summaryrefslogtreecommitdiff
path: root/sw
AgeCommit message (Collapse)AuthorFilesLines
2021-04-20tdf#141525 Fix "Line with Arrow/Circle" insertion error.Jeff Huang1-0/+8
When you insert "Line with Arrow/Circle" from "Lines and Arrows", it will show "Line Starts with Arrow" instead of "Line with Arrow/Circle". This patch can fix this issue. (This issue exists at least since 6.3 series.) Change-Id: I827b9582951fe359d89245970e73d63f7697553f Reviewed-on: https://gerrit.libreoffice.org/c/core/+/113963 Tested-by: Jenkins Reviewed-by: Mark Hung <marklh9@gmail.com> (cherry picked from commit 2b498e380a91a536d177376e854cf3a6e2fb0b35) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/114179 Reviewed-by: Po-Yen Huang <jeff@mail.ossii.com.tw> Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org> (cherry picked from commit 4ee3d213e8d3c4afdc3714b50def117dfe5c8023) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/114275 Reviewed-by: Michael Stahl <michael.stahl@allotropia.de>
2021-03-30tdf#140556 master document view not launching subdocument from context menuCaolán McNamara1-1/+1
Change-Id: I0ae61e51266ae050e340c1af99e0dd6a481e6ff9 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/113282 Tested-by: Jenkins Reviewed-by: Michael Stahl <michael.stahl@allotropia.de>
2021-03-29sw fr_FR AutoCorrect: fix crash using "Apply and Edit changes"Julien Nabet1-1/+1
Bad calling of SvxAutoCorrect::FnAddNonBrkSpace() in French documents, see tdf#140849 for a similar problem. Change-Id: I19b964fb8fe437c7f98342a873a456ffded75d82 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112120 (cherry picked from commit df41967a29c4f8243d1df2d9b23d7c41d4a13819) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112792 Tested-by: Jenkins Reviewed-by: László Németh <nemeth@numbertext.org> Reviewed-by: Caolán McNamara <caolanm@redhat.com>
2021-03-29tdf#122717: fix handling of zero width/height linesXisco Fauli2-0/+13
The code was introduced in 627c2469843c9461b665c4571f1214aca7fc36a4 < tdf#96674 drawingML import: fix handling of zero width/height lines > and later on removed by 36bade04d3780bc54c51b46bb0b63e69789658a5 < tdf106792 Get rid of SvxShapePolyPolygonBezier > with the comment "I doubt that the additional code to make a line not exactly hor/ver is needed. Checked and it is not needed, thus removed the change from tdf#96674 in shape.cxx." it turned out, it's still needed Change-Id: Ib64ee17227e3e588e94381abeabe5a2ff2e0b7d1 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/113102 Tested-by: Jenkins Reviewed-by: Miklos Vajna <vmiklos@collabora.com> (cherry picked from commit 3e4eb070787d4d44b3bdc95046e5b231dbbef42b) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/113119 Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org> (cherry picked from commit 11d344f6271e4171f8006c2cb29372f612cbdc00) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/113123 Tested-by: Xisco Fauli <xiscofauli@libreoffice.org>
2021-03-26tdf#125936 writerfilter: treat escapement in numbering like stylesJustin Luth2-0/+6
deferCharacterProperties wasn't occurring in the numbering import (and so it was affecting the first run of the body text). But just like character styles, it would be better to just consider this auto-superscript instead of to defer it and calculate based on the fontsize - since that really isn't known until layout time, and so only works with direct formating. cherry-picked from 7.2's 2a7617653fb3d31e44e5cbcf8daf4b4c2a24e2c8 Change-Id: I9ce5a31c173089603316f4c3389e5f2e5dbe165a Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112987 Tested-by: Jenkins Reviewed-by: Justin Luth <justin_luth@sil.org> Reviewed-by: Miklos Vajna <vmiklos@collabora.com> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/113076
2021-03-26tdf#141079 Restore double click behavior for script fieldsSamuel Mehrbrodt3-2/+36
Change-Id: I5daa730740ea042bdae56b832f2557ec974339b3 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112668 Tested-by: Jenkins Reviewed-by: Samuel Mehrbrodt <samuel.mehrbrodt@allotropia.de> (cherry picked from commit 5302268de6a33716c7746aa13232746ad2f2b561) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/113072 Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
