summaryrefslogtreecommitdiff
path: root/svl
AgeCommit message (Collapse)Author
2022-03-11new loplugin:trivialdestructorNoel Grandin
look for potentially trivial destructors that can then be elided Change-Id: I435c251bd4291b5864c20d68f88676faac7c43fb Reviewed-on: https://gerrit.libreoffice.org/c/core/+/131318 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
2022-03-07do not pass XComponentContext to officecfg::...::get() callsLuboš Luňák
It's used only for the ConfigurationWrapper singleton, so it's used only the first time and then ignored. It also causes calls to comphelper::getProcessComponentContext() for every single invocation despite the value not being needed, and the calls may not be cheap (it's ~5% CPU during ODS save because relatively frequent calls to officecfg::Office::Common::Save::ODF::DefaultVersion::get()). Change-Id: I02c17a1a9cb498aeef220ddd5a0bde5523cb0ffb Reviewed-on: https://gerrit.libreoffice.org/c/core/+/131056 Tested-by: Jenkins Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
2022-03-06make a bunch of SvtBroadcaster/SvtListeners functions inlineLuboš Luňák
These classes are used extensively during some operations in Calc, and not even LTO can inline these when they end up in different binaries. Change-Id: I5b17a004c6cc039508b76d557d58714c83f237a5 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/131071 Tested-by: Jenkins Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
2022-03-06do not destroy broadcasters and then recreate again (tdf#134268)Luboš Luňák
Sorting ends tells all listeners on all the sorted cells to end listening to stop updates, then sorts the cells and starts listening again. This will cause all broadcasters for the sorted cells to temporarily stop having any listeners, so they'll be deleted and removed from the mdds vector (which may additionally cause moving large parts of the mdds vector repeatedly). And since all listeners will want to listen again after the sort, this will all need to be reconstructed. To avoid this, temporarily block this removal and then later just checks and remove any possibly left-over broadcasters that ended up with no listeners. Change-Id: Ie2d41d9acd7b657cf31a445870ce7f18d28d5ebb Reviewed-on: https://gerrit.libreoffice.org/c/core/+/131069 Tested-by: Jenkins Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
2022-03-03Recheck modules sv* with IWYUGabor Kelemen
See tdf#42949 for motivation Change-Id: I25779cbfb1aa93c31d6e12ac95e136b3bdbbc058 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/130403 Tested-by: Jenkins Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
2022-02-27use more SfxItemSet::CloneAsValueNoel Grandin
to reduce heap allocations Change-Id: Ia755c3e7f9610a5441a447cc74ea38ebcef068bc Reviewed-on: https://gerrit.libreoffice.org/c/core/+/130066 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
2022-02-22merge poolio.cxx into itempool.cxxLuboš Luňák
It doesn't make sense to have few SfxItemPool functions in a separate source file, and without LTO those tiny functions do not get inlined, which is confusing when profiling. Change-Id: I0ba891ba51c7d8dd3eba3f4cae0f7819f9b033e7 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/130327 Tested-by: Jenkins Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
2022-02-21replace SfxPoolItem::LookupHashCode() with Lookup() (tdf#135215)Luboš Luňák
The LookupHashCode() way still loops over an array while calling virtual functions for a type that is actually known. Overriding instead a whole new Lookup() function that does the loop avoids the virtual calls, allowing compiler optimizations such as inlining, or class-specific optimizations like the hash code. So this still improves performance with ScPatternAttr a bit, for tdf#135215 it saves about 40% of load time, but this is scraping the bottom of the barrel, as this is still a linear search (the IsSortable() approach, if it worked, would be still 3x faster). Change-Id: I8fe5f70cabb77e2f6619d169beee8a3b4da46213 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/130228 Tested-by: Jenkins Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
2022-02-20fix usage of std::lower_bound() in SfxItemPool (tdf#81765)Luboš Luňák
