of passive FTP, which created wxSocketClient directly and so
didn't use the correct flags when used from a worker thread.
See #17937.
Building samples one by one has become a significant pessimization
nowadays, when multicore machines are much more common than single core
ones, so get rid of the shell loop which built the samples one by one
and use make targets to let make build as many targets at once as
desired.
Closes https://github.com/wxWidgets/wxWidgets/pull/536
When WM_ENDSESSION was received by the application, the wxApp object
itself was shut down by calling OnExit() on it before all the TLWs were
destroyed, which could be completely unexpected as during normal
shutdown the order of events is exactly the reverse.
In practice, this resulted in crashes in any application whose main
window close event handler or dtor touched wxTheApp in any way (e.g. to
save any configuration in the global wxConfig object destroyed by
wxApp::OnExit()).
See #9590 (sorry for missing the point before, ATS).