2021-03-26tdf#140590 Writer crash in Tools>Chapter NumberingNoel Grandin1-1/+1
and an assert in vcl to catch it a little higher up the stack. Change-Id: Iaa2e70f901f93ca9f678118e4f5a8bc1b6eda20a Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112459 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk> Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112866 (cherry picked from commit ebbf550f7087b173bdcf600a02bb87c0b655e787) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/113124
2021-03-19xmlsecurity: improve handling of multiple X509Data elementsMichael Stahl1-1/+2
Combine everything related to a certificate in a new struct X509Data. The CertDigest is not actually written in the X509Data element but in xades:Cert, so try to find the matching entry in XSecController::setX509CertDigest(). There was a confusing interaction with PGP signatures, where ouGpgKeyID was used for import, but export wrote the value from ouCertDigest instead - this needed fixing. The main point of this is enforcing a constraint from xmldsig-core 4.5.4: All certificates appearing in an X509Data element MUST relate to the validation key by either containing it or being part of a certification chain that terminates in a certificate containing the validation key. Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111254 Tested-by: Jenkins Reviewed-by: Michael Stahl <michael.stahl@allotropia.de> (cherry picked from commit 9e82509b09f5fe2eb77bcdb8fd193c71923abb67) xmlsecurity: improve handling of multiple certificates per X509Data It turns out that an X509Data element can contain an arbitrary number of each of its child elements. How exactly certificates of an issuer chain may or should be distributed across multiple X509Data elements isn't terribly obvious. One thing that is clear is that any element that refers to or contains one particular certificate has to be a child of the same X509Data element, although in no particular order, so try to match the 2 such elements that the parser supports in XSecController::setX509Data(). Presumably the only way it makes sense to have multiple signing certificates is if they all contain the same key but are signed by different CAs. This case isn't handled currently; CheckX509Data() will complain there's not a single chain and validation of the certificates will fail. Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111500 Tested-by: Jenkins Reviewed-by: Michael Stahl <michael.stahl@allotropia.de> (cherry picked from commit 5af5ea893bcb8a8eb472ac11133da10e5a604e66) xmlsecurity: add EqualDistinguishedNames() Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111545 Tested-by: Jenkins Reviewed-by: Michael Stahl <michael.stahl@allotropia.de> (cherry picked from commit 1d3da3486d827dd5e7a3bf1c7a533f5aa9860e42) xmlsecurity: avoid exception in DigitalSignaturesDialog::getCertificate() Fallback to PGP if there's no X509 signing certificate because CheckX509Data() failed prevents the dialog from popping up. To avoid confusing the user in this situation, the dialog should show no certificate, which is already the case. Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111664 Tested-by: Jenkins Reviewed-by: Michael Stahl <michael.stahl@allotropia.de> (cherry picked from commit 90b725675c2964f4a151d802d9afedd8bc2ae1a7) xmlsecurity: fix crash in DocumentDigitalSignatures::isAuthorTrusted() If the argument is null. This function also should use EqualDistinguishedNames(). Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111667 Tested-by: Jenkins Reviewed-by: Michael Stahl <michael.stahl@allotropia.de> (cherry picked from commit ca98e505cd69bf95d8ddb9387cf3f8e03ae4577d) Change-Id: I9633a980b0c18d58dfce24fc59396a833498a77d Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111910 Tested-by: Jenkins Reviewed-by: Caolán McNamara <caolanm@redhat.com>
2021-03-18tdf#118693: no need to use convertMm100ToTwip() for line shapes anymoreXisco Fauli3-10/+39
