summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2012-09-21fdo#42450: fix crash in SwXTextSection::attach:Michael Stahl1-0/+8
Inserting the section can fail if the given text range is not valid. Change-Id: Ib6ba3b02dd581dce08b646b841354073caf894b2 (cherry picked from commit 5bc6c7b2e170a35914d7cd07347c77a9c9d23664 and commit 2940a697dce2292e9c5b118feb8ec4159e3606fa) Reviewed-on: https://gerrit.libreoffice.org/602 Reviewed-by: Norbert Thiebaud <nthiebaud@gmail.com> Reviewed-by: Miklos Vajna <vmiklos@suse.cz> Tested-by: Miklos Vajna <vmiklos@suse.cz> (cherry picked from commit 525127d5b605bf34a8ff0b71f4b169fa7b72c435) Reviewed-on: https://gerrit.libreoffice.org/607 Reviewed-by: Fridrich Strba <fridrich@documentfoundation.org> Tested-by: Fridrich Strba <fridrich@documentfoundation.org>
2012-09-20fdo#54609, fdo#45366: Exception while importing xlsx.Muthu Subramanian1-7/+15
* Handle any exception thrown during document properties import. This is not so critical so as to stop the import. We anyways check for hasElements(). * Also lclGetRelatedStreams might throw IllegalArgumentException Change-Id: I5206ff1ab07f2853e6a20aebea3f892460b881ad Signed-off-by: Kohei Yoshida <kohei.yoshida@gmail.com> Signed-off-by: Markus Mohrhard <markus.mohrhard@googlemail.com> Signed-off-by: Michael Meeks <michael.meeks@suse.com>
2012-09-20fdo#37529 clear DataColumns read-only information when we clear DataColumnsLionel Elie Mamane1-0/+3
Change-Id: I6f3b4f1646df1244d97ae98f56ba054bc26e3314 Reviewed-on: https://gerrit.libreoffice.org/608 Reviewed-by: Michael Stahl <mstahl@redhat.com> Tested-by: Michael Stahl <mstahl@redhat.com> Signed-off-by: Miklos Vajna <vmiklos@suse.cz> Signed-off-by: Michael Meeks <michael.meeks@suse.com>
2012-09-19import values and formulas from lotus-123, fdo#55066Markus Mohrhard1-3/+3
Change-Id: Ia403c849ef537c7f7804f907889110b8096c0bb6 Signed-off-by: Michael Meeks <michael.meeks@suse.com> Signed-off-by: Fridrich Strba <fridrich.strba@graduateinstitute.ch> Signed-off-by: Kohei Yoshida <kohei.yoshida@gmail.com>
2012-09-17fdo#44721 protect against negative calculated height valueLionel Elie Mamane1-2/+4
Change-Id: I4faedd535421eaf1e3e425f4731b3bb8831e7360 Reviewed-on: https://gerrit.libreoffice.org/624 Reviewed-by: Michael Meeks <michael.meeks@suse.com> Tested-by: Michael Meeks <michael.meeks@suse.com>
2012-09-12bump product version to 3.6.2.1+, release number to 1Fridrich Štrba2-22/+22
Change-Id: I784eede0b5aa06c90bd0caa427f9c16aae07ba79 Signed-off-by: Petr Mladek <pmladek@suse.cz>
2012-09-11Branch libreoffice-3-6-2Petr Mladek0-0/+0
This is 'libreoffice-3-6-2' - the stable branch for the 3.6.2 release. Only very safe changes, reviewed by three people are allowed. If you want to commit more complicated fix for the next 3.6.x release, please use the 'libreoffice-3-6' branch. If you want to build something cool, unstable, and risky, use master. Change-Id: Idd0d2b186e9113a4a3a30ed9ae03918577ed30ec
2012-09-11missing part of fix for fdo#54721Noel Power1-1/+1
commit id cea414bfe8fe356793778bdf09f5ca3b3a42daa0 for "fix vba not tracking currently selected doc correctly" was missing necessary change in sfx2 Change-Id: I1697c005374b3d26dc63c78702a46c6d9879adaa Signed-off-by: Petr Mladek <pmladek@suse.cz>
2012-09-11fdo#54721 fix vba not tracking currently selected doc correctlyNoel Power1-8/+3
Change-Id: I43b17dfc621d0aba76cdd33ebf97c457fa1da912 Signed-off-by: Petr Mladek <pmladek@suse.cz>
2012-09-11fdo#53588 Assume cdir being in the last 1k of the file to speed-up searchAndras Timar1-0/+1
Change-Id: I25291c52d9f0117c70c48781c2eb575bb6b374ff Signed-off-by: Fridrich Štrba <fridrich.strba@bluewin.ch>
