The size is expressed in logical coordinates and needs to use bitmap's
scaled size, otherwise it would appear twice as large in both
dimensions.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@75350 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
Generalize the code in wxTopLevelWindowBase dtor checking for the children of
the TLW being destroyed pending for deletion themselves to work when the child
TLW is an indirect child, i.e. was created with a child window of this TLW as
parent and not this TLW itself.
Closes#15743.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@75342 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
IsDirty() still returned true even after undoing all the commands which was
wrong, as there was nothing to save in this case.
Closes#15722.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@75310 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
wxMenu::Remove() was still broken in wxMSW after r75250, even with the fix in
r75290, as wxMSW code relied on the item still being present in
wxMenu::m_items.
Delay removing it from there until after DoRemove() call to fix this.
See #3424.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@75297 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
The menu items were not removed from the menu any longer when they were
deleted or destroyed after the changes in this revision.
Fix this by calling the public Remove(), which does the right thing, instead
of the private DoRemove(), which only does part of the job of removing the
item, in DoDelete() and DoDestroy().
See #3424.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@75290 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
The assert complained about a bug in wxMenu code itself, but actually could be
triggered by simply trying to remove a non-existent item from the menu.
Fix this by checking for the item existence, and also removing it from the
internal menu data structure, in wxMenuBase::Remove() and only dealing with
the item itself in DoRemove().
Closes#3424.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@75250 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
This was broken because Select() never returned wxSOCKET_CONNECTION_FLAG which
is supposed to be set when a connection is accepted.
Closes#15669.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@75205 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
We detected that the label was invalid and gave a debug warning message about
it but then still proceeded to crash by accessing the data beyond the end of
the string. Don't do this.
Closes#15665.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@75180 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
This ensures that the files created by wxTextCtrl::SaveFile() can be read back
by wxTextCtrl::LoadFile() as previously the files were saved using the current
locale encoding but read back using Latin1 (after first trying, and failing,
to read them as UTF-8).
This is a backwards incompatible change but it ensures that wxTextCtrl can
load its own files and is also consistent with the use of UTF-8 by default in
other places. Finally, and perhaps most importantly, this ensures that the
file contents can always be saved, i.e. there is no risk of conversion errors
any more.
Closes#15611.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@75073 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
Don't provide our own (not fully functional) definitions of them and always
use the system versions as we don't support OS X 10.2 which was the last
platform where these functions didn't exist/work.
See #15580.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@75024 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
No real changes, just use a self-explanatory "lastWasCR" variable instead of
storing the last character read which wasn't used for anything except
comparing it with '\r'.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@75010 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
This undoes r74888, r74913 and r74969 because all of them were still not
enough to make ICC build work correctly. So instead of trying to find all the
places where we need to test for __INTELC__ in addition to __VISUALC__, just
revert to defining __VISUALC__ for ICC under Windows as it uses its CRT anyhow
and in the vast majority of cases we are actually testing for the CRT and not
the compiler anyhow.
Closes#15359.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74999 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
Using wxFD_FILE_MUST_EXIST in combination with an extra panel in wxFileDialog wasn't working properly. Simply remove the unused flag under OS X. Caused problems with Cocoa only (not Carbon), and at least under 10.6 and 10.8.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74971 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
While "!a & b" is the same as "!(a & b)" when "b" is a single bit, the latter
is much more standard and more clear, so write this test (added in r71704)
like this.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74952 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
This change slightly improves wxBufferedDC and wxBufferedPaintDC behaviour
when the origin of the DC is shifted, but they still don't work quite right in
this case as they don't use the buffer area of correct size in this case,
which results in cropping the final bitmap (as can be seen in the drawing
sample, which was modified to show wxBufferedPaintDC in action).
Closes#15497.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74951 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
This allows the program to distinguish between some other, unspecified,
warnings and this one which can and does happen whenever too many changes
occur too quickly but which has a clearly defined work around: the state kept
inside the program just needs to be refreshed by rescanning the directory anew.
See #12847.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74950 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
They were compared inconsistently with the numbers and strings, -1 is supposed
to be returned if the first element is less than the second one, not 1.
Closes#15406.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74949 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
Explicitly use wxConvLibc with wxTextInputStream to make sure we correctly
decode non-ASCII data in the subprocess output.
This is a hack, the real solution would be to make wxTextInputStream work
properly with wxConvAuto.
See #14720.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74946 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
We can't rely on our own globals being already constructed if we're called
during another global initialization, so use the usual trick with wrapping
them in accessor functions to ensure that they are.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74935 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
The code trying to avoid forwarding duplicate events to wxDocManager was over
eager and in some situations filtered out the events which hadn't been sent to
it yet and were, in fact, not handled at all. This could be seen, for example,
by running the docview sample with "--sdi" command line option, creating one
child frame and then trying to create another one from the parent frame menu:
this failed because the existence of a valid child was considered to be enough
for the event to have been already processed in it which was false in this case.
Unfortunately there is no obvious fix to this problem, notably because of the
very roundabout way the toolbar events are processed in MDI windows: the
toolbar itself is a child of the parent frame but the events from it are still
sent to the currently active child frame by wxMDIParentFrameBase. So we can't
rely on any kind of parent-of-originating-window checks.
Instead, remember the last event handled in the child and avoid processing the
same event in wxDocManager again. This should at least avoid the false
positives (like the one fixed by this commit), although it could still result
in false negatives (i.e. some duplicated events) if an event handler generated
other events while skipping the original one. This is a lesser evil though and
should be relatively rare in practice, so live with this ugliness until
someone comes with another idea of fixing the bug described above.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74919 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
This is confusing and makes it more difficult to test for the "real" MSVC,
test for __INTELC__ explicitly wherever needed instead.
Also document __INTELC__ in our list of compilers.
Closes#15359.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74888 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
Doing so has negligible cost and can be useful in some situations (e.g.
when using symbolic msgids despite gettext best practices).
Patch by tmsorensen.
Fixes#15522.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74849 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
Replace GetString(), which always returns something (possibly the
original string) with GetTranslatedString() that returns either a
pointer to translated string or NULL.
This simplifies the code a bit, all handling of missing translations is
now done in wxGetTranslation().
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74836 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
The GetUntranslatedString() hack keeps a global copy of all strings, so
that it can return a const reference as wxGetTranslation() return value.
A global wxHashSet instance shared by all threads won't do, even guarded
with a critical section, because it may internally copy values on any
insert and thus invalidate pointers that may still be used on another
thread.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74833 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
This helper method falls back on the alignment of the column if the renderer
alignment is not specified. This is almost always what should be used instead
of GetAlignment() to determine the alignment that really should be used in the
drawing code.
In particular, using GetEffectiveAlignment() in wxDataViewCustomRenderer fixes
the problem with bitmap columns ignoring column alignment for their bitmaps.
Closes#15498.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74825 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
wxGridBagSizer lay out algorithm needs at least a single row and a single
column to work, so simply don't run it at all if there is nothing to lay out.
Closes#15475.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74810 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
This makes it behaviour consistent with wxFile::Open().
Also don't use Detach() in Close(), again for consistency with wxFile, even if
this has no user-visible effects at all.
See #15494.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74799 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
Always move the dates invalid due to DST (i.e. falling into the "missing" hour
on the DST start date) forward, as GNU libc does, even when using a different
CRT implementation, such as MSVC one which moves the invalid dates backwards.
This seems more expected and also fixes an especially bad problem which
happened due to moving the date backwards in Brazilian time zone where DST
starts at midnight as doing this changed the day and totally broke ParseDate()
assumption that setting wxDateTime to 00:00:00 at the given date really did
set it to this date.
Closes#15419.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74777 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
Finding a button in a child dialog when looking for it in the current window
is totally unexpected and can result in subtle bugs, see #15442.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74721 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
We need to do this when using any port under OS X, not just wxOSX, e.g. the
-psn_xxx arguments passed by the Finder when launching the application, must
also be ignored when using wxGTK.
Closes#15432.
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@74703 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