The function expects elements starting from smallest to the biggest, and finds the first one that is not smaller than the one searched for. That means that all items remaining will not be smaller, and thus end of search is when items compare larger. Comparing remaining items as smaller means searching until the end. Change-Id: If5cf5c18951abf987ddbbf201f49cfb195e36d32 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/130220 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk> Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
2022-02-18avoid an allocation in WhichRangesContainer::MergeRangeNoel Grandin
speeds up loading a large chart by 5% Change-Id: Idd8566012a0049d429e38b589782fc6d76eb3a5a Reviewed-on: https://gerrit.libreoffice.org/c/core/+/130132 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
2022-02-18improve the SfxItemSet::CloneAsValue checkNoel Grandin
to prevent object slicing. Which reveals a problems with commit 044fa30a4c77013c87a7e2a6dd9022a2f6599778 Author: Noel Grandin <noelgrandin@gmail.com> Date: Thu Sep 23 18:44:42 2021 +0200 no need to allocate this SfxItemSet on the heap and commit 044fa30a4c77013c87a7e2a6dd9022a2f6599778 Author: Noel Grandin <noelgrandin@gmail.com> Date: Thu Sep 23 18:44:42 2021 +0200 no need to allocate this SfxItemSet on the heap so revert the problematic bits of those commits Change-Id: I5eeba1d5bdb91f4e539850516f2b1c11e69ff2c1 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/130127 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
2022-02-16Extend loplugin:stringview to OUStringBuffer::copyStephan Bergmann
(Somewhat oddly, there is no OStringBuffer::copy counterpart.) This required some modification to StringView::VisitCXXConstructExpr to avoid > In file included from odk/qa/checkapi/checkapi.cxx:29: > In file included from workdir/CustomTarget/odk/allheaders/allheaders.hxx:351: > In file included from instdir/sdk/include/rtl/math.hxx:31: > instdir/sdk/include/rtl/ustrbuf.hxx:1687:16: error: rather than copy, pass with a view using subView() [loplugin:stringview] > return copy( beginIndex, getLength() - beginIndex ); > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ while building CppunitTest_odk_checkapi as external C++03 code, where the returned value is wrapped in a CXXConstructExpr. And testing for that case required a new CompilerTest_compilerplugins_clang-c++03 that uses gb_CXX03FLAGS and needs to not set LIBO_INTERNAL_ONLY (via gb_CompilerTest_set_external_code), as compiling as C++03 would otherwise generate lots of errors like unknown char16_t at include/sal/types.h:118. (There was a choice whether to name the new test "-c++03" or "-external", but the issue it tests is caused more by the code being compiled with C++03 than by this being external code, see above.) Change-Id: I873a9c5a70d3ea949cf13a169d46920b71282712 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/130036 Tested-by: Jenkins Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2022-02-08speed up SfxItemPool searches with items that can't use IsSortable()Luboš Luňák
As pointed out by 585e0ac43b9bd8a2f714903034e435c84ae3fc96, some item types cannot be used as IsSortable(), because they are modified after having been inserted in the pool, which breaks the sorted order. But it's possible to at least somewhat improve performance of these items by explicitly providing a hash code and using that first for comparisons when looking up items, which may be cheaper than calling operator==. With ScPatternAttr such comparisons seem to take only 60% of the original time, reducing loading time of some documents by 25%. Change-Id: I41f4dda472fb6db066742976672f2c08b9aeef63 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/129667 Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk> Reviewed-by: Luboš Luňák <l.lunak@collabora.com> Tested-by: Jenkins
2022-02-07tdf#145868 sd replace: if search changes, restart find/replaceJustin Luth
