Age | Commit message (Collapse) | Author |
|
The scenario of fdo#63421 (loading data and re-dragging the same field)
is not needed, simple data is sufficient and crash happened also when
dragging (removing) a field from a pane and dropping it anywhere else.
Multiple errors:
* getAccessibleChildCount() must return the real current count of
children, not what mpFieldWindow says; AtkListener::updateChildList()
uses this value to repopulate its own list; a child is added after it
is added to mpFieldWindow but removed before it is removed from
mpFieldWindow;
* LostFocus() uses an index of -1 if the last child was already removed
and the field was dropped after dragging it away from a pane, handle
that but it still does not look right
* RemoveField() called CommitChange() with
AccessibleEventObject::NewValue set instead of OldValue, leading to
AtkListener::handleChildAdded() being called instead of
handleChildRemoved()
Apparently this never worked since 2002.
Change-Id: Idfb59d947002d2212bc67b414daecb65c55edae8
|
|
by handling created ScPrintRageSaver objects more careful.
Found by: Chao Huang
Patch by: Chao Huang, alg
Review by: alg(cherry picked from commit 803b94a2391352ccef46a6f8fdcfa9dc748bb737)
Change-Id: I0523bb870ed6a4ee301722e6e80dc9ac497a7299
|
|
Patch-by: Tan Li <litan.test@gmail.com>
Review-by: hdu@apache.org
(cherry picked from commit 3258f70090ffb706d7dcfb66b16b404d86f8cfbf)
Conflicts:
sc/source/filter/excel/xistyle.cxx
Change-Id: Id50781e163ac33d3cd1b670808824c7c9b8718d3
|
|
Change-Id: I1f53a4f4b9beee501e29149ed88e3ce5d63d9a3c
|
|
Change-Id: Ia5a73ec2acd3663a35632f268b59c0646cd356ea
|
|
Change-Id: Ie8dad05f8629f8acea735c6ff71851d22b997a37
|
|
Change-Id: Ib7f28a27ea99e76f439673e4971706ed9bb7a488
|
|
Change-Id: I74e2b8299455753d4663f344506912302317b681
|
|
Change-Id: I165c75693c4d4acb331e93905392fa7b6f7bc1c0
|
|
Change-Id: Ifa50e460768ffff4b1096d795ab694fea62a6c29
|
|
It was about time.
Change-Id: I87225be540c22b3031c6aee3f3fd9d26795abef8
|
|
Change-Id: Icaf310182dcdd2a5d06c2d4bcbb98e3850fb2d80
|
|
It'll simply call set_auto() for now.
Change-Id: I580b2ecbd34ffe7674f64ade912877ef07a60aca
|
|
With this, string cells from ODS docs get imported. For now, format runs
are ignored.
Change-Id: Ia9530357d3afc6cf4a0c52362bc770560debfd61
|
|
Change-Id: I9c1d0e904464ff2489678972aa5c6b6e4246cd81
|
|
This is because orcus imports cells before it imports the string table.
Change-Id: I1b85be2d9832b9a9a52961ece735de8cd980b893
|
|
Also, let's not leak memory by returning a new'ed instance from get_styles().
Orcus never manages the life cycle of objects returned from client code.
Change-Id: Id92eb10dbe88c2a55b5d04e408a01fa9c25fa9a1
|
|
But this occasionally crashes.
Change-Id: I0a18d0e210639b43b89b966a54de541b9a43e329
|
|
This fixes the symptom of the crash but not the underlying cause why a
subtotal count would be wrong.
Change-Id: I3782b5e39f18bc65ffe510b847ffa7969a26cd37
|
|
The document https://issues.apache.org/ooo/attachment.cgi?id=34418
attached to https://issues.apache.org/ooo/show_bug.cgi?id=62493 crashes
in a dbgutil build due to an invalid subtotal count that leads to an out
of bounds access in a vector, in this case with offset -1
This fixes the symptom, not the cause why the subtotal count would be
wrong.
Change-Id: If96c2eadba8753ea3e3166db0a81441818c89b2d
|
|
Change-Id: I7480e8d606e1eafde45635e36dfafa4232f529cb
Reviewed-on: https://gerrit.libreoffice.org/3217
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: Ibf587e1ea94765f0322a5b847bcf82d1eabc1b64
|
|
Change-Id: I93ac358e259a1fb310dfeed75bc60910405e7d48
Reviewed-on: https://gerrit.libreoffice.org/3306
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
to avoid a memory leak.
Found by: Chao Huang
Patch by: Chao Huang, alg
Review by: Chao Huang, alg
(cherry picked from commit e7168710bfd30b5d682d21d81c116cd33cf22505)
Conflicts:
sc/source/filter/xml/xmlfonte.cxx
Change-Id: I75475bb59b3a7040938826d095dcd81e19d61b46
|
|
Change-Id: I0542d9918d9dbbfdc002b7fadec1bb461cbd2831
Reviewed-on: https://gerrit.libreoffice.org/3291
Reviewed-by: David Tardon <dtardon@redhat.com>
Tested-by: David Tardon <dtardon@redhat.com>
|
|
Change-Id: Ie78b7ce399ba34485146ca7622c59d31f8105d02
Reviewed-on: https://gerrit.libreoffice.org/3229
Reviewed-by: David Tardon <dtardon@redhat.com>
Tested-by: David Tardon <dtardon@redhat.com>
|
|
I plan to use it to deliver all */uiconfig files, not just .ui, as a
preliminary step to get rid of postprocess/packconfig.
Change-Id: Ie7b4434b2f247165e3ab69a4d0c193418720a149
Reviewed-on: https://gerrit.libreoffice.org/3225
Reviewed-by: David Tardon <dtardon@redhat.com>
Tested-by: David Tardon <dtardon@redhat.com>
|
|
Change-Id: Icb0000bde3723c1b37713d0f26ef8305c4a199b8
|
|
also checked and fixed memory leaks caused by not deleting
SfxStyleSheetIterator instances.
Found by: Chao Huang
Patch by: Chao Huang
Review by: alg
(cherry picked from commit aee8221b3fbe64d8e88aa3a1b630214807cd86f0)
Conflicts:
sc/source/core/data/stlsheet.cxx
sc/source/filter/xml/xmlfonte.cxx
sfx2/source/doc/objcont.cxx
svl/source/items/style.cxx
sw/source/core/unocore/unostyle.cxx
Resolves: #120077# ALG: Reverted r1354011,
adapted all usages of SfxStyleSheetIterator and CreateIterator to use
boost::shared_ptr
(cherry picked from commit ec1819f218b94bed1556653e1b2b816a65da61ef)
Conflicts:
sc/source/core/data/stlsheet.cxx
sc/source/filter/xml/xmlfonte.cxx
sfx2/source/doc/objcont.cxx
svl/inc/svl/style.hxx
svl/source/items/style.cxx
sw/source/core/unocore/unoframe.cxx
sw/source/core/unocore/unostyle.cxx
26937d4d65562b14de0aea59a7b01cc4334f73c1
Change-Id: If99bf3ee399292ec269d68b36ee78ed63f033215
|
|
Joren has improved this, but the behavior was unexpected when the user did not
start in a cell that contained the string to replace. This commits changes
it further that when the user hits Replace:
* in a cell that does not contain the string to replace, it jumps to the cell
that contains it, replaces the string, and stays there
* in a cell that contains only one occurrence of the string to replace, it
replaces it, and jumps to the next cell that contains such a string
* in a cell that contains multiple occurrences, it replaces it, but does not
move the cursor (so that the further occurences in that cell can still be
replaced)
This seems to be the most intuitive behavior.
Change-Id: If6c10069b8dff933a035780732a7d7b6e0740383
|
|
Change-Id: I4e2567ba2a9000763c54b7d38ab4ba66a13a77f4
|
|
Change-Id: Iaa48ebe1799960532483812e54aa08c625882388
|
|
Without these linking fails with unresolved auto_ptr<
formula::FormulaTokenArray> symbols which is clearly a compiler bug.
Change-Id: I14faf8e318522f913370ac186b5114ae5db997ad
|
|
dialog is visually overly narrow now so reorient it to the normal button
orientation and, keeping "delete all" prominent, put the rest of the options
into two columns
Change-Id: I6a8692df612d9f97d45621a0bcc9d9a6e75a4312
|
|
Change-Id: Ia1735ae39618fb8b0e4749e1970adfe83192f755
Reviewed-on: https://gerrit.libreoffice.org/3197
Reviewed-by: Jan Holesovsky <kendy@suse.cz>
Tested-by: Jan Holesovsky <kendy@suse.cz>
|
|
Change-Id: I193f1cad5938550c576909802c945448ee41d17e
|
|
Change-Id: I8edc9210bc172cd285fe4dce54d30606ecc7c716
|
|
Change-Id: I1c84fe3c57fdd72880b16a2d69aa44cda0fc4f4f
|
|
Change-Id: I892d2856c28536a7828c292c8fb3cd8b30fe8ecb
|
|
Change-Id: I9b55140e619a16a22dd10ff67abfe27fbd58bc05
|
|
Change-Id: I67d587c0241a1d8377aee199a30bfcc919e3faad
|
|
Change-Id: I2d9baed5032e267ac5053267d367723999526c8f
|
|
Change-Id: Ifa5ded07e3af153e747fdbd82a75560dfb7e2b75
|
|
Change-Id: I8f07e1a37f32222aa25721704e513ef1aa3092e3
Reviewed-on: https://gerrit.libreoffice.org/3281
Tested-by: Caolán McNamara <caolanm@redhat.com>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: I5f6a159b3c925bbd9b3a5820968fff8f95a70925
|
|
Change-Id: I62122e298f1434aa9e44c8f3fdfb7f710513617d
|
|
With this, we get file system abstraction for free.
Change-Id: Iec988712599f506444cf9c7e3ebbd59d38f083ee
|
|
Change-Id: Ife3c14a446273654d720eb38349cc42d1ed76893
|
|
The reason there appears to be two dialogs, is because there
actually is. The original dialog doesn't get deleted under
windows.
If we use the actual returned type of the dialog here rather than a lower
baseclass then the right thing happens. Original code works under Linux but not
Windows.
Change-Id: Iaf99fa4e16f743b1f5a3d29f42a140491f3d67b0
|
|
Change-Id: Ib53b42a340da04276ce005efb22d19fda617e29a
|