It was introduced in 11129d89b152db54c86bb2bda58c24b8abb6c5a8 < tdf#85232 WPG import: fix handling of line shapes > and later in 36bade04d3780bc54c51b46bb0b63e69789658a5 < tdf106792 Get rid of SvxShapePolyPolygonBezier > ForceMetricToItemPoolMetric was added to SvxShapePolyPolygon::setPropertyValueImpl to convert from 100thmm to twips as can be read in the comment in testTdf85232 With this change, xShape->getPosition().X in testTdf85232 is 2267, which was already in twips Change-Id: I30b757885327a477213f96f8f84541971f435164 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112663 Tested-by: Jenkins Reviewed-by: Miklos Vajna <vmiklos@collabora.com> (cherry picked from commit c9e5640c8fcad7beb42a66f9bee0252eee9fe323) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112619 Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org> Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112671
2021-03-16Revert "tdf#114076: Expand ClipRange to next PixelBound" ...Xisco Fauli1-51/+1
and "tdf#124272 use ClipRegion's geometry if not a rectangle" This commit reverts c1230cede19ae3633e51c7ca780cb34d9dbaa20f and 362c1cf2bd580f6dc8bf27bdcd79174111bc1b5c tdf#114076, or any of its duplicates, is not reproducible in master if c1230cede19ae3633e51c7ca780cb34d9dbaa20f ("tdf#114076: Expand ClipRange to next PixelBound") is reverted. (Tested on Linux and Windows) So, if we revert it, we no longer need 362c1cf2bd580f6dc8bf27bdcd79174111bc1b5c ("tdf#124272 use ClipRegion's geometry if not a rectangle"), which was a follow-up fix for a regression introduced by the first commit. This also fixes tdf#129085 and all the duplicates, which were introduced by the follow-up commit. I plan to add the unittests in a different commit Change-Id: Ie4328c15b24b521127c1b653bd621bfc92ac39cf Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112340 Tested-by: Jenkins Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org> Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112398 Reviewed-by: Armin Le Grand <Armin.Le.Grand@me.com> (cherry picked from commit 8926d9282fe84793ce909471b5caec3583fd3ec9) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112530 Reviewed-by: Caolán McNamara <caolanm@redhat.com>
2021-03-11tdf#140597 DOCX import: fix missing tblPrEx borderLászló Németh2-0/+13
of first table cells, caused by the workaround for tdf#138612. Now property set of a new cell is a copy of the table exception property set of the table row, as needed for the import of the table style inheritance. Regression from commit f319d6b543c2367546bc80d138e56ed03731e265 (tdf#138612 DOCX import: fix lost part of split table cell). Change-Id: Iaf6637e757fbfeef7651a4300a7f65a23615f5c5 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112247 Tested-by: Jenkins Reviewed-by: László Németh <nemeth@numbertext.org> (cherry picked from commit 53884e8fe92597e909e4fa5599192783c3d31a56) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112225 Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
2021-03-11tdf#133473 Change SDRATTR_SHADOWSIZEX flag.Gülşah Köse3-0/+22
SDRATTR_SHADOWSIZEX holds a percentage value. So we shouldn't use METRIC_ITEM. Change-Id: I9f1a20d5871f60d2899ef7d525f4cf837b0a9088 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111311 Tested-by: Jenkins Reviewed-by: Gülşah Köse <gulsah.kose@collabora.com> Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112241 Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112262
2021-03-03tdf#140552 RTF export: fix hyperlink, in footnote, in hyperlinkMiklos Vajna3-4/+57
Regression from commit 7d42346ba77c9c4df241ea40eaf550993ca18783 (tdf#90421 RTF export: ignore hyperlinks without an URL, 2015-04-21), URLs can be nested in the footnote case, which requires a stack. Otherwise the inner URL clears "the" URL and we don't close the outer field as we believe it's empty, so it was not started. Change-Id: I9f87ddbb7e597c413bf836eb9b58beb76722361f Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111794 Tested-by: Jenkins Reviewed-by: Miklos Vajna <vmiklos@collabora.com> (cherry picked from commit 5a74baa4f033f84c4bbcec869a68eef149f77161) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111778 Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org> Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111823 Reviewed-by: Michael Stahl <michael.stahl@allotropia.de>
2021-03-01tdf#140137 Don't throw exception when w:gridCol is missing "w" attrAron Budea2-0/+5
