Age | Commit message (Collapse) | Author |
|
Change-Id: Ie5f55feb39798dff27e0c6b41d5b88b6e9bf19bc
|
|
Change-Id: I4c2d0447526bda4ccbe2fa7edaf8be1886d214f0
|
|
Just remove the inner rectangle. Now it looks nice.
Change-Id: Icd3d3971292f958d1ffdd3dc6a7697005b5f9047
|
|
Change-Id: I89f1f5771eb7d96da233a8c7386b906b2a3a9c3e
|
|
Change-Id: I9906b4e58fd8672e51f6aa2e56ec68bb28772500
|
|
Also put a missing space before MAX
Change-Id: Ie2def48ec66d4de91bd16c39ed821c4028b6d1c6
|
|
Change-Id: Ib92adfc83c963ecfcfa7fdbc2ec71f363a181330
|
|
Change-Id: Ifd53e37d23c29efb2f084a39b48c88ea4bc0d70a
|
|
Change-Id: I77eaf1dad9ce4c855b5f88bd8568944fd4128227
|
|
Makes it possible to see the messages immediately on Windows; otherwise the
SAL_DEBUG() etc. is useless as it shows on screen too late.
Change-Id: I701922aa4fcacf66d6b044e0badca7ef1f9cac94
|
|
Apparently it can happen in JunitTest_sw_unoapi that between the
getAccessibleChildCount() and the getAccessibleChild() calls that the
children all disappear; try not to assert in that case.
Change-Id: Iec8e373281950bf684fa76d69f8c4625ecd1255a
|
|
This does not appear to work so well: for a row with multiple cells,
when the second cell frame's SwAccessible is disposed it will send an
event, and the AtkListener will then request all children of the row,
and this causes a new SwAccessible to be created for the first cell
frame that was disposed previously.
Unlike most other events, Dispose events cannot be sent delayed.
The SwAccessibles are disposed from ~SwCellFrm already, that should be
sufficient.
Change-Id: I6da72605e5386540f8d82535fca151e97d70869f
|
|
... with accessibility enabled: the SwCellFrms are deleted, but the
entry in the SwAccessibleMap::mpFrmMap is not removed, because Dispose
returns early since the connections to the nodes have been severed already.
This then crashes in some questionable code that iterates over and uses
the frame pointers in mpFrmMap that was added in
76c549eb01dcb7b5bf28a271ce00e386f3d388ba (where previously only the
WeakReferences were used out of that map).
Change-Id: I13885c6ab9a92e34e95925f275e15c20fdc600b7
|
|
... before firing events from them; use a thread-safe WeakReference.
(regression from 76c549eb01dcb7b5bf28a271ce00e386f3d388ba)
Change-Id: Ie060d27cc44415e9a75b75027f5510577ac17a6e
|
|
Change-Id: Ibca921eff3647507ffb7e9011a84aeefbec58d7a
|
|
maItemListRect serves as a sort of bounding box for what items /could/
be visible on screen, however it is only used for determining the id
of an item we have just clicked on -- given that we already know that
we are clicking inside the visible area this is an unnecessary check.
In fact this check is erronous as we no longer limit ourselves to "whole"
rows as of 806919adf9c9bafbaba92c2d2ab35d2e2f9863f8 -- i.e. the previous
assumption was that a view large enough for n rows will show precisely
n (whole) rows, whereas we might be showing (n-1) complete rows, and a
fraction of a row at the top and bottom of the screen, i.e. a total of
n+1 rows, where maItemRect only encloses the top n rows, meaning we
were erronously ignoring clicks on the (partly visible) lowest n+1 row.
Change-Id: Ia52ed2e0d217a9f077cef86ee9c1847252844995
|
|
Change-Id: I96e84e3ec756216fabe9739e6bd53d2a89d186b5
|
|
Change-Id: Ifd89b00c525ed4d75fcffe92845c4ca2f73705a8
|
|
Don't be fooled even when the document doesn't have a drawing layer
initialized. Sometimes a note creates caption on demand later, but if
the drawing layer isn't there the caption will not get created, which
ultimately causes this crash.
Change-Id: I37f4902fa84de91c9f793dc352127d9345a725e3
|
|
Change-Id: I7ac9792da93a1b5231ddac5398823667b28adf92
|
|
There is a good reason for that: File lists of
some file pickers handle double click event on
mouse down. So if a start center thumbnail is
exactly behind the file on which the user double-
clicked, it will receive the mouse up event, and
will also open.
