- 05 Haz, 2019 11 kayıt (commit)
-
-
Christian Lohmaier yazdı
Change-Id: Ie796a5c8f36b33c61eb69b6b68f7439f9359526f (cherry picked from commit b1a17c2d)
-
Michael Stahl yazdı
The first problem here is that the table isn't fully formatted; it fails with: warn:legacy.osl:22975:22975:sw/source/core/layout/tabfrm.cxx:2639: debug assertion: <SwTabFrame::MakeAll()> - format of table lowers suppressed by fix i44910 The situation is that there is a big table split across pages; the first line of the table would fit onto the previous page so its follow frame moves backward and then the table frame tries to split again. During SwTabFrame::Split(), all the frames in the table are formatted, and at that point a footnote that was on the next page is moved to this page. A nested table frame also splits, such that it fits inside the page... but then the split of the outer table fails by 5 twips, because the moved footnote has reduced the space available for the outer table. The footnote is anchored in the inner table's follow frame, which would be moved to the next page anyway, taking the footnote with it. Fix this in lcl_RecalcSplitLine() by checking for footnotes that are anchored in the follow frame of the top-level table being split, and adding their height to the available space on the page. Fixing the first problem avoids the crash as well; the crash happens since 18765b9f and it's rather hard to avoid it in a situation where formatting starts at the end and recurses into an unformatted table preceding it, which isn't supposed to happen. Change-Id: I85286583c1c4930468a1c283afc98504cd35bb71 Reviewed-on: https://gerrit.libreoffice.org/73465 Tested-by: Jenkins Reviewed-by:
Michael Stahl <Michael.Stahl@cib.de> (cherry picked from commit e37ffdd1) Reviewed-on: https://gerrit.libreoffice.org/73535
-
László Németh yazdı
Default auto values must be outside of the new enlarged range of the superscript/subscript percent values. Note: the raising limit was modified to 13999 from 14400, because the RTF unit test tdf112208_hangingIndent.rtf lost its hanging from the bigger value. Change-Id: I3a7e9715a27570278ee4ee70f9fc9ad29457e100 Reviewed-on: https://gerrit.libreoffice.org/73166 Tested-by: Jenkins Reviewed-by:
László Németh <nemeth@numbertext.org> (cherry picked from commit 32262b0a) Reviewed-on: https://gerrit.libreoffice.org/73460
-
Tamas Bunth yazdı
Which is mapped to sal_Int32. Change-Id: Ibf12e92a20034440fa990ed0c6f1196f4ca3f40f Reviewed-on: https://gerrit.libreoffice.org/73217 Tested-by: Jenkins Reviewed-by:
Andras Timar <andras.timar@collabora.com> (cherry picked from commit 3a4c4482) Reviewed-on: https://gerrit.libreoffice.org/73459Reviewed-by:
Xisco Faulí <xiscofauli@libreoffice.org>
-
Mark Hung yazdı
Lines per page and chars per line can not be set before the max values determined, otherwise we may get incorrect result because of truncating to incorrect max values. Change-Id: I9829af0dcfd744f4383a5f9034270b00fdfb15b8 Reviewed-on: https://gerrit.libreoffice.org/73280 Tested-by: Jenkins Reviewed-by:
Mark Hung <marklh9@gmail.com> (cherry picked from commit c1c3a3c6) Reviewed-on: https://gerrit.libreoffice.org/73458Reviewed-by:
-
László Németh yazdı
Deleted fields in tracked changes, for example page number fields, imported as not deleted text fields. Change-Id: Id35c3ad7d293039fc8a2d0b564572dc459a25d0f Reviewed-on: https://gerrit.libreoffice.org/73397 Tested-by: Jenkins Reviewed-by:
-
Mark Hung yazdı
Non-BMP characters use two sal_Unicodes. It is necessary to advnace with OUString::iterateCodePoints safely instead of nCharPos + 1. Change-Id: Iea1005f05421b1021a9b3ff611439357d4dcba9e Reviewed-on: https://gerrit.libreoffice.org/73326 Tested-by: Jenkins Reviewed-by:
Khaled Hosny <khaledhosny@eglug.org> (cherry picked from commit 16faad36) Reviewed-on: https://gerrit.libreoffice.org/73461Reviewed-by:
-
Olivier Hallot yazdı
Change-Id: Ia77b5a407693802f408cd7d28d50962ae1c746e7 Reviewed-on: https://gerrit.libreoffice.org/73308 Tested-by: Jenkins Reviewed-by:
Heiko Tietze <tietze.heiko@gmail.com> (cherry picked from commit 27cd31a7) Reviewed-on: https://gerrit.libreoffice.org/73456
-
Olivier Hallot yazdı
Draw and Impress have different menus names Change-Id: I8e0dec96e3b52f87bb5e7430f21c96f3b4a681a5 Reviewed-on: https://gerrit.libreoffice.org/73117Reviewed-by:
Olivier Hallot <olivier.hallot@libreoffice.org>
-
Tor Lillqvist yazdı
Change-Id: I17d95cce0dd69e2ba8b4ebc7045544fc5659a249 Reviewed-on: https://gerrit.libreoffice.org/73469Reviewed-by:
Tor Lillqvist <tml@collabora.com> Tested-by:
-
Tor Lillqvist yazdı
The DISABLE_DYNLOADING case in ImpFilterLibCacheEntry::GetImportFunction() had apparently been broken from the start, but nobody never noticed as the common formats (PNG, JPEG, GIF etc) use a more "internal"code path. The problem was noticed now with a TIFF image in a presentation. The fix for the import case seems to work, but I could not test the fix for the export case. Change-Id: Icbd33c7cb654136a562f18b06d5b0896e701bf3d Reviewed-on: https://gerrit.libreoffice.org/73468 Tested-by: Jenkins Reviewed-by:
-
- 04 Haz, 2019 17 kayıt (commit)
-
-
Mike Kaganski yazdı
Regression from commit 1fe24bb1 Change-Id: I5597fe6a7f7c54ad9bf2634eba5245e2e4a1efbf Reviewed-on: https://gerrit.libreoffice.org/73430 Tested-by: Jenkins Reviewed-by:
Mike Kaganski <mike.kaganski@collabora.com> (cherry picked from commit 166aafde) Reviewed-on: https://gerrit.libreoffice.org/73452
-
Caolán McNamara yazdı
Change-Id: I802d4c92924b379f11c73ddfe758f4c6a63feb37 Reviewed-on: https://gerrit.libreoffice.org/73454 Tested-by: Jenkins Reviewed-by:
Caolán McNamara <caolanm@redhat.com> Tested-by:
-
Miklos Vajna yazdı
Regression from commit a4cb27f6 (refactor ScMenuFloatingWindow to use RenderContext, 2015-05-14), the problem was that the change had two effects: 1) It removed incremental rendering that now causes flicker. 2) It added support for painting into a rendercontext, not on the window itself. Given that here a hierarchy of widgets have to be painted at once (to avoid flicker), use vcl::Window::RequestDoubleBuffering() instead of the much simpler vcl::BufferDevice. This improves the situation on Windows, but explicitly blacklist the focus rectangle case (gtk3) which is currently not compatible with this higher level double-buffering (and there we already have lower-level double-buffering anyway, similar to OpenGL). (cherry picked from commit 4dadbff7) Change-Id: I77aa951e14db668166dc8757e36c037ee98243bf Reviewed-on: https://gerrit.libreoffice.org/73451 Tested-by: Jenkins Reviewed-by:
Miklos Vajna <vmiklos@collabora.com>
-
Tamás Zolnai yazdı
... for bullet customization Change-Id: Ica66dcc0d3f379886788d426941b98d170444799 Reviewed-on: https://gerrit.libreoffice.org/73306 Tested-by: Jenkins Reviewed-by:
Tamás Zolnai <tamas.zolnai@collabora.com> (cherry picked from commit 7dc6525f) Reviewed-on: https://gerrit.libreoffice.org/73318Reviewed-by:
-
wishawa yazdı
* Update dictionaries from branch 'libreoffice-6-3' - Remove abbreviations from Thai hunspell dictionary. Reasons: 1. They are not present in other spellcheck dictionaries (such as en). 2. To add all abreviations would not be much different from adding all possible combinations of 2-3 Thai consonants. This would make misspelled abbreviations come as correct to the spellcheck anyway. 3. Adding only the most common ones can cause confusion to those using less common ones, as the spellchecker will suggest the more common abbreviation. It is also nearly impossible to determine if an abbreviation is common that it should be added or not. Change-Id: I8b7bce70f552ab3b356e1ab2c101e9e31a8ec271 Reviewed-on: https://gerrit.libreoffice.org/73216Reviewed-by: -
Noel Grandin yazdı
Use a more useful expansion strategy when re-sizing the array, we use this 150% in a lot of other places. This takes the opening time from more than 4m (I lost patience) to under 5s. Change-Id: I8511662947d737c26ec49503a75af8d4cf447f67 Reviewed-on: https://gerrit.libreoffice.org/73394 Tested-by: Jenkins Reviewed-by:
Noel Grandin <noel.grandin@collabora.co.uk> (cherry picked from commit 79927a93) Reviewed-on: https://gerrit.libreoffice.org/73433Reviewed-by:
-
Eike Rathke yazdı
If possible here means if it can't be an operator (is not an ASCII value) and is not any other separator. This restores the "working by chance" behaviour for some locales that use NO-BREAK SPACE as group separator. It never worked for locales that use one of the "ordinary" comma or dot or apostrophe group separator and will not. Change-Id: I8b7ba1b502b6b7367ffb2199d2f3922ab605a659 Reviewed-on: https://gerrit.libreoffice.org/73422Reviewed-by:
Eike Rathke <erack@redhat.com> Tested-by: Jenkins (cherry picked from commit 4e3e0c67) Reviewed-on: https://gerrit.libreoffice.org/73436
-
Miklos Vajna yazdı
Relative size should have priority over image size from image headers. (cherry picked from commit a628c01e) Change-Id: I80cce79ff87c7eb359cb292526859435cb85f929 Reviewed-on: https://gerrit.libreoffice.org/73434 Tested-by: Jenkins Reviewed-by:
-
heiko tietze yazdı
Zero remembered searches crash the app when no result is found minimum set to 1 Change-Id: Ib2b4b195d018f0f11238699458066f4d128f8997 Reviewed-on: https://gerrit.libreoffice.org/72459 Tested-by: Jenkins Reviewed-by:
-
Noel Grandin yazdı
When we shut down, we destroy the various caches, in the process of which SalBitmap calls back into it's owning cache, causing a SIGSEGV. Found while loading files from tdf#83426 Change-Id: I53db1621a0fdb75a8e66582662b0e2666499192b Reviewed-on: https://gerrit.libreoffice.org/73387 Tested-by: Jenkins Reviewed-by:
-
Noel Grandin yazdı
WPXSvInputStreamImpl was hammering on getFilePos pretty hard, and getFilePos uses a mutex, which is slow when it is called from every single read. So switch to using std::atomic to access position. This is specifically fixing the performance of queryTypeByDescriptor when called from a basic macro on a local test file. This takes my test macro from 8s to 4s. Change-Id: Iab707a374359e2ee0e92425b2d9a903d67cb53d4 Reviewed-on: https://gerrit.libreoffice.org/73377 Tested-by: Jenkins Reviewed-by:
-
Noel Grandin yazdı