2149e924cbc32c370128c5f87a4f55c50c99e6bd added a division-by-zero check, which caused Writer to throw an error on the bugdoc. Since the file could be loaded fine before, let's return to a working version, with the check included. The cause is the following in document.xml (originating from a non-MS generator): <w:tblGrid> <w:gridCol/> <w:gridCol/> </w:tblGrid> Word still splits such tables differently, but that difference was always there in Writer. Change-Id: I6d91a736f460394a76f035298a238c41da201cb3 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111723 Tested-by: Jenkins Reviewed-by: Caolán McNamara <caolanm@redhat.com> Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111736 Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111741
2021-02-27tdf#140528 Crash in writer when switching from Numbering to PositionNoel Grandin5-8/+26
regression from commit e5ecb998fd78137aec32ad2cc996eaa4bb3e499d Date: Mon Dec 5 09:52:43 2016 +0000 avoid coverity#1371269 Missing move assignment operator rather write this code in a more obvious fashion, and bypass the use of the very weird operator= method, which barely copies anything at all. Change-Id: I548d8f73224c2625ed6389861551038b18396b0f Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111677 Tested-by: Jenkins Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk> (cherry picked from commit 9ab2284660857f52efc9b68b9d2dd8ea768d6916) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111696 Reviewed-by: Caolán McNamara <caolanm@redhat.com>
2021-02-26tdf#134940 sw: fix AutoCorrect of arrow "-->"László Németh1-0/+10
Fix bad replacement of "-->" with "–>" instead of "→" (i.e. premature replacement of "--" to n-dash) since '>' was added to IsAutoCorrectChar(). Regression from commit 57f07b1d7378d218648667c5b1315cc8ad905875 "tdf#133524 AutoCorrect: support double angle quotes". Change-Id: I06f0cddb48d13c8e230dab964f79f588799ed4ab Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111527 Tested-by: László Németh <nemeth@numbertext.org> Reviewed-by: László Németh <nemeth@numbertext.org> (cherry picked from commit d1be3d80d0ca5ccd7639ede379a1befc48dc73f2) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111580 Tested-by: Jenkins Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
2021-02-26tdf#134619 docxexport: don't skip font properties in NumberingLevelJustin Luth3-4/+18
This partially reverts LO 6.4.5 commit 598ca431de96d8bfcf18fa2945e9e30f98387474 The output set can contain more than just the font name. Things like colour and fontsize were being lost when the "else" clause was added. Since it looks like the main intent of the else clause was to avoid specifying a font name twice, erase that from the property set before writing out the rest of the font properties. There is a unit test that enforces that. (I was not able to reproduce the original problem that this was trying to fix. The entire commit is hard to revert, and I will assume that the rest of the commit is useful/accurate, so just fix up a careless portion of the commit.) Change-Id: I772e40e8bd75c0589f3308d4d7470229855aed8f Reviewed-on: https://gerrit.libreoffice.org/c/core/+/110034 Tested-by: Jenkins Reviewed-by: Vasily Melenchuk <vasily.melenchuk@cib.de> Reviewed-by: Justin Luth <justin_luth@sil.org> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/110169 Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org> (cherry picked from commit 5031a2932e9146db8b163ecfcaaf29e59cc4cc9d) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/110558 Reviewed-by: Michael Stahl <michael.stahl@allotropia.de>
2021-02-25tdf#140572 docx export: avoid DocDefault superscript crashJustin Luth3-1/+6
This fixes a crash situation introduced in LO 6.4 by commit d71cf6390a89ea6a4fab724e3a7996f28ca33661. The crash was occurring when the DocDefaults were being exported. Since there was no character or paragraph context, the lookup of the related fontsize was crashing. Moving this variable closer to where it is used "just happens" to fix the crash. Because the DocDefault is considered a style, we are just exporting the superscript/subscript as a default size, and so don't need to know the font size. Thus a style-context position can't cause the crash again. So it isn't a great fix, and leaves the possibility for the same thing to happen in other situations (like RES_CHRATR_BACKGROUND), but at least it quickly solves this known problem, and