Change-Id: I9fc9ceff0874d0486d5b568c9b35143c8d8aabf8
|
|
Change-Id: I3055df538fc4dfa625aa942ad0d0e6b6299c53b3
|
|
Change-Id: I6d7f6e035bb0dcc86f8aebfdfa4b7f9521bc8820
|
|
Change-Id: If22324ea65a45ac93f6c4b9cbaf9cfb11ef91195
|
|
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.9.sdk/System/Library/Frameworks/Carbon.framework/Versions/A/Frameworks/HIToolbox.framework/Versions/A/Headers/Events.h
says that 0x27 is kVK_ANSI_Quote, which we call KEY_QUOTERIGHT,
i.e. the plain ASCII apostrophe, while 0x32 is not the ASCII tilde but
kVK_ANSI_Grave, which we call KEY_QUOTELEFT.
Makes the Command-' binding (see Kohei's
56cce6fd8151a916e1be540ac18724b46d1b15fc) work at least for the US
English keyboard. But not for the Finnish keyboard, for instance,
where the key labeled with an apostrophe (or "ASCII-style" single
quote) generates a different keycode than the single quote key on the
US English keyboard. (Which, I guess, isn't surprising as it after all
is a different physical key.)
In other words, situation normal, using keycodes is a horror as
always...
Change-Id: Ibf8f994370fec75849391e1dba30c7b7d54d237f
|
|
Change-Id: I2acd44d54966dd227f5dc23ca4acde3c50ca8347
|
|
Change-Id: I9fd2980a32bd0a3eb4849a53f29cc10b9a371f4c
|
|
Change-Id: Id320bc41fefff17035c8a30115dc94af5cbf9503
|
|
this is like it was pre ui conversion
Change-Id: Ie0fe230893dc448531a44c6d79c00b4a29120cba
|
|
Change-Id: I3b6d7091922ed345d24c15c5a52d37521792d551
|
|
Change-Id: If872737d2a7db5d9053a256f7e3d57007c24230e
|
|
Change-Id: I9dfe76d422ad732a081442b95a995d8d395a098e
|
|
Change-Id: I8914e0dc253c636a1393672f7a20bcb5d39d4276
|
|
Change-Id: I8e163e63bfbbbaa5f7f4de2151fea41d735417d0
|
|
Change-Id: If417d8091896c9f6c9650cf7d6036ffacbcc0884
|
|
Change-Id: I6510f486707b08ea8f611105f40cc4c7e1581a7f
|
|
Change-Id: I9c2aecd5dea617e7cd69c1fc005b6a9ee237f202
|
|
Change-Id: Id6bac3a8dd46d7994c9e64c9e0c16e8c7ac036ac
|
|
Change-Id: Ibfed531e48848ea711be429d6e5c08094473f781
|
|
Change-Id: I91220e8ebffa0090fa7e9146df853f16b6783491
|
|
Change-Id: If0454ace5ef21b9d32299f3edea115a2726aac00
|
|
Change-Id: I651c0e5830008fd08ba886ccf1be99d40868bb72
|
|
Change-Id: Ic117de10be2181a80e65ea48c888789456abfb0a
|
|
Change-Id: I7a3680aa324c4242572b71c91ec5bc4f85eff9e9
|
|
Change-Id: I7b1ca054006500d468da73d2d85eafaad85dda67
|
|
Change-Id: Ia2e09c92673b97fd48c4ede7ebb9ff47da821926
|
|
Change-Id: If1e0e78e1b62c547bd0f03e727e1856227b32821
|
|
Change-Id: I7a2acd99f3a8cb143c2ad1c978dfa4b02bf54464
Signed-off-by: Stephan Bergmann <sbergman@redhat.com>
|
|
... during DOCX import of shapes with text, so that we don't crash when
the currently always active
setServiceName("com.sun.star.text.TextFrame") call is not present.
Change-Id: I5b005583ddcee81b9683e3b34e8f0a2e5faa4f95
|
|
Fix build error on windows with --enable-dbgutil autogen.sh option.
We need to link freetype248_D.lib,when we use --enable-dbgutil option.
(Suffix "_D" is needed.)
Change-Id: Ib2927e777d9b9bb615e1df62e428cf17cc8bc26e
Reviewed-on: https://gerrit.libreoffice.org/9248
Tested-by: Michael Stahl <mstahl@redhat.com>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
|