REPLACE is really a replaceAndFind instead of a findAndReplace. Thus, when you changed your search parameter and did a replace, it replaced the previously searched for item, and then found the first instance of the new search parameter. That of course is just wrong. So make sure to verify that the previous search matches the current search competely. However, that doesn't mean that the entire searchItem matches, since we don't want to restart the search just because the replace parameter changed. In my testing, this wasn't an issue for REPLACE_ALL. So the only time we need to worry about the last search result is in a replace once situation. P.S. This commit exposed that mpSearchItem can point to a destructed SvxSearchItem, so this patches unit test will crash if the other 7.4 commit is missing. Change-Id: I7be14d64534018718145c6ac5f8629ff5f2e5611 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/129385 Tested-by: Jenkins Reviewed-by: Justin Luth <jluth@mail.com> Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
2022-02-06No need to call makeStringAndClear to pass a string viewMike Kaganski
Change-Id: Id9e18e66f3b3c583d7cb22ee0a5d6272ca2f7ea9 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/129556 Tested-by: Jenkins Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
2022-02-06Simplify code a bitMike Kaganski
Change-Id: I6f1df008e2b51ff42d058f2c4f37393066798c93 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/129555 Tested-by: Jenkins Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
2022-02-03no point in mapping SID to itselfNoel Grandin
and once we remove that self mapping, no need to call GetWhich for those IDs either Change-Id: Ia881328a29bb022dace8d5f25c57279a381e0377 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/129321 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
2022-02-03SfxMetricItem should take signed valueNoel Grandin
since it extends SfxInt32Item been this way since commit 5f51e579fc2e3207166b053382ca14b95082728c Date: Wed Apr 11 18:28:13 2007 +0000 INTEGRATION: CWS hedaburemove01 (1.1.2); FILE ADDED 2007/02/09 16:12:44 vg 1.1.2.1: #72503# get rid of hedabu procedure: Moving headers to svtools/inc/svtools and correspondent necessary changes Change-Id: Id23909c5cb842b858baf2f6918298607a79adcec Reviewed-on: https://gerrit.libreoffice.org/c/core/+/129402 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
2022-02-02convert OfaRefItem to a normal classNoel Grandin
only instantiated with one type, so just turn it into a normal class Change-Id: If3ae908f3e226ae9f4d3b81a7a7d9ba492ccda4f Reviewed-on: https://gerrit.libreoffice.org/c/core/+/129283 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
2022-01-29fix loading file with very large number of stylesNoel Grandin
where nCurrentPosition was wrapping around Change-Id: I839215f3138d58884f03509b6341147ef28edb4c Reviewed-on: https://gerrit.libreoffice.org/c/core/+/129108 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
2022-01-21fix opengl slide transitionNoel Grandin
regression from commit 62efb188668a3296591dcfa3658185e2f982e356 use comphelper::WeakComponentImplHelper in SlideShowVie Switch notifyEach to leave the mutex locked after being called, because that seems to compose better - because after a call it is in the same state that was when we entered the call. Change-Id: I42e80cc7be1b65ed8cab24ab7c11210e056d916d Reviewed-on: https://gerrit.libreoffice.org/c/core/+/128723 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
2022-01-14make comphelper::OInterfaceContainerHelper4 more threadsafeNoel Grandin
(*) make all the methods that require an external mutex take a std::unique_lock as a parameter, so that call sites cannot forget (*) make the forEach method drop the lock when firing listener methods, to reduce the odds of deadlock Change-Id: I0a80e3b3d1c1c03b7de4a658d31fcc2847690903 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/128415 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
2022-01-02Introduce OUString::unacquired(const OUStringBuffer&)Mike Kaganski
... and avoid OUStringBuffer::toString when the temporary is used for checking current buffer content Change-Id: I114178f3e74ca3e4a3e517763f9eaab4797b7deb Reviewed-on: https://gerrit.libreoffice.org/c/core/+/127478 Tested-by: Jenkins Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
2021-12-26Avoid OUStringBuffer::toString where possibleMike Kaganski
Change-Id: I99abbe97a48b2077e28d6221fb70036e5e412657 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/127479 Tested-by: Jenkins Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
2021-12-16Fix typosAndrea Gelmini