is easily backportable, as well as following best practice of keeping variables as close to where they are used as possible - eliminating some unnecessary lookups. I did look for a more generic fix (or whether other properties could be affected) and didn't come with any clear examples or a better fix. Change-Id: I03dc3d74a5148973c065cadb09c5c79068f12df1 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111403 Tested-by: Jenkins Tested-by: Justin Luth <justin_luth@sil.org> Reviewed-by: Justin Luth <justin_luth@sil.org> (cherry picked from commit 76d2a9fd461dde817fbb1c5d4aa830cb2c42b504) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111458 Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
2021-01-29sw: fix SwTOXMgr::UpdateOrInsertTOX() for user-definedMichael Stahl1-13/+4
The main problem is that when editing an existing user-defined index, the "Create From" changes aren't applied, because it does pCurTOX->SetCreate() and not pNewTOX->SetCreate(). But that wasn't obvious as the function is obviosly very confused about its 3 different TOX variables. pTOX is just used at the end, so just move it there. pCurTOX can be const. Then there is an odd condition on pSh->HasSelection() which ends up doing DelRight() but fortunately it's dead code since commit e9da29679bce3b544add9233a4aca2b19b78da1a #i97572# cleared the selection already. The pNewTOX = pCurTOX there seems pointless because pNewTOX is already a copy of *pCurTOX. Change-Id: I1e9c8178e314a188a50272410fbe566e4a566742 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/110021 Tested-by: Jenkins Reviewed-by: Michael Stahl <michael.stahl@allotropia.de> (cherry picked from commit 960e88835e0b2f7ff5fc57393b74017198c314e3) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/110064 Reviewed-by: Thorsten Behrens <thorsten.behrens@allotropia.de>
2021-01-27tdf#134605 DOCX export: fix unexpected loss of footer marginMiklos Vajna3-0/+21
Regression from commit 51534ac2b9747975945acb6a1e1ba5cc6d73f5c2 (tdf#127778 DOCX import: fix unexpected heading on non-first page ..., 2020-05-11), which replaced an import fix with an export fix. What was missed there is the import fix handled both headers and footers, while the export fix only handled headers. Fix the problem by doing the same mapping for footers: if a follow page style has no footer, but the first page style has, then take the margin from that one. The import side did the opposite of this already. (cherry picked from commit 80b360085f5974fbb3abe49a88b20f4d39704fb5) Conflicts: sw/qa/extras/ooxmlexport/ooxmlexport16.cxx Change-Id: Ibb191db4136bce0ad5e4bf14dcea15f5955bd298 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/110008 Tested-by: Jenkins Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
2021-01-20tdf#135377 keep the correct index selectedCaolán McNamara1-0/+12
Change-Id: I8737b64300e4bb3a3441c1ec43620b2acea31d2c Reviewed-on: https://gerrit.libreoffice.org/c/core/+/109589 Tested-by: Jenkins Reviewed-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>
2021-01-20tdf#119610 tdf#112634: Don't show broken connections - RevBJustin Luth1-2/+4
This changes LO 6.0 commit 75a881829f19439245cdb859fc16d59461992f79 to use a light-weight check to see if a database exists. The previous way made a connection, which is extremely expensive if there is a password, or if network traffic needs to timeout, etc. And if there are multiple registered databases like that... Change-Id: I980cb6979cfc7cae8f1251f3459718192459aaee Reviewed-on: https://gerrit.libreoffice.org/c/core/+/109242 Tested-by: Jenkins Reviewed-by: Justin Luth <justin_luth@sil.org> (cherry picked from commit 647c9de0338350c1b5eff4b9adf08bd06930e36b) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/109402 Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org> (cherry picked from commit 6c580008d397f5f5f4a0bfd879308018b5d2a54f) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/109582 Reviewed-by: Michael Stahl <michael.stahl@allotropia.de>
2021-01-18tdf#136704 sw autofmt: prevent crash if no nextNodeJustin Luth1-4/+7