2012-09-11fdo#54718 fix opcode detection in basic resulting in failed/unregcognized codeNoel Power1-3/+3
fix is followup to bf5b493104d2dd4ab964f1fcb845200eaefbbcd5 which didn't adjust the runtime to cater for the new enum layout Change-Id: I6613fb8aacd8a70947c4fff556fb3e2d33c1113e Signed-off-by: Petr Mladek <pmladek@suse.cz>
2012-09-11Enable Evolution Adress Book support in official GNU/Linux buildLionel Elie Mamane1-0/+1
Conflicts: distro-configs/LibreOfficeLinux.conf Change-Id: Id05436f38477d73be308310c7ef0ec3baca01bc9 Signed-off-by: Petr Mladek <pmladek@suse.cz>
2012-09-11fdo#53113 various docx shape export fixesMiklos Vajna6-11/+145
- docx export: initial shape text support - vml export: handle custom segment types - docx: export fillBlip shape property Change-Id: I8453d76cfb91386c8beaedc874e09a291234babb Reviewed-on: https://gerrit.libreoffice.org/411 Reviewed-by: Bosdonnat Cedric <cedric.bosdonnat@free.fr> Tested-by: Bosdonnat Cedric <cedric.bosdonnat@free.fr>
2012-09-10Calculate positions of cell-anchored objects upon ods import (fdo#54695)Kohei Yoshida1-0/+3
Since we no longer re-calc row heights on ods import, we need to do this manually, in order to position cell-anchored objects correctly. Previously we were getting this for free since the row height recalc code path did it as part of it. Change-Id: I8ab5dd1fe7cd8a45b7968e101c893b442d7ce132 Signed-off-by: Noel Power <noel.power@suse.com>
2012-09-10gbuild: disable symbols on --enable-dbgutil --disable-symbolsMichael Stahl4-9/+17
Due to the setup of gb_DEBUGLEVEL in gbuild.mk, gb_SYMBOL was always enabled when --enable-dbgutil is set, with no way to override it. Fix that by turning configure's ENABLE_SYMBOLS into a tri-state, where the new "FALSE" value, set by an explicit --disable-symbols, overrides any implicit way of enabling symbols. But by default an --enable-dbgutil still enables gb_SYMBOL. Change-Id: I94c609863980ed1ab9c73d7a4861c394442b531d (cherry picked from commit d9cbc837f84c415f5949a6893764dae8268f0d85) Reviewed-on: https://gerrit.libreoffice.org/572 Reviewed-by: Petr Mladek <pmladek@suse.cz> Tested-by: Petr Mladek <pmladek@suse.cz>
2012-09-10fdo#52246 let helpex use gsicheck's tag checkerAndras Bartek8-454/+558
So the build will not break on translations errors in help. English segment will be used, when the localized segment has fatal error in it, e.g. a tag error that results in invalid xml. Change-Id: I8a74649d1f00f6379a7244f2582d27b8a437cd2f Signed-off-by: Andras Timar <atimar@suse.com>
2012-09-10deb#681185: wait for InternalIPC::ProcessingDone in oosplashStephan Bergmann1-0/+11
Change-Id: I9606ec34ad0d8d24dca71d060bf5e1d37152903c (cherry picked from commit 5c804dce946a928adcaf97533f5345b33e688ff5) Reviewed-on: https://gerrit.libreoffice.org/569 Reviewed-by: Michael Meeks <michael.meeks@suse.com> Tested-by: Michael Meeks <michael.meeks@suse.com>
2012-09-10fdo#38913: Prevent invalid parameter handler crashesStephan Bergmann1-0/+29
It appears that on Windows at least some jvm.dll versions can cause calls to _fileno(NULL), which leads to a call of the invalid parameter handler (see <http://msdn.microsoft.com/en-us/library/ksazx244%28v=vs.80%29.aspx> "Parameter Validation: Visual Studio 2005"). The default handler causes the application to crash, so install a "harmless" one instead. Change-Id: Id6a3ffb63f70b0c65546bc933e994c8dbf35203c (cherry picked from commit a82e532ce006c54b2740de74d1da5d11307da7c1) Reviewed-on: https://gerrit.libreoffice.org/564 Reviewed-by: Michael Meeks <michael.meeks@suse.com> Tested-by: Michael Meeks <michael.meeks@suse.com>