Change-Id: I66feced8bed05c7859e36a6d2f746a7faf30c7a4 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/126915 Reviewed-by: Julien Nabet <serval2412@yahoo.fr> Tested-by: Jenkins
2021-12-15tdf#145381 handle closing brackets in URLs correctlyArmin Le Grand (Allotropia)
The task presents an URL that ends with a closing bracket. If pasted to LO, the closing bracket got interpreted as not being part of the URL due to the heuristical interpretation of URLs in urihelper. Adapted this to handle matching brackets, so that an closing and ending bracket will be added to the uri text when there is a matching pair. Added unit test to testFindFirstURLInText with simplified uri example. Change-Id: I58dd460a37d0066ff46845832eabd2a790e4ccd1 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/126832 Tested-by: Jenkins Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2021-12-10Generally determine Rdb content from gb_*_set_componentfile callsStephan Bergmann
...instead of by listing the content somewhat redundantly in the Rdb_*.mk files, to avoid duplication of logic for components that are only built conditionally (and thus should only be included conditionally in the corresponding Rdb). To achieve that, add an "rdb" parameter to gb_ComponentTarget_ComponentTarget (and to the gb_*_set_componentfile macros that internally call gb_ComponentTarget_ComponentTarget), which is used to make the appropriate gb_Rdb_add_component call internally from within gb_ComponentTarget_ComponentTarget. (As a special case, gb_CppunitTest_set_componentfile shall not call gb_Rdb_add_component, as that has already been done by the corresponding gb_Library_set_componentfile call, so allow the gb_ComponentTarget_ComponentTarget "rdb" parameter to be empty to support that special case.) Most Rdb_*.mk files are thus mostly empty now. One exception is i18npool/Rdb_saxparser.mk, which duplicates some of the Rdb_services content as needed during the build in CustomTarget_i18npool/localedata. 1c9a40299d328c78c035ca63ccdf22c5c669a03b "gbuild: create services.rdb from built components" had already tried to do something similar (in addition to other things) under a new --enable-services-rdb-from-build option. However, that approach had four drawbacks that this approach here addresses (and which thus partly reverts 1c9a40299d328c78c035ca63ccdf22c5c669a03b): 1 Rdb_services shall not contain the component files of all libraries that are built. While that commit filtered out the component files that go into Rdb_ure/services (ure/Rdb_ure.mk), it failed to filter out the component files that go into others like Rdb_postgresql-sdbc (connectivity/Rdb_postgresql-sdbc.mk). 2 The code added by that commit to Makefile.gbuild codified the knowledge that there is an Rdb_services, which is brittle. 3 The code added by that commit to solenv/gbuild/Rdb.mk codified the knowledge (for gb_Rdb__URECOMPONENTS) that there is an Rdb_ure/services, which is brittle. 4 Introducing an --enable-services-rdb-from-build option needlessly provided two different ways how the content of Rdb_services is assembled. The changes done here would leave --enable-services-rdb-from-build as a misnomer, as it no longer controls how Rdb_services is assembled. I thus renamed it to --enable-customtarget-components, as that is apparently what it still does now. Change-Id: Ia5e8df4b640146c77421fcec6daa11a9cd260265 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/126577 Tested-by: Jenkins Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2021-12-06make it clearer that ScHint is modifiedLuboš Luňák
It's somewhat confusing that an accessor is provided to give a reference to internal data and then the object is modified indirectly using the reference. It appears to be only for performance reasons, so I thought that inlining the ctor and ctor could help the compiler to optimize this, but apparently it can't move this outside of the loop, so at least make it clearer. Change-Id: I72cf15d1446daa559ac4079b9478e53694d7d198 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/126394 Tested-by: Jenkins Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
2021-11-28use more OInterfaceContainerHelper3 in svlNoel Grandin
Change-Id: I93443f87fe9b68157cc9a126f2c31f917a016a1c Reviewed-on: https://gerrit.libreoffice.org/c/core/+/125975 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