The second GetNextNode() was a nullptr that was being dereferenced. Now, when enter is pressed after some text followed by a colon, that line is made into a heading3, and the new blank line is properly using the text body paragraph style. Change-Id: If10dc85e490ac57fc150b3d541f5033dfd5293bb Reviewed-on: https://gerrit.libreoffice.org/c/core/+/109413 Tested-by: Jenkins Reviewed-by: Justin Luth <justin_luth@sil.org> Reviewed-by: Michael Stahl <michael.stahl@allotropia.de> (cherry picked from commit fea7b531a4c9e8326a644cccc26325b16585675a) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/109520
2021-01-16tdf#135014 sw_redlinehide: fix missing frames when removing fieldmarkMichael Stahl2-0/+4
A fieldmark was deleted. UpdateFramesForRemoveDeleteRedline() deleted the MergedPara but its start node was before the start node of the fieldmark, and then MakeFrames() didn't find a frame on the preceding node and did nothing. Similar problem likely possible with redline. Change-Id: I532f9a67c0268f3287736a61da4cc9fefec7b8e4 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/109307 Tested-by: Jenkins Reviewed-by: Michael Stahl <michael.stahl@allotropia.de> (cherry picked from commit 32e104c1d6be0ffe6ed6c4e08af868c87b3c258c) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/109347 Reviewed-by: Caolán McNamara <caolanm@redhat.com>
2021-01-15tdf#135014 sw: fix crash exporting numbering to DOCXMichael Stahl9-16/+52
It can happen that the default SwNumRule of a SwList isn't used by anything directly, but there are other SwNumRule associated with that SwList and then the DOCX export needs to export it as an abstract numbering definition. (regression from 632ee9aae6d5f3cf08b6d6b2789310c20db713b7) Change-Id: I6b1851980464aaa95bf731a60b7d11ab91cec7b6 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/109303 Tested-by: Jenkins Reviewed-by: Michael Stahl <michael.stahl@allotropia.de> (cherry picked from commit cd1c9f5167e797807d6726219f06190657f58372) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/109335 Reviewed-by: Caolán McNamara <caolanm@redhat.com>
2021-01-13tdf#139580: combine these identical blocks together into a functionCaolán McNamara2-28/+29
Change-Id: I783c7f649d0624de9fb46c2b8379fc47ec311f63 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/103917 Tested-by: Jenkins Reviewed-by: Caolán McNamara <caolanm@redhat.com> Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/109226
2021-01-11Related tdf#123632: fix crash in accdoc (sw)Julien Nabet1-1/+1
Make sure to use gtk3 rendering - Open https://bugs.documentfoundation.org/attachment.cgi?id=149492 - Click F5 to open the navigator - Go to Images - Try double click the third and last image at least 15 times -> crash, here's part of bt: Thread 1 "soffice.bin" received signal SIGSEGV, Segmentation fault. SwFrame::IsInTab (this=0x0) at sw/source/core/inc/frame.hxx:938 938 if ( mbInfInvalid ) (gdb) bt 0 SwFrame::IsInTab() const (this=0x0) at sw/source/core/inc/frame.hxx:938 1 0x00007fd7e78caf9c in SwAccessibleDocument::getExtendedAttributes() (this=0x7e4e090) at sw/source/core/access/accdoc.cxx:596 2 0x00007fd7e78cbf9d in non-virtual thunk to SwAccessibleDocument::getExtendedAttributes() () at sw/source/core/access/accdoc.cxx:708 3 0x00007fd7f18f90ad in attribute_set_new_from_extended_attributes(com::sun::star::uno::Reference<com::sun::star::accessibility::XAccessibleExtendedAttributes> const&) (rExtendedAttributes=uno::Reference to (SwAccessibleDocument *) 0x7e4e180) at vcl/unx/gtk3/a11y/gtk3atktextattributes.cxx:1229 4 0x00007fd7f19113f3 in wrapper_get_attributes(_AtkObject*) (atk_obj=0x7df37d0) at vcl/unx/gtk3/a11y/gtk3atkwrapper.cxx:394 I put "Related" because I don't know if it corresponds to the initial bug but I used the attached file of the bugtracker Change-Id: Ief9fda9f7bcf277d18490169eee2e43fb046a1bf Reviewed-on: https://gerrit.libreoffice.org/c/core/+/109051 Tested-by: Jenkins Reviewed-by: Michael Stahl <michael.stahl@allotropia.de> (cherry picked from commit 003b45624788610fba98e3f3334f99140017e472) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/109002 Reviewed-by: Caolán McNamara <caolanm@redhat.com>
2021-01-07Resolves: tdf#139126 DBL_MAX is a valid value, just not for WriterEike Rathke1-1/+3