Size the OUStringBuffer in OrcusFormatDetect::detect better at construction, to avoid re-alloc. This is specifically fixing the performance of queryTypeByDescriptor when called from a basic macro on a local test file. This takes my test macro from 9s to 8s Change-Id: Id4c2bf386de172b1deed98efece8c1f040c599a5 Reviewed-on: https://gerrit.libreoffice.org/73376 Tested-by: Jenkins Reviewed-by:
-
Noel Grandin yazdı
SfxMedium was creating a temporary file and copying the input file, when CloseInStream_Impl was called from the destructor. Very very bad for performance. This is specifically fixing the performance of queryTypeByDescriptor when called from a basic macro on a local test file. This takes my test macro from 16.1s to 9s. Change-Id: If52f8e0587c7b11666893f6cb79b3180d45cce43 Reviewed-on: https://gerrit.libreoffice.org/73375 Tested-by: Jenkins Reviewed-by:
-
Noel Grandin yazdı
Skip creating an intermediary buffer in XStream_impl::readBytes, and just read directly into the destination. This is specifically fixing the performance of queryTypeByDescriptor when called from a basic macro on a local test file. This takes my test macro from 17.1s to 16.1s. Change-Id: Iaa7d38c6a90a3b3f01a4b748c4512dd8fda690c7 Reviewed-on: https://gerrit.libreoffice.org/73374 Tested-by: Jenkins Reviewed-by:
-
Roman Kuznetsov yazdı
on Outline & Numbering tab in Paragraph style dialog Change-Id: I2e97dfae832c497c5d08f00eb8c6e6b97884b717 Reviewed-on: https://gerrit.libreoffice.org/72551 Tested-by: Jenkins Reviewed-by:
Adolfo Jayme Barrientos <fitojb@ubuntu.com>
-
Jan-Marek Glogowski yazdı
We just have one driving timer. Not much to sync with. Gets rid of these annoying scheduler restart messages. The idea of coarse, synced timer events sounds nice to have, but LO currently can't handle it in a sensible sense.. Change-Id: I169c2bb582aca0e03b00e24dc364ac9b19162c3b Reviewed-on: https://gerrit.libreoffice.org/73392 Tested-by: Jenkins Reviewed-by:
Jan-Marek Glogowski <glogow@fbihome.de> (cherry picked from commit 677db601) Reviewed-on: https://gerrit.libreoffice.org/73411Reviewed-by:
Michael Weghorn <m.weghorn@posteo.de>
-
Jan-Marek Glogowski yazdı
Change-Id: I708476919ce7750df6a691d75fe594c3f44d9126 Reviewed-on: https://gerrit.libreoffice.org/73173 Tested-by: Jenkins Reviewed-by:
-
- 03 Haz, 2019 8 kayıt (commit)
-
-
Michael Stahl yazdı
Empty section frames cause trouble for the assert added in e14056e6: * existing ones may be skipped by MoveBwd(), so a SwTextFrame moves to the previous page but the SwSectionFrame that was its mpPrev remains on this page; e.g. ooo110854-1.rtf * a SwSectionFrame moves backward, then some SwTextFrame in it splits and the follow tries to move forward, resulting in *new* SwSectionFrames on this page that are empty (because the follow is eventually joined again); e.g. abi3213-1.rtf So if there's a section frame, assume it's SNAFU and don't assert. Change-Id: Ibdca2aa39fae123583e5edf5173f80e8b70ef335 Reviewed-on: https://gerrit.libreoffice.org/73333 Tested-by: Jenkins Reviewed-by:
Thorsten Behrens <Thorsten.Behrens@CIB.de>
-
Caolán McNamara yazdı
so turn on sorting on first sort request Change-Id: I897f64c0695be8cca6d93e234a7d7c4718c94f78 Reviewed-on: https://gerrit.libreoffice.org/73345 Tested-by: Jenkins Reviewed-by:
-
László Németh yazdı
Non-en_US locale setting resulted broken XLSX export with incorrect localized range separator. The previous commit 19976f07 "tdf#112567 XLSX export: fix broken built-in names" only avoided of the bad export of the XSLX documents with correct definedName. Note: this commit fixes the bad XLSX documents during the next import/export cycle. Change-Id: I9101c5df0f29c7a42fd49f4c2765dcf47a711916 Reviewed-on: https://gerrit.libreoffice.org/73220 Tested-by: Jenkins Reviewed-by:
-
Eike Rathke yazdı
Change-Id: Ie9781713f04076174cdbbc34bf802df45c169727 Reviewed-on: https://gerrit.libreoffice.org/73389Reviewed-by:
-
Michael Weghorn yazdı
Handle special values for offset in 'Qt5AccessibleWidget::attributes' the same way that the base class's 'QAccessibleTextWidget::attributes' does, s. [1]. In particular, an offset matching the text length can be passed e.g. by "accerciser" or screen readers, which previously resulted in an 'IndexOutOfBoundsException' being thrown when the index was later checked in the call to 'VCLXAccessibleStatusBarItem::getCharacterAttributes'. See also 'IAccessibleText::attributes' documentation at [2] and the page on special offset values referenced there [3]. [1] https://code.qt.io/cgit/qt/qtbase.git/tree/src/widgets/accessible/qaccessiblewidgets.cpp?h=5.12.4#n791 [2] https://accessibility.linuxfoundation.org/a11yspecs/ia2/docs/html/interface_i_accessible_text.html#a29e5c8f69ec13c683ed6bca53333e6a5 [3] https://accessibility.linuxfoundation.org/a11yspecs/ia2/docs/html/_general_info.html#_specialOffsets Change-Id: I623995aeb689b31c5b49fb3ace8e4dd4c18927d2 Reviewed-on: https://gerrit.libreoffice.org/73225 Tested-by: Jenkins Reviewed-by:
Katarina Behrens <Katarina.Behrens@cib.de> (cherry picked from commit 93c81657) Reviewed-on: https://gerrit.libreoffice.org/73378Reviewed-by:
-
Xisco Fauli yazdı
Use same logic as in vcl/unx/generic/gdi/gdiimpl.cxx Regression from 16091ff8 See https://bugs.documentfoundation.org/show_bug.cgi?id=125506#c13 Thanks to Thorsten Wagner for pinpointing Change-Id: I7a7a8c4b3355f5621ba1603939a3757cd03e7777 Reviewed-on: https://gerrit.libreoffice.org/73319 Tested-by: Jenkins Reviewed-by:
-
Jan-Marek Glogowski yazdı
Grep'ing the Qt source code, most of the places call toInt() on the QVariants returned from Qt::ImCursorPosition and Qt::ImAnchorPosition (toUInt() is probably a bug). So actually deliver these (and fix the 32bit build with the sal_Int32 ambiguity - there is no "signed long" QVariant constructor). Change-Id: Id6b8fbabb8156692592bdef48437db140c6b7cca Reviewed-on: https://gerrit.libreoffice.org/73356Tested-by:
Rene Engelhard <rene@debian.org> Reviewed-by:
-
Olivier Hallot yazdı
* Update helpcontent2 from branch 'libreoffice-6-3' - tdf#125645 Fix reference in SWITCH function help page Change-Id: I34cd365829187ea29fd45c43aa08d0b6a3cfdd4d Reviewed-on: https://gerrit.libreoffice.org/73352 Tested-by: Jenkins Reviewed-by:
-
- 02 Haz, 2019 4 kayıt (commit)
-
-
Caolán McNamara yazdı
Change-Id: If31cf3c02ab714f45ebbc0139e0e77102b5eb2ba Reviewed-on: https://gerrit.libreoffice.org/73342Reviewed-by:
-
Caolán McNamara yazdı
The gtk one does, but the gen one overwrite the default handler instead of adding to it Change-Id: I40c760821cb46283c00624797e97d601babd6200 Reviewed-on: https://gerrit.libreoffice.org/73340Reviewed-by:
-
Caolán McNamara yazdı
Change-Id: I13019ae1b991b5910c50e760a8a825bcf50509d2 Reviewed-on: https://gerrit.libreoffice.org/73336Reviewed-by:
-
Caolán McNamara yazdı
Change-Id: If4ee4e09f5549ef347c7e97b1c2e68b3971b8b19 Reviewed-on: https://gerrit.libreoffice.org/73284 Tested-by: Jenkins Reviewed-by:
-