2021-11-23make a bunch of svl::SharedString functions inlineLuboš Luňák
They are tiny and they are used in performance-critical parts of Calc. Change-Id: If227b11ac7929dd1369545a590d8ef1a977185f4 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/125698 Tested-by: Jenkins Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
2021-11-20tdf#133835 speedup calc autofilter (1)Noel Grandin
avoid a copy in INetURLHistory::QueryUrl, saves 10% Change-Id: I662a4e48f198e61f1cef52b1e920613bce9c9766 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/125563 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
2021-11-19loplugin:stringliteraldefine in svlNoel Grandin
Change-Id: I49a8062c30bdce5b0d9bd27e421d85ba61d5592e Reviewed-on: https://gerrit.libreoffice.org/c/core/+/125544 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
2021-11-18tdf#140086 svl: restore PasswordContainer to single-instanceMichael Stahl
(regression from 7256ff08bc46840bb85fa255ace6541dca91329e) Change-Id: Ib640dea001fc787279761ca72bbc3db46d0102c2 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/125485 Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk> Reviewed-by: Michael Stahl <michael.stahl@allotropia.de> Tested-by: Jenkins
2021-11-18rtl::Static->thread-safe static in svlNoel Grandin
Change-Id: Idadc4dc77eb681a8b8923ffacf37ddbe1d8245e0 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/125425 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
2021-11-15svl: use std::rotate() in SfxUndoManager::ImplUndo()Miklos Vajna
Instead of manually moving out, moving a range and then moving in. Change-Id: Iaff461e1fcee3936c8ddc02bf471a804e7881aef Reviewed-on: https://gerrit.libreoffice.org/c/core/+/125219 Reviewed-by: Miklos Vajna <vmiklos@collabora.com> Tested-by: Jenkins
2021-11-12sw, out of order undo: allow multiple actions from other viewsMiklos Vajna
Previously we assumed that the action to be executed is always exactly the top of the undo stack minus 1 element. Extend this, so that in case an other view appends two or more elements to the undo stack, we still find our undo action. Obviously only do this if all those undo actions are independent from us. This requires replacing the swap in svl/ with a move-out + move a range down + move in construct. Change-Id: Ic12d32d6eb5e77618d99eddb4fa096802f32d655 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/125076 Reviewed-by: Miklos Vajna <vmiklos@collabora.com> Tested-by: Jenkins
2021-11-11sw, out of order undo: allow a subset of a non-empty redo listMiklos Vajna
Specifically, we used to not allow out of order undo at all if the redo list was non-empty. This relaxes that condition a bit. Out of order undo is OK with a non-empty redo list, in case all undo actions in the redo list are either 1) owned by the current view or 2) independent from the undo action to be executed I.e. if view1 has lots of type undo actions and an view2 adds a single type undo action on top of it, then allow view 1 to execute multiple of its typing undo actions, not just a single one. Change-Id: I2f5d9404a9994ed74b65233d2a315976c27b28b2 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/125023 Reviewed-by: Miklos Vajna <vmiklos@collabora.com> Tested-by: Jenkins
2021-11-10sw: allow undo of typing in 2 views independent from each otherMiklos Vajna
Undoing out of order is dangerous by default, so limit this to a very specific case as a start, that allows growing in follow-up commits. For now, allow out of order undo if: 1) redo stack is empty 2) we're in LOK mode (different views represent different users) 3) we undo a single action (count is 1) 4) the top undo action doesn't belong to the current view 5) the top and the previous undo actions are independent Which only requires that SwUndoInsert::UndoImpl() is independent for two different paragraphs, which seems to be the case. Independent undo actions opt in for this, currently the only such allowed undo action is SwUndoInsert ("typing"), which adds characters to a single text node. Even those are only considered independent if they operate on different text nodes. On the positive side, this allows out of order undo in the frequent case where two users collaborate on a long document and they just type some new content into the document at different paragraphs. Change-Id: Ibb4551e8f7046b4947491b8bf751eaa0cbb2d060 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/124949 Reviewed-by: Miklos Vajna <vmiklos@collabora.com> Tested-by: Jenkins