2012-09-10fdo#53557: use metadata's idea of column count to allocate a rowLionel Elie Mamane1-2/+3
m_pColumnNames contains only the columns that come straight from a table, not calculated columns Change-Id: I2706790e8c1447d1c4e144bbdef773deeca3e2e8 Reviewed-on: https://gerrit.libreoffice.org/522 Reviewed-by: Michael Meeks <michael.meeks@suse.com> Tested-by: Michael Meeks <michael.meeks@suse.com>
2012-09-10fdo#47368: Fix for infinite recursionMichael Meeks1-9/+3
See bug for discussion. Change-Id: Id45dd2902bd2cb68618c3e7324867f226c025698 Signed-off-by: Tor Lillqvist <tml@iki.fi>
2012-09-10fdo#50584 bundle MS Visual C++ 2010 Runtime so LO can work with JRE7Andras Timar2-8/+20
(cherry picked from commit 8b749cb16dde14b199a61456741b586ddd88a584 plus 238eba337cedda17a9ed1ce7e91755c66446195a "make the vc2010 redist optional") Change-Id: If22bed6728121ec66ecd89ad4f869e980aeb79b7 Signed-off-by: Stephan Bergmann <sbergman@redhat.com>
2012-09-10fdo#47368: Remove event listener also in Invalidate()Tor Lillqvist2-0/+9
Change-Id: Ifa3d7ff400137b4560fa0ff890cd211147c5c234 Signed-off-by: Michael Meeks <michael.meeks@suse.com>
2012-09-10don't use unitialized global variable, fdo#53713Markus Mohrhard3-6/+15
Change-Id: I9e5150fb4c5681abe9c786f4c2d80f5a10e8a66b (cherry picked from commit 2ac26e6b5248f2f3fde8dc4341b01ef15b6c3a29) Signed-off-by: David Tardon <dtardon@redhat.com>
2012-09-07fdo#54450 install evolution DB driver correctlyDavid Tardon4-6/+12
Change-Id: I0b6eb67014a61e22981f1f6be488fa762d639554 Signed-off-by: Michael Meeks <michael.meeks@suse.com>
2012-09-07fdo#54620 do not use vector iterator after insertDavid Tardon1-6/+7
Change-Id: I5e809f38a500ab818e9acef80b419dfece7a29fa (cherry picked from commit 06ae26b5676c29c4d1cc9c65f5bbaede9483a21d) Signed-off-by: Caolán McNamara <caolanm@redhat.com>
2012-09-07Related: fdo#53337 Maybe this is a typeinfo visibility problemCaolán McNamara2-0/+14
Change-Id: I024de5b400d0f7a443d4cce7bfe4e88841e6049b Signed-off-by: Michael Meeks <michael.meeks@suse.com>
2012-09-07multiline tooltip for very long (eg. footnote) texts #i42424#László Németh1-2/+4
Change-Id: Ib6383220c3cdd6fef55dee82fcfdc3472da45163 Signed-off-by: Andras Timar <atimar@suse.com>
2012-09-07resolved fdo#52240 fdo#52137 fdo#52288 user editable date patternsEike Rathke14-46/+429
Implemented user editable date acceptance patterns. The introduction of strict date parsing using locale dependent date acceptance patterns in 3.6.0 wasn't always welcomed. Besides that not every locale had patterns for incomplete (only day and month) date input, users also complained about not being able to key in dates on numeric keypads if the locale's date separator wasn't '/' or '-' This commit implements a "Date acceptance patterns" edit field under Tools->Options->LanguageSettings->Languages that follows the selected locale and enables the user to add patterns. Example de-DE locale: * default patterns: D.M.Y;D.M. * to enable additional input on numeric keypad: D.M.Y;D.M.;D-M-Y;D-M * if 3-4 shall not result in a date, D-M- could be used instead of D-M * note that to enter an ISO 8601 Y-M-D date with a D-M-Y pattern active one needs to enter a year >31 or with at least 3 digits, e.g. 011 (cherry picked from commit bf10f4d62a5fe308ea47f9a0aac4f6f7e264ae3e) Conflicts: cui/source/options/optgdlg.hrc svl/source/numbers/zforfind.cxx unotools/source/config/syslocaleoptions.cxx Change-Id: Ic1ce91b1f9d29f1837d56b45ba0ae16f6d9cb17c Reviewed-on: https://gerrit.libreoffice.org/511 Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com> Reviewed-by: Petr Mladek <pmladek@suse.cz> Tested-by: Petr Mladek <pmladek@suse.cz> Reviewed-by: Caolán McNamara <caolanm@redhat.com>
2012-09-06Inadvertent omit of Events node from Jobs-onlineupdate.xcuStephan Bergmann1-1/+1
...causing update check to not run due to the missing JobList entry. Change-Id: If855259ac86340763e4b0fa8266f011f8980d61b Reviewed-on: https://gerrit.libreoffice.org/570 Reviewed-by: Jan Holesovsky <kendy@suse.cz> Tested-by: Jan Holesovsky <kendy@suse.cz>
2012-09-06assign correct valueMarkus Mohrhard1-1/+1
(cherry picked from commit 160a88e18d868433813f3f09b340dcaf68f276b2) Signed-off-by: Stephan Bergmann <sbergman@redhat.com>
2012-09-06Resolves: fdo#54486 ensure ordinal format has non-ordinal prefixCaolán McNamara1-33/+69
Make sure that the ordinal format and the non-ordinal format match at the start, so that the expectation can be verified that there is some trailing "ordinal suffix" which can be extracted. I'm not at all sure that this will be true for all languages. Change-Id: Ib89a51cdd743db12bfe6e723a9acb61d9fe5e6ce (cherry picked from commit a05357ab69712bec53c2d8d17efbbf25907ff9b8) Signed-off-by: Michael Stahl <mstahl@redhat.com>
2012-09-06fdo#38242: Make autocorrect of 1st to 1 + superscripted st work againCaolán McNamara1-1/+1
It surely doesn't matter if the result of normalization is unchanged (cherry picked from commit defe079d455ccc958fd0128e8a8cf0e4aeb5cd9c) Change-Id: I87dfd0dadee33897f5df4b0dc532166a8bd9d7e5 Signed-off-by: Michael Stahl <mstahl@redhat.com>
2012-09-06Headers/Footers: don't create temporary parts of document (perf problems)Cédric Bosdonnat3-55/+33
Clicking on an empty header/footer only displays the separator as creating empty header/footer could cause important layout recomputing on long documents. Change-Id: I800c5dd27744772833dee0da4269e7591bd0a10b Signed-off-by: Petr Mladek <pmladek@suse.cz> Signed-off-by: Michael Stahl <mstahl@redhat.com> Signed-off-by: Rainer Bielefeld <LibreOffice@bielefeldundbuss.de> Signed-off-by: Roman Eisele <bugs@eikota.de>
2012-09-06fdo#46141: Don't use a timer to show header/footer separatorsCédric Bosdonnat5-35/+91
Clicking on the header/footer area will show the separator. Clicking on an empty header/footer area will add a temporary one that we try hard to remove when clicking outside Change-Id: I8f29f8fd80b2d808257636bf94ec4e2e46f6028e Signed-off-by: Petr Mladek <pmladek@suse.cz> Signed-off-by: Michael Stahl <mstahl@redhat.com> Signed-off-by: Rainer Bielefeld <LibreOffice@bielefeldundbuss.de> Signed-off-by: Roman Eisele <bugs@eikota.de>
2012-09-05better import of conditional format from old ODF structureMarkus Mohrhard6-183/+155
The old ODF storage is style based and so the sam cond format can be divided up into several single stlyes which resulted in several new style cond formats. Now we check for old stlye cond formats if there is a equal cond format and in this case just extend the area. This should make it easier to transform old documents into the new range based cond formats. Conflicts: sc/inc/conditio.hxx sc/source/filter/xml/xmlstyli.cxx sc/source/filter/xml/xmlstyli.hxx Change-Id: I51a5148922e19e6860de9915abfc59d49b18d96e Signed-off-by: Kohei Yoshida <kohei.yoshida@gmail.com>
2012-09-05HelpIndexer, HelpLinker are build-time toolsStephan Bergmann1-2/+2
Change-Id: I43186271f36f67a95a7e3766e1998ea42531e596 (cherry picked from commit beab021ffe57cf6004971d4f285984b6debe5b5f) Reviewed-on: https://gerrit.libreoffice.org/562 Reviewed-by: Michael Stahl <mstahl@redhat.com> Tested-by: Michael Stahl <mstahl@redhat.com>
2012-09-05fdo#53009: Compile extension help in gbuildStephan Bergmann5-52/+141
...as had been done in the old build system (solenv/inc/extension_helplink.mk). Especially for bundled extensions, this removes the need to compile the help data per user on first start. gb_Extension_add_helpfile(s) replaces gb_Extension_localize_help, and takes care of all the steps (localization, compilation, inclusion in .oxt), even for the en-US data (which was handled with additional gb_Extension_add_file calls before). (cherry picked from commit b23bb8e0de3dbdc2c66c3dedf70dfd318868f76c with additional fixes from follow-up commits 152ffaa15f963f92b1564946a648c53d6b5c808c "Call HelpIndexer/Linker with gb_Helper_set_ld_path" and a297372210396260da57f34da3790f76682603cc "Quote .ddf content (potentially containing stuff like '%2F')") Conflicts: solenv/gbuild/Extension.mk solenv/gbuild/ExtensionTarget.mk Change-Id: Ie4bab66d3cad2b713780a23bf2606ca56cfff37f Reviewed-on: https://gerrit.libreoffice.org/563 Reviewed-by: Michael Stahl <mstahl@redhat.com> Tested-by: Michael Stahl <mstahl@redhat.com>
2012-09-05fdo#54546: Force version 2.1 of the sRGB profile for PDF/A-1aFridrich Štrba1-0/+2
(cherry picked from commit 24f691867d02e853153a53e49276c2a8c30ea1fe) Change-Id: I7c40c37fbe344f1e46ea4a09fb99a5ac82ffd577 Signed-off-by: Lionel Elie Mamane <lionel@mamane.lu>
2012-09-05idlc: clear include file set in Idlc::reset():Michael Stahl1-0/+2
Resetting the set between files reduces the size of the generated offapi.d from 41M to 4.5M. Change-Id: I221e6dfb75cbadb5d970f18eccfc85ffdb83ce6c (cherry picked from commit 39c3a4d6644ae78783aa8877557e4c021cba7973) Reviewed-on: https://gerrit.libreoffice.org/566 Reviewed-by: Stephan Bergmann <sbergman@redhat.com> Tested-by: Stephan Bergmann <sbergman@redhat.com>
2012-09-05fdo#53872: reportdesign: fix section drawpage crash:Michael Stahl3-8/+23
In order to wrap the SdrPage's UNO object completely, set its mxUnoPage member to the OSection wrapper instance in OSection::init; only OSection should have access to it. Also initialize m_xDrawPage_Tunnel (thanks Lionel for the hint). (regression from 05218c101df486302bf4cfe8be23ad840daa3f73) Change-Id: I048ddafc31e946853e56e6a403ddc9487cfbcf0e Signed-off-by: Lionel Elie Mamane <lionel@mamane.lu>
2012-08-31Don't export the internal field name suffix '*' to ods.Kohei Yoshida1-2/+5
These '*' suffix is used internally to make duplicate field names unique. However, we were exporting the raw names (names with the '*'s) by mistake when saving to ods. We need to stop this ASAP before someone gets hurt. Change-Id: Iaff7a6343cf202f5fb9855fb7dfe59b872920e4f (cherry picked from commit fe0ef1b2f13d1c2cc8e4f777911c158390183053) Signed-off-by: Eike Rathke <erack@redhat.com>
2012-08-31fdo#54303: Prevent crash on load.Kohei Yoshida1-1/+5
Another simple null pointer check, to save the day. This unfortunately prevents the properties of the skipped field from being loaded properly, but it's still better than a immediate crash. Change-Id: I49398d80eb3bf0834b7e9830ac8be3ac776cb354 Signed-off-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
2012-08-31handle missing cond format entries more gracefully, related fdo#52340Markus Mohrhard1-0/+6
Change-Id: Id1cf7d99a13c04541e87ad00c5418dd4f766d268 Signed-off-by: Kohei Yoshida <kohei.yoshida@gmail.com>
2012-08-30fdo#53938: Check for null pointer, and don't proceed when error occurs.Kohei Yoshida3-5/+12
This prevents the crash, the same error dialog from appearing 3 times in a row, and also prevents the pivot table outout from becoming empty on error condition. Change-Id: I09a72f78561f3f0446a95732fb9c242c1144878a Signed-off-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
2012-08-30Update desc of mergelib configure optionTomas Chvatal1-4/+13
Also do not be fatal on clash with binfilter and just warn the user that mergelib was disabled in favor of binfilter if he has both enabled. Change-Id: I3b683a948e00b1134c10881c5ec442e6380f2822 Signed-off-by: Matúš Kukan <matus.kukan@gmail.com>
2012-08-30also handle the cond format code for xlsb the new way, fdo#53424Markus Mohrhard1-0/+1
Change-Id: I8d54e5fee29952635c3d4c2d5a12203668e2ff28
2012-08-30fdo#53887 convert filter value as LIKE pattern only on (NOT) LIKELionel Elie Mamane1-2/+6
Change-Id: I24d1f3af5d428e735a94fc9a69960244d0225513 Reviewed-on: https://gerrit.libreoffice.org/513 Reviewed-by: Björn Michaelsen <bjoern.michaelsen@canonical.com> Reviewed-by: Andras Timar <atimar@suse.com> Tested-by: Andras Timar <atimar@suse.com>
2012-08-30fdo#38791: Do not use java.awt, causes problems on Mac OS XStephan Bergmann3-23/+24
On Mac OS X, it apparently suffices to execute "new java.awt.Dimension()" to run into "Apple AWT Java VM was loaded on first thread -- can't start AWT" (see <http://developer.apple.com/library/mac/#technotes/tn2005/tn2147.html> "JNI Development on Mac OS X - Thread-Safe JNI Programming - Calling AWT/Swing From AppKit"). The solution here is simple, in that uses of java.awt.Dimension can be replaced with com.sun.star.awt.Size without loss of functionality. However, there are still occurrences of java.awt.Image and java.awt.Toolkit lurking (reportbuilder/java/com/sun/star/report/SOImageService.java, reportbuilder/java/com/sun/star/report/pentaho/output/ImageProducer.java, reportbuilder/java/com/sun/star/report/pentaho/output/OfficeDocumentReportTarget.java) for which it might be sheer luck that they do not seem to cause trouble yet. Change-Id: I33e9f74c50ebddc11bd1d9a48c55dc8f8700345d Signed-off-by: Lionel Elie Mamane <lionel@mamane.lu>
2012-08-29fdo#52052 fix RTF import of page breaks on landscape pagesMiklos Vajna2-6/+48
The problem was that we tried to insert a page break before reaching the first section break, where section properties are sent. Additionally, the continuous section break at the end of the doc caused trouble, so ignore it explicitly. (cherry picked from commit 1efa576ef88141c4deb5da9818537e053dc6517b) Change-Id: I22bc355994991beeadb41d26b44ce3e2beedbdb2 Reviewed-on: https://gerrit.libreoffice.org/497 Reviewed-by: Norbert Thiebaud <nthiebaud@gmail.com> Tested-by: Norbert Thiebaud <nthiebaud@gmail.com>
2012-08-29fdo#53968 etc.: Fix more "doesn't start after upgrade" problemsStephan Bergmann13-121/+179
* fdo#53968 revealed that multiple soffice.bin instances can run removeTree in parallel. Therefore, demoted failures from exceptions to SAL_WARNs. (And keeping fingers crossed.) * a8cdce148c76c93c5d41820610d6e6ac175e03a7 "fdo#53655: Ignore failure to remove directories (as happens on Windows XP)" was due to a forgotten osl::Directory::close before calling osl::Directory::remove after all. * UserInstallations have been seen in the wild where no extensions were installed per-user (any longer), but user/uno_packages/cache/registry/ com.sun.star.comp.deployment.component.PackageRegistryBackend/*.rdb files contained data nevertheless. To reliably clean out any old junk, refreshBundledExtensionsDir has been extended to cleanExtensionsCache which in tandem with an extended Desktop::SynchronizeExtensionRepositories now cleanly re-installs all bundled, shared, and per-user extensions after a LO upgrade. Change-Id: Ic6b5b6c1945d76eb3a65b6cd4512a657b7a835a0 Reviewed-on: https://gerrit.libreoffice.org/477 Reviewed-by: Norbert Thiebaud <nthiebaud@gmail.com> Tested-by: Norbert Thiebaud <nthiebaud@gmail.com>