Restore the old side effect behaviour where "1.79769313486232E+308" was not converted back to DBL_MAX, Writer doesn't check cell value after import for this "special value", *cough*. Reviewed-on: https://gerrit.libreoffice.org/c/core/+/108875 Tested-by: Jenkins Reviewed-by: Eike Rathke <erack@redhat.com> (cherry picked from commit 0e37ded8d4aea25e5d9f7325fba0597f509147bc) Conflicts: sw/source/filter/xml/xmltbli.cxx Change-Id: I31cf598d5f91d1f727d5f1f0e936a3505ea1b9e0 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/108900 Tested-by: Jenkins Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
2020-12-23tdf#138457 Cursor position wrong when using functions in a writer tableCaolán McNamara1-0/+8
Change-Id: I70156d8492afbab2c5993e1627a024ef8e07c492 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/108176 Tested-by: Jenkins Reviewed-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>
2020-12-21tdf#138953: use original (cropped, but unrotated) object size in spPrMike Kaganski5-18/+45
This not only fixes the regression from b226383a83e41bbced9fc2a02dc09a449401ec97, but also makes the written size more correct than before, when it was slightly larger compared to original object size. Corrected unit test for tdf#116371 reflect that: the object in ODT is 241.78 mm x 240.61 mm. It previously was exported as 241.88 x 240.70; now the exported size is closer: 241.79 x 240.63. Change-Id: Ibfe85c7cd98c089e58af8d7f3848990af8e1100f Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107957 Tested-by: Jenkins Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com> (cherry picked from commit 3dc2e629b247873bfbd3190c11152d8d2bab1a03) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107997 Reviewed-by: Caolán McNamara <caolanm@redhat.com>
2020-12-19tdf#82802 sw conditional style conditions not saved for TextBodyJustin Luth3-7/+5
Someone thought it would be a good idea for the built-in paragraph style Text Body to be special and support conditional formatting. It has been broken for import/export since 2015. It was mostly fixed in duplicate issue tdf#103091, but the side case of Text Body was missed. Partially reverting commit a5b4cb3f836c991d0647f55e1ef4920ce6115eac. xNewStyle didn't need to be a member variable since it is assigned to GetStyle anyway. Change-Id: I102d86c0b1f28e42b2c1350abc247b6b1f8acd4e Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107352 Tested-by: Justin Luth <justin_luth@sil.org> Reviewed-by: Justin Luth <justin_luth@sil.org> Reviewed-by: Michael Stahl <michael.stahl@cib.de> (cherry picked from commit 4b5aaa9edfd58ec8f73de127410656f884f944d7) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/108000 Tested-by: Jenkins
2020-12-17tdf#128501: DOC export: fix lost bitmap fill for OOXML custom shapesXisco Fauli2-0/+17
this fix is based on 7032be2e9edd82dad2d67f1582aaa57676bda4a1 which fixes the same problem for PPT filter Change-Id: Id62c29892dd3fce42d27e2e46a7933154cb973f6 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107003 Tested-by: Jenkins Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org> (cherry picked from commit 2f9325b270fab10f6900aec30ca27135363c4c69) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107094 Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
2020-12-17tdf#136816: fix hidden legend entries in "Column and Line" chartsTünde Tóth2-0/+13
Regression from commit: 300e65cc47f3d6ae1563350757dbfadc080d7452 (tdf#123268 fix lost chart if all legend entries are hidden) Change-Id: Id59cd8d681dada123feadbe7910be7fbc7ec37f6 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107757 Tested-by: Jenkins Tested-by: László Németh <nemeth@numbertext.org> Reviewed-by: László Németh <nemeth@numbertext.org> (cherry picked from commit 1eb25c11500a235aa141a327a5489f6861e60a89) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107802 Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org> Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107874
2020-12-16tdf#138892 writerfilter: cancel style list if bNoNumberingJustin Luth4-0/+58
This fixes 6.1 regression from tdf#95377's commit e24e2d2fb02239753c1520a0458b44683ea4cc2e. Starting in 7.0 (tdf#131321), paragraph styles kept their numbering property. But even before that, non-paragraphs marked by bNoNumbering were having their direct numbering removed, and then finishParagraph applied para-style numbering since there was no direct numbering. So, we need to pass the bNoNumbering to finishