Get rid of ms_isInitialized as it must be kept synchronized with
ms_loaderDll.IsLoaded() anyhow, and it's simpler to not have it at all
rather than ensuring this.
Also ensure that calling Initialize() again, after doing it first
unsuccessfully, doesn't assert because ms_loaderDll is already loaded,
by only leaving it with a valid handle if the initialization succeeded.
Closes#19041.
Under MSW, don't set the state to State_Cancelled as soon as Cancel()
was called, as the request was still used from the other threads
afterwards, resulting in race conditions and crashes.
Fix this by just removing the SetState(State_Cancelled) call from the
main thread, as it was redundant anyhow. This also makes the behaviour
correspond to the documentation, which indicates that Cancel() works
asynchronously.
Also ensure, for all backends, that we actually cancel the request only
once, even if public Cancel() is called multiple times. This required
renaming the existing wxWebRequestImpl::Cancel() to DoCancel().
No real changes, just use the same name as in the other backends for
consistency (we could also rename m_sessionImpl in the other ones to
m_sessionCURL and m_sessionURLSession respectively, but this would have
been more work and the latter name is really not great).
Rename Init() to Open() as we need this method to return bool to
indicate its success in order to avoid using non-initialized handle
later. Init() is also reserved, by convention, for the common part of
all class ctors in wx code.
Remove m_initialized entirely, it doesn't seem to be obviously better to
cache the failure to create a session than to retry doing it every time
(in fact, it would seem to be worse) and not having it is simpler.
This commit is best viewed ignoring white space.
Don't force the application code to deal with wxObjectDataPtr<> or,
worse, calling {Inc,Dec}Ref() manually by hiding it inside the wx
objects themselves and giving the value-like semantics to them.
There should be no real changes in the behaviour, but the API does
change significantly. Notably, wxWebRequest is not a wxEvtHandler itself
any longer, as this would be incompatible with the value semantics, and
an event handler needs to be specified when creating it, so that it
could be notified about the request state changes.
Having wxWebSessionFactory part of the public API implies keeping
compatibility with the possible ways of implementing it which is too
restrictive for no good reason, so move this class to the private header
and don't document it nor wxWebSession::RegisterFactory() (which is now
private).
Check if edge webview runtime is available in the call to
wxWebView::IsBackendAvailable() instead of only doing it during
process initialization. This allows an application to install
the Edge WebView2 Runtime during runtime and use the edge webview
afterwards without restarting the process.
Setting EnableContextMenu and EnableAccessToDevTools where
ignored before the webview loaded content. This would often
be set directly after creation and work like this with other
backends.
Fix the problem of the access-denied error when using a 'new
window.ActiveXObject("Microsoft.XMLHTTP")' to access URIs from custom
WebView scheme handler of the Internet Explorer WebView.
The solution here is to implement PARSE_SECURITY_URL and
PARSE_SECURITY_DOMAIN in ParseUrl of the IInternetProtocolInfo interface
of the Internet Explorer Control.
The direct implementation of VirtualProtocol::QueryInterface() instead
of using IMPLEMENT_IUNKNOWN_METHODS works, but is not nice. The better
approach would be probably enhancing ADD_RAW_IID to support multiple
inheritance.
Closes#17893.
Apply the utility from https://github.com/codespell-project/codespell/
to fix spelling issues in the headers under both include and interface
directories and add a file with a couple of exceptions.
The exact command line used was:
$ codespell -w -I misc/scripts/codespell.ignore -i 3 in*
This results in a single intelligible error instead of a hundred of less
clear ones in case our code is compiled against an out of date SDK
version.
There doesn't seem to be any way to check the version directly, so just
check that the interface which hadn't been defined in the previous
version is defined now to check for it indirectly.
Creating wxPaintDC for a window outside of any wxEVT_PAINT handler
already resulted in assert failures and crash due to using the empty
wxDidCreatePaintDC stack, but the assert message was not really clear,
so improve it by stating explicitly that wxPaintDC can only be created
from wxEVT_PAINT handlers.
Also check that wxPaintDC is being created for the correct window: this
wasn't detected at all before, but could still result in a lot of grief,
so check for this too.
Finally, create a new private header with the paint data stack variable
declaration instead of using "extern" to declare it manually in wxDC
code.
Sockets returned by wxSocket::Accept() are non-blocking by default and
the only way to use them safely in worker threads is by switching them
to the blocking mode by calling SetFlags(wxSOCKET_BLOCK).
However this didn't work correctly since at least 2.8 days, as turning
wxSOCKET_BLOCK on didn't unregister the socket from the event loop, with
which it had been registered on creation. Fix this by doing this now,
which ensures that the main thread doesn't get any notifications about
the socket if it's used, in a blocking way, in a worker thread.
Note that making the new socket blocking after accpeting is still pretty
inefficient and pre-creating the socket as blocking and using
AcceptWith() is still preferable, but at least it does work now.
Closes#12886.
In addition to unblocking and registering the socket, also support using
this function to make the socket blocking and unregistering it from the
event loop, if its flags include wxSOCKET_BLOCK.
This was already half-done by wxMSW, which took wxSOCKET_BLOCK presence
into account in its implementation, but not by the Unix implementation.
Now do it under all platforms, as this will be useful for switching a
previously non-blocking socket to blocking mode.
Finally, rename the function to better reflect what it really does.
See #12886.
Some native dialogs do not scale correctly (color picker, font picker,
open file with custom controls). ALl other native dialogs do scale correctly
(open file, open directory, find replace, print).
Change the DPI Awareness Context temporarily to SystemAware, so Windows handles
the scaling.
These metrics can change when the DPI of a Window changes, so we can not keep a
static reference.
According to documentation, the second parameter (uiParam) should be set to the
size of the NONCLIENTMETRICS object.
If accepting a socket connection failed, wxSocketImpl::Accept() used
close() to close the socket even under MSW, but it can be only used for
the file descriptors there and closesocket() must be used instead for
the sockets.
Add new (private) wxCloseSocket define and use it both here, to fix the
bug, and elsewhere to make the code more clear.
Closes#18407.
Since the changes of 6ae7aa4443, the
windows were shown when their geometry was restored as a side effect of
calling ::SetWindowPlacement(). This was unexpected and resulted in
flicker on startup, so fix this by explicitly passing SW_HIDE to
SetWindowPlacement() if the window is currently hidden and storing the
real show command inside wxTLW itself, where it will be used when it's
finally shown.
Save both the normal window geometry and its maximized position instead
of saving just its current position. This fixes restoring geometry of
the maximized windows as previously they were always restored on the
primary monitor, as their original position was lost.
Use the native {Get,Set}WindowPlacement() functions for a MSW-specific
wxTLWGeometry implementation to achieve this.
Closes#16335.
Use __ms_u_long instead of just u_long with Cygwin to avoid mismatch
between (64 bit) Cygwin long and (still 32 bit, even in 64 bit build)
Windows API long.
Add wxMSWWinStyleUpdater and wxMSWWinExStyleUpdater helper classes which
allow writing code changing GWL_STYLE and GWL_EXSTYLE bits,
respectively, in a shorter and more clear way.
There should be no real changes in behaviour.
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.