2021-11-07Resolves: tdf#145386 Use "Default" for LANGUAGE_PROCESS_OR_USER_DEFAULTEike Rathke
There's no, specifically not in Writer, handling of the LCID 0x0400 LANGUAGE_PROCESS_OR_USER_DEFAULT language/locale concept other than the number formatter mapping it to LANGUAGE_SYSTEM. Use the LANGUAGE_SYSTEM "Default" string in UI (status bar, status menu, language list) but keep the LCID, and don't append the resolved locale string as it is also displayed both in the Font Western and CJK listboxes. This ends up as two list entries, like * Default - English (UK) first entry * Default last entry of which the second would be selected. Change-Id: I8d9e4171bee6bbe9d1c9dcfb7a5fa8fc92ea1a2c Reviewed-on: https://gerrit.libreoffice.org/c/core/+/124449 Reviewed-by: Eike Rathke <erack@redhat.com> Tested-by: Jenkins
2021-10-31Prepare for removal of non-const operator[] from Sequence in svlMike Kaganski
Change-Id: I6b71a075de5d5ac002dc48cd2bb21ff5bf5dd072 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/124395 Tested-by: Jenkins Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
2021-10-29Fix misuses of NULL across Windows-only codeStephan Bergmann
...which defines NULL as a plain 0 integer literal instead of the GNU __null extension, so clang-cl's -Wnull-conversion cannot kick in. These findings are from an experimental build done with clang-cl and a modified > --- a/clang/lib/Headers/stddef.h > +++ b/clang/lib/Headers/stddef.h > @@ -83,6 +83,10 @@ typedef __WCHAR_TYPE__ wchar_t; > # if !defined(__MINGW32__) && !defined(_MSC_VER) > # define NULL __null > # else > -# define NULL 0 > +# if __cplusplus >= 201103L > +# define NULL nullptr > +# else > +# define NULL 0 > +# endif > # endif > #else > # define NULL ((void*)0) However, that build also ran into lots of places where 3rd-party code in external/ and Windows system headers caused issues when NULL is nullptr (which I worked around with various hacky patches for that build), so this is unfortunately not something that can easily be enabled generally. Change-Id: I10674464498a9bc63578d9e6cc32ddde23ab4f30 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/124419 Tested-by: Jenkins Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2021-10-26tdf#131025 swtable: don't apply number format to non-number textJustin Luth
Applying a numbering style to text causes export to save that out as a number (valued as zero). That is not good because the ODF spec says that a number overrides a string. So don't accept a numbering format on non-number text. Why is this change good? -the cell previously had no direct formatting (by definition). -the cell's previous old format was text (tested). -any numbering format applied obviously isn't correct (by definition). -any previous formatting has already been overwritten with numformat. -the default numbering is appropriate for text. -empty cells still get the numbering format (tested). -odd human-designed formats are accepted as intentional (tested). What are the concerns? -the scope of this change is HUGE, way beyond this bug. -on both my dev box and patch box I saw occassional crashes. -the bug was "fixed" by a different import commit that ensured different languages were treated consistently. So this patch is no longer critical, just nice to have to avoid exporting out-of-spec content. Change-Id: Id3dc5f803c3cf4875bc0cab52d1019a18679da77 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/123904 Tested-by: Jenkins Reviewed-by: Justin Luth <justin_luth@sil.org> Reviewed-by: Eike Rathke <erack@redhat.com>
2021-10-25Restore behaviour SvNumberFormatterServiceObj keep LANGUAGE_SYSTEM unresolvedEike Rathke
Semantics were temporarily changed with commit bba6a4ed92a0feb288a9dedd648d623bee02a3ce CommitDate: Mon Oct 25 11:18:31 2021 +0200 Get rid of fuzziness in MsLangId::Conversion::convertIsoNamesToLanguage() Change-Id: I8bbf77998bea81c5691ba518c7ae25093b0df421 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/124141 Reviewed-by: Eike Rathke <erack@redhat.com> Tested-by: Jenkins
2021-10-25Get rid of fuzziness in MsLangId::Conversion::convertIsoNamesToLanguage()Eike Rathke
With on-the-fly LanguageTag the fuzzy fallbacks for a few languages aren't needed anymore. Proper fallbacks should be obtained when needed, like when configuring default locales, or configured values used after. Change-Id: I8b85b8099e085508435036ac846db6c3e516dc23 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/124127 Reviewed-by: Eike Rathke <erack@redhat.com> Tested-by: Jenkins
2021-10-25Fix stack-use-after-scopeStephan Bergmann
...during CppunitTest_svl_qa_cppunit, after df42cb6552b20372f62b5a361709670db80e4ed4 "Optimize assignment from OUStringLiteral to OUString", > ==918==ERROR: AddressSanitizer: stack-use-after-scope on address 0x2b2e203b5900 at pc 0x2b2e1b9004f8 bp 0x7ffc06726270 sp 0x7ffc06726268 > READ of size 4 at 0x2b2e203b5900 thread T0 > #0 0x2b2e1b9004f7 in void rtl::str::release<_rtl_uString>(_rtl_uString*) /sal/rtl/strtmpl.hxx:1064:9 > #1 0x2b2e1b8d916c in rtl_uString_release /sal/rtl/ustring.cxx:1785:12 > #2 0x2b2e36263ec5 in rtl::OUString::~OUString() /include/rtl/ustring.hxx:493:9 > #3 0x2b2e3622c5ff in (anonymous namespace)::Test::testTdf103060() /svl/qa/unit/svl.cxx:553:1 [...] > [2304, 2320) 'EXPECTED_G3' (line 550) <== Memory access at offset 2304 is inside this variable (<https://ci.libreoffice.org/job/lo_ubsan/2176/>). (aa2064c5c5f23f6f4b7bc44e12345b37f66995bc "Improve loplugin:stringliteralvar" had failed to introduce those OUStringLiteral variables as static.) Change-Id: I59168979fcc4b055d17d1d4f315577eef1027505 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/124134 Tested-by: Jenkins Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2021-10-16Simplify vector initialization in svlJulien Nabet
Change-Id: I8ff1eb008f3173791c7c1020db08d29451998f42 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/123699 Tested-by: Jenkins Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
2021-10-15Remove non-const Sequence::begin()/end() in internal codeMike Kaganski
... to avoid hidden cost of multiple COW checks, because they call getArray() internally. This obsoletes [loplugin:sequenceloop]. Also rename toNonConstRange to asNonConstRange, to reflect that the result is a view of the sequence, not an independent object. TODO: also drop non-const operator[], but introduce operator[] in SequenceRange. Change-Id: Idd5fd7a3400fe65274d2a6343025e2ef8911635d Reviewed-on: https://gerrit.libreoffice.org/c/core/+/123518 Tested-by: Jenkins Reviewed-by: Stephan Bergmann <sbergman@redhat.com> Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
2021-10-14Avoid COW overhead using css::uno::SequenceMike Kaganski
The scenarios are: 1. Calling sequence's begin() and end() in pairs to pass to algorithms (both calls use getArray(), which does the COW checks) 2. In addition to #1, calling end() again when checking result of find algorithms, and/or begin() to calculate result's distance 3. Using non-const sequences in range-based for loops, which internally do #1 4. Assigning sequence to another sequence variable, and then modifying one of them In many cases, the sequences could be made const, or treated as const for the purposes of the algorithms (using std::as_const, std::cbegin, and std::cend). Where algorithm modifies the sequence, it was changed to only call getArray() once. For that, css::uno::toNonConstRange was introduced, which returns a struct (sublclass of std::pair) with two iterators [begin, end], that are calculated using one call to begin() and one call to getLength(). To handle #4, css::uno::Sequence::swap was introduced, that swaps the internal pointer to uno_Sequence. So when a local Sequence variable should be assigned to another variable, and the latter will be modified further, it's now possible to use swap instead, so the two sequences are kept independent. The modified places were found by temporarily removing non-const end(). Change-Id: I8fe2787f200eecb70744e8b77fbdf7a49653f628 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/123542 Tested-by: Jenkins Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
2021-10-12Revert "Use placement new to avoid one of the allocation calls..."Stephan Bergmann
