- 03 Nis, 2019 6 kayıt (commit)
-
-
Michael Stahl yazdı
Fixes CVE-2019-9636 CVE-2019-5010 CVE-2018-14647 Change-Id: If0a115960aed1ee90b63e6716c844669f0ec91e5 Reviewed-on: https://gerrit.libreoffice.org/70182 Tested-by: Jenkins Reviewed-by:
Michael Stahl <Michael.Stahl@cib.de> (cherry picked from commit 3c7354a8) Reviewed-on: https://gerrit.libreoffice.org/70192Reviewed-by:
Thorsten Behrens <Thorsten.Behrens@CIB.de>
-
László Németh yazdı
Now correction of the ASCII quotation mark to the typographical one (’) capitalizes the preceding single lowercase i, too, as in MSO, if the language of the text is English. Change-Id: Ic7586f07aa13f441e25494ff4cc11c672ac4a67a Reviewed-on: https://gerrit.libreoffice.org/70088 Tested-by: Jenkins Reviewed-by:
László Németh <nemeth@numbertext.org> (cherry picked from commit 376e3a48) Reviewed-on: https://gerrit.libreoffice.org/70190Reviewed-by:
Adolfo Jayme Barrientos <fitojb@ubuntu.com>
-
Dennis Francis yazdı
is not enough. This needs to be done recursively, ie notify the listeners, notify the listeners of the listeners and so on. ScDocument::BroadcastCells() seems to do exactly that, so lets use it here intead of collect+notify. Change-Id: Ifa273fea5f08af661958cd9b4c3d01d9044b8727 Reviewed-on: https://gerrit.libreoffice.org/70044Reviewed-by:
Dennis Francis <dennis.francis@collabora.com> Tested-by:
Eike Rathke <erack@redhat.com>
-
Takeshi Abe yazdı
starting from 2019-05-01, which has been announced officially. This fills the provisional slot acknowledged at cacbb0fa. Change-Id: Ifb12e6afaad4c66d455f664b46ec946e80324e87 Reviewed-on: https://gerrit.libreoffice.org/70157Reviewed-by:
-
Jaromir Wysoglad yazdı
The ANOVA sum of squares is now hardcoded to use '$Sheet?' in the source cell addless every time. Change-Id: I215b719be11ecfc2ae20c5df8e86876bc22d50f0 Reviewed-on: https://gerrit.libreoffice.org/69899 Tested-by: Jenkins Reviewed-by:
-
Caolán McNamara yazdı
rather than the meaning of nothing-selected which is -1 Change-Id: I3be01fa8d6b57f61aa660be5912f479e79cff2b2 Reviewed-on: https://gerrit.libreoffice.org/70143 Tested-by: Jenkins Reviewed-by:
Xisco Faulí <xiscofauli@libreoffice.org>
-
- 02 Nis, 2019 6 kayıt (commit)
-
-
Armin Le Grand yazdı
By error the ClipRegion's geometry was replaced by it's BoundRect expanded to PixelBounds. If the ClipRegion is not a rectangle, this will create wrong results. To do both - extend to PixelBounds and have the original geometry included - use the PolyPolygon topology as needed (see comment in code for details) Change-Id: If3f75223144eba8eb23909a7c701ad544346099b Reviewed-on: https://gerrit.libreoffice.org/70146 Tested-by: Jenkins Reviewed-by:
Armin Le Grand <Armin.Le.Grand@me.com> (cherry picked from commit 362c1cf2) Reviewed-on: https://gerrit.libreoffice.org/70154Reviewed-by:
-
Michael Stahl yazdı
There is a surprising recursion with 2 SwTextSlot on the stack, with 2 consecutive fields, and then the m_pOldGrammarCheckList is not the SwTextNode's but a SubList of that, so the assert triggers spuriously. SwExpandPortion::Paint(SwTextPaintInfo const&) const (this=0x7d61cc0, rInf=...) at sw/source/core/text/porexp.cxx:66 SwFieldPortion::Paint(SwTextPaintInfo const&) const (this=0x7d61cc0, rInf=...) at sw/source/core/text/porfld.cxx:437 SwLinePortion::PrePaint(SwTextPaintInfo const&, SwLinePortion const*) const (this=0x7d61cc0, rInf=..., pLast=0x6fa49d0) at sw/source/core/text/porlin.cxx:138 SwExpandPortion::Paint(SwTextPaintInfo const&) const (this=0x6fa49d0, rInf=...) at sw/source/core/text/porexp.cxx:78 Change-Id: I915f31ce6de57b6a4a051598428441ae6c86d264 Reviewed-on: https://gerrit.libreoffice.org/70131 Tested-by: Jenkins Reviewed-by:
Caolán McNamara <caolanm@redhat.com> Tested-by:
-
Noel Grandin yazdı
on my machine, this takes the time from 12.3s to 8.9s This also includes ba748771 tdf#122926 FILEOPEN ODT Slower compared to the past (part2) Further improvement: on my machine, this takes the time from 8.9s to 5.7s Change-Id: Ibdc7e3c34d1acb0d1f7f784720552fa6c7977fa5 Reviewed-on: https://gerrit.libreoffice.org/70074Reviewed-by:
Noel Grandin <noel.grandin@collabora.co.uk> Tested-by:
-
Thomas Viehmann yazdı
Previously __file__ was set after executing the module. This meant, however that it was not available during execution. This patch moves setting __file__ before the execution. As __file__ is a string (and not bytes), we remove the encoding. Change-Id: I63fe34b6f5d30f53aab16627d413c87b9de81992 Reviewed-on: https://gerrit.libreoffice.org/70048 Tested-by: Jenkins Reviewed-by:
Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de> (cherry picked from commit dbbf20a4) Reviewed-on: https://gerrit.libreoffice.org/70119Reviewed-by:
-
Caolán McNamara yazdı
Change-Id: I4e801ec234e86be935a79c9e29f0e5091904f12e Reviewed-on: https://gerrit.libreoffice.org/70053 Tested-by: Jenkins Reviewed-by:
-
Caolán McNamara yazdı
and do the work that depends on size happen at size-changed instead of first-draw Change-Id: Ia2605ba9bb875be9248a009a6f26f01cbbbd4632 Reviewed-on: https://gerrit.libreoffice.org/69928 Tested-by: Jenkins Reviewed-by:
-
- 01 Nis, 2019 2 kayıt (commit)
-
-
Aleksei Nikiforov yazdı
Move deleteLater() call to the different event loop. This fixes issue with frames not being disposed of deterministically, and previously opened window not reactivating. Also hide modal dialog window before unsetting modal mode. Unsetting modal mode may require to toggle window visibility. Window will be hidden soon anyway, and additional generated events might make finding correct focused window harder. Change-Id: Id7839f817075785287b09f6ac79eb3fb211726aa Reviewed-on: https://gerrit.libreoffice.org/68852 Tested-by: Jenkins Reviewed-by:
Katarina Behrens <Katarina.Behrens@cib.de> Reviewed-on: https://gerrit.libreoffice.org/70078Reviewed-by:
-
Noel Grandin yazdı
regression from commit 3a9d3f27 Date: Mon Apr 23 09:38:41 2018 +0200 loplugin:useuniqueptr in TextDoc Change-Id: I51ab5de7571e4ec358442e54d590adf88fbeb12d Reviewed-on: https://gerrit.libreoffice.org/70061 Tested-by: Jenkins Reviewed-by:
-
- 30 Mar, 2019 3 kayıt (commit)
-
-
Jan-Marek Glogowski yazdı
LO does its own layouting and also all the scaling, so disable the automatic HiDPI scaling explicitly. Otherwise fonts will be rendered in LoDPI and just scaled as an image, if UI scaling is enabled in KDE, resulting in really blocky / ugly fonts. Change-Id: I38503ce27c1671e80d0749d21c6c6dcff1a808a6 Reviewed-on: https://gerrit.libreoffice.org/69941 Tested-by: Jenkins Reviewed-by:
Jan-Marek Glogowski <glogow@fbihome.de> (cherry picked from commit b4698a22) Reviewed-on: https://gerrit.libreoffice.org/69984Reviewed-by:
-
Caolán McNamara yazdı
Change-Id: I0fd8738d42ad33f029381dae0ec9f41e7912b72c Reviewed-on: https://gerrit.libreoffice.org/69918 Tested-by: Jenkins Reviewed-by:
-
Miklos Vajna yazdı
Visible at e.g. in Calc: Format Cells/Borders/Line Styles. The problem was that first commit 2e99e4e1 (opengl: use MVP matrix in vertex shaders, pixel offsets, 2015-07-08) introduced the concept of pixel offsets, setting the value (implicitly) to 0 in OpenGLSalGraphicsImpl::DrawTrapezoid(), but using 0.5 in OpenGLSalGraphicsImpl::FlushLinesOrTriangles(). This is fine, but then later commit 2003076c (opengl: batch draw polypolygons, 2016-05-29) changed OpenGLSalGraphicsImpl::drawPolyPolygon() to use deferred drawing instead of DrawTrapezoid(), without doing any translation of the input polypolygon. This resulted in loss of those polygons when used in the above mentioned dialog, which has a listbox of bitmaps, where each line style preview is drawn on a virtual device with a height of 1px. So at the end the 1px offset meant the previews were simply missing. ('make CppunitTest_vcl_gen SAL_USE_VCLPLUGIN=gen SAL_FORCEGL=1' is needed on Linux to see the test failing without the fix.) (cherry picked from commit 75e152a7) Change-Id: Ia9f3d6e7cb38a43fe2f8a41746b538af68add43c Reviewed-on: https://gerrit.libreoffice.org/69931 Tested-by: Jenkins Reviewed-by:
Tomaž Vajngerl <quikee@gmail.com>
-
- 29 Mar, 2019 8 kayıt (commit)
-
-
Serge Krot yazdı
Change-Id: Ic0c4b6f7480a4c6c3f53bd04e285cb0cab172531 Reviewed-on: https://gerrit.libreoffice.org/69888 Tested-by: Jenkins Reviewed-by:
-
Caolán McNamara yazdı
and ensure the modify arrives after the input text has been filtered Change-Id: I16c0bd095700e7d9858e3e74884b5a1ad3eb7e4c Reviewed-on: https://gerrit.libreoffice.org/69764 Tested-by: Jenkins Reviewed-by:
-
László Németh yazdı
to show/handle them in Calc AutoFilter. Partial revert of "tdf#116666 fix Hungarian sorting" (commit 7b1eb631). Change-Id: Ib135f1eac77780e901fb4c427c3f1bb8a5e090b9 Reviewed-on: https://gerrit.libreoffice.org/68014 Tested-by: Jenkins Reviewed-by:
-
Christian Lohmaier yazdı
Change-Id: I1307c30f732e0ea12dc65e1f80503a0145371566
-
Christian Lohmaier yazdı
* Update translations from branch 'libreoffice-6-2' - update translations for 6.2.3 rc1 and force-fix errors using pocheck Change-Id: I76da2876015e25484eb696c3278818f0b4f032d1 -
Stephan Bergmann yazdı
...for compatible changes 61c4f96d "Support AdoptOpenJDK" and 41507db5 "Support Amazon Corretto". Changing <updated> causes jfw_startVM and jfw_getSelectedJRE (both jvmfwk/source/framework.cxx) to fail with JFW_E_INVALID_SETTINGS, which in turn causes functionality that requires a JVM to issue a GUI error dialog stating that the user must select a new JRE in the Options dialog. While that behavior makes sense if a JRE was selected that would no longer be supported by the modified javavendors_*.xml, it is just annoying if an already selected JRE is still supported. And a compatible change to javavendors_*.xml implies that an already selected JRE will still be supported. Now, on master towards LO 6.3, javavendors_*.xml have meanwhile been changed incompatibly with 49c21e31 "Javavendors: Set minimum version to 1.6", so keep <updated> updated there. And the updated <updated> from 61c4f96d "Support AdoptOpenJDK" are included in the published LO 6.2.1, so don't change those back. But the updated <updated> from the subsequent 41507db5 "Support Amazon Corretto" (cherry-picked to libreoffice-6-2 towards LO 6.2.3 as be9c3ff7) have not yet been included in a published LO, so they can be changed back to the previous values. Change-Id: I59113db8c3c6c449947692597bc9a3b2c4504a26 Reviewed-on: https://gerrit.libreoffice.org/69730 Tested-by: Jenkins Reviewed-by:
-
Dennis Francis yazdı
at specified position. At least in case of ScUndoPaste, during undo operation, the content from changed area in main-document is removed before actual undo-paste, so when cloned formula-cell from undo-doc is inserted to main-doc, there is no "old" cell to detach, so "rNewSharedRows" is not populated which is needed for proper re-establishment of listeners if there are formula/ formula-groups adjacent to insertion position. Change-Id: Id2bfb65a2a7a5abb03effbdcc4405015fddc401c Next: fix the vector version of this case and write unit tests. Reviewed-on: https://gerrit.libreoffice.org/69737 Tested-by: Jenkins Reviewed-by:
-
Serge Krot yazdı
Change-Id: Ib241edd07e4c6781d80db274f73146bda310d8c0 Reviewed-on: https://gerrit.libreoffice.org/69827 Tested-by: Jenkins Reviewed-by:
-
- 28 Mar, 2019 6 kayıt (commit)
-
-
Caolán McNamara yazdı
we were emitting page-entered but not emitting the page-leave signal Change-Id: I64131631016331cf102d08244982148ac8c3c69c Reviewed-on: https://gerrit.libreoffice.org/69686 Tested-by: Jenkins Tested-by:
Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>
-
Eike Rathke yazdı
This is a combination of 4 commits. Add unit tests for tdf#123736 and related unshared/regrouped listening Reviewed-on: https://gerrit.libreoffice.org/69372 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
Normally when ScInterpreter has bMatrixFormula set, pMyFormulaCell is set as well (done in the ctor). But in some rare cases, this may not be so, and some functions such as ScRandom() already check for this. But not all do, tdf#123479 specifically crashes because ec974965 added AssertFormulaMatrix() to force bMatrixFormula be set without pMyFormulaCell being set, and ScColumn() doesn't handle this case. Fix this by trying to pass ScFormulaCell* to ScInterpreter when AssertFormulaMatrix() is used, this should ensure the result in the formula editor fits better the edited formula. Since there still may be cases when the cell is not set (e.g. editing a new formula), also handle that case gracefully. Ideally ScSimpleFormulaCalculator should pass ScMarkData to ScInterpreter in such cases so that those functions use that when they can't use GetMatColsRows(), but currently the handling of selections is rather poor in the formula edit dialog: Non-array formulas are simply entered in one cell and the selection is ignored, in case of a multi-selection there's an error dialog only after the dialog is closed, and the result field of the dialog is rather small and doesn't scroll, so e.g. matrix result of ScRandom() wouldn't show more than one item anyway. Given that tdf#123479 is a priority bug, better just fix it and possibly handle selections better somewhen later. Change-Id: I5fcbe1e358fac3623d4917eb0ead8eae00a1e153 Reviewed-on: https://gerrit.libreoffice.org/69161Reviewed-by:
Markus Mohrhard <markus.mohrhard@googlemail.com> Tested-by:
Luboš Luňák <l.lunak@collabora.com> (cherry picked from commit ade1df09) Reviewed-on: https://gerrit.libreoffice.org/69849 Tested-by: Jenkins Reviewed-by:
-
Miklos Vajna yazdı
Regression from commit 50223ea6 (tdf#98987 sw: add AddVerticalFrameOffsets compat mode, 2016-03-31), SwFEShell::ImpEndCreate() was not adapted to call SwTextFrame::GetBaseVertOffsetForFly() when determining the vertical position of the inserted shape. The call can be unconditional, the returned value is already 0 when the DocumentSettingId::ADD_VERTICAL_FLY_OFFSETS compat setting is false. (cherry picked from commit 4218caf1) Conflicts: sw/qa/extras/uiwriter/uiwriter2.cxx Change-Id: Iec6af5a6d1ff3466e08377853cc8ca84f33a76d1 Reviewed-on: https://gerrit.libreoffice.org/69766 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
Apparently the Color class uses transparency as the opposite of the normally used meaning of opacity, so transparency 255 means transparent. Change-Id: I3a76c2c93d98ad3c850017d3fd569b04dc6c19c8 Reviewed-on: https://gerrit.libreoffice.org/69211 Tested-by: Jenkins Reviewed-by:
Miklos Vajna <vmiklos@collabora.com>
-
Luboš Luňák yazdı
If the bitmap to be written is non-8bit, e.g. in case of using OpenGL (on Linux the test requires "SAL_FORCEGL=1 SAL_USE_VCLPLUGIN=gen", on Windows it's probably the default case), then the bitmap is not in native format for 8bit. This was done correctly by 45e8e0fb that introduced it but somehow it got lost later (I can't even find where). Reviewed-on: https://gerrit.libreoffice.org/69213 Tested-by: Jenkins Reviewed-by:
-
- 27 Mar, 2019 3 kayıt (commit)
-
-
Caolán McNamara yazdı
Change-Id: Ie14600e9f9a1e1c4e99c7a872f5d677453481888 Reviewed-on: https://gerrit.libreoffice.org/69666 Tested-by: Jenkins Reviewed-by:
-
Julien Nabet yazdı
See https://bugs.documentfoundation.org/show_bug.cgi?id=123999#c14 Change-Id: I47ff6171c873c701794f179b924005c43551ba21 (cherry picked from commit 3149ffea) Reviewed-on: https://gerrit.libreoffice.org/69786Reviewed-by:
Lionel Elie Mamane <lionel@mamane.lu> Tested-by: Jenkins
-
Luboš Luňák yazdı
If the bitmap to be drawn will need to be scale for drawing, OpenGL handles this by scaling, but the scaling is of poor quality. Other backends scale in such a case with a good quality, and e.g. opengl's drawAlphaBitmap() also does the same check, so copy that check to drawBitmap() as well. Change-Id: If6a457c69c6676d03fa4046b9910683f51479d21 Reviewed-on: https://gerrit.libreoffice.org/69556 Tested-by: Jenkins Reviewed-by:
Michael Meeks <michael.meeks@collabora.com>
-
- 26 Mar, 2019 6 kayıt (commit)
-
-
Thorsten Behrens yazdı
Temp fix, since update check is an extension, and extensions can't currently modify notebookbars. This needs implementation of something like https://wiki.openoffice.org/wiki/Framework/Article/Addon_Menu_Toolbar_Merging for notebookbar popup menus. Such that Addons.xcs provide some merge points with notebookbar menu IDs, we can then plug into existing popup menus. Change-Id: If30525fec1e7be0429e351588614618c5f56cdd4 Reviewed-on: https://gerrit.libreoffice.org/69735Reviewed-by:
-
Samuel Mehrbrodt yazdı
Reintroduce the UpdateFields call, but guard it so that it does not modify the doc (as it was before) Change-Id: I23e814f0b5d900d145dc82272034f03e66c079ed Reviewed-on: https://gerrit.libreoffice.org/69696 Tested-by: Jenkins Reviewed-by:
-
Eike Rathke yazdı
This was done long ago for the tr-TR locale as well. Change-Id: I5bf8595f6d49adb7fd76b3c4924c4d72b3b8ea5e Reviewed-on: https://gerrit.libreoffice.org/69717Reviewed-by:
-
Eike Rathke yazdı
This is a combination of 7 commits. Resolves: tdf#123714 tdf#123736 all split formula groups; tdf#120013 related Add all split shared formula groups to regrouping and mark for listening, even if the references had not be to adjusted. This partly also resolves tdf#120013 but there's more to that, a remaining partial group is not updated. Reviewed-on: https://gerrit.libreoffice.org/68951Reviewed-by:
-
Mike Kaganski yazdı
https://aws.amazon.com/corretto Change-Id: I7f9ef25e1e81ededce426bad661adc6880feb9d3 Reviewed-on: https://gerrit.libreoffice.org/67289 (cherry picked from commit 41507db5) Reviewed-on: https://gerrit.libreoffice.org/69688 Tested-by: Jenkins Reviewed-by:
Stephan Bergmann <sbergman@redhat.com>
-
Noel Grandin yazdı
regression from commit 6be7e2e9 Date: Thu Aug 30 13:54:33 2018 +0200 pass SdrUndoAction around by std::unique_ptr Looks like previously this memory was just leaked, so fix that and the crash Change-Id: I5eb5c104dccd82cef30af533ce1fe8546ced0ea0 Reviewed-on: https://gerrit.libreoffice.org/69648 Tested-by: Jenkins Reviewed-by:
-