This reverts commit 503ab1ca9ae11978d9717557546c01ff598aaf88, plus follow-up 17915ab5202a4d7456e9bc031c3f6a72bc861844 "fix ubsan alloc-dealloc-mismatch". It failed to properly destroy the object assembly, and caused e.g. CppunitTest_svl_items to fail with > ==850754==ERROR: AddressSanitizer: new-delete-type-mismatch on 0x6060000024e0 in thread T0: > object passed to delete has wrong type: > size of the allocated type: 64 bytes; > size of the deallocated type: 56 bytes. > #0 in operator delete(void*, unsigned long) at /home/sbergman/github.com/llvm/llvm-project/compiler-rt/lib/asan/asan_new_delete.cpp:164:3 (workdir/LinkTarget/Executable/cppunittester +0x330ae2) > #1 in SfxItemSet::~SfxItemSet() at svl/source/items/itemset.cxx:202:1 (instdir/program/libsvllo.so +0x110ccf6) > #2 in std::default_delete<SfxItemSet>::operator()(SfxItemSet*) const at /home/sbergman/gcc/trunk/inst/lib/gcc/x86_64-pc-linux-gnu/12.0.0/../../../../include/c++/12.0.0/bits/unique_ptr.h:85:2 (instdir/program/libsvllo.so +0x1142a28) > #3 in std::_Sp_counted_deleter<SfxItemSet*, std::default_delete<SfxItemSet>, std::allocator<void>, (__gnu_cxx::_Lock_policy)2>::_M_dispose() at /home/sbergman/gcc/trunk/inst/lib/gcc/x86_64-pc-linux-gnu/12.0.0/../../../../include/c++/12.0.0/bits/shared_ptr_base.h:442:9 (instdir/program/libsvllo.so +0x12696e4) > #4 in std::_Sp_counted_base<(__gnu_cxx::_Lock_policy)2>::_M_release() at /home/sbergman/gcc/trunk/inst/lib/gcc/x86_64-pc-linux-gnu/12.0.0/../../../../include/c++/12.0.0/bits/shared_ptr_base.h:168:6 (instdir/program/libsvllo.so +0xe500b5) [...] > 0x6060000024e0 is located 0 bytes inside of 64-byte region [0x6060000024e0,0x606000002520) > allocated by thread T0 here: > #0 in operator new(unsigned long) at /home/sbergman/github.com/llvm/llvm-project/compiler-rt/lib/asan/asan_new_delete.cpp:95:3 (workdir/LinkTarget/Executable/cppunittester +0x32fe7d) > #1 in SfxItemSet::Clone(bool, SfxItemPool*) const at svl/source/items/itemset.cxx:1270:34 (instdir/program/libsvllo.so +0x1127854) > #2 in (anonymous namespace)::Node::setItemSet(SfxItemSet const&) at svl/source/items/stylepool.cxx:65:107 (instdir/program/libsvllo.so +0x1212179) > #3 in StylePoolImpl::insertItemSet(SfxItemSet const&, rtl::OUString const*) at svl/source/items/stylepool.cxx:417:19 (instdir/program/libsvllo.so +0x12103e1) > #4 in StylePool::insertItemSet(SfxItemSet const&, rtl::OUString const*) at svl/source/items/stylepool.cxx:456:17 (instdir/program/libsvllo.so +0x1212ffb) [...] in Clang ASan builds done with -fsized-deallocation. Change-Id: I3ccba7e7d9712ecabf38a0149252d3cd70cdb446 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/123446 Tested-by: Jenkins Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2021-10-11In O[U]StringBuffer, make string_view params replacements for OUString onesStephan Bergmann
...for LIBO_INTERNAL_ONLY, instead of having them as additional overloads. That way, loplugin:bufferadd and loplugin:stringviewparam found many further opportunities for simplification (all addressed here). Some notes: * There is no longer an implicit conversion from O[U]String to O[U]StringBuffer (as that goes via user-defined conversions through string_view now), which was most noticeable in copy initializations like OStringBuffer buf = someStr; that had to be changed to direct initialization, OStringBuffer buf(someStr); But then again, it wasn't too many places that were affected and I think we can live with that. * I made the O[U]StringBuffer ctors taking string_view non-explicit, mainly to get them in line with their counterparts taking O[U]String. * I added an OUStringBuffer::lastIndexOf string_view overload that was missing (relative to OUStringBuffer::indexOf). * loplugin:stringconstant needed some addition to keep the compilerplugins/clang/test/stringconstant.cxx checks related to OStringBuffer::append and OStringBuffer::insert working. * loplugin:stringviewparam no longer needs the special O[U]StringBuffer-related code that had been introduced in 1250aecd71fabde4dba990bfceb61bbe8e06b8ea "loplugin:stringviewparam extend to new.." Change-Id: Ib1bb8c4632d99b744e742605a9fef6eae959fd72 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/122904 Tested-by: Jenkins Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
generated by cgit (git 2.34.1) at 2024-07-15 14:42:16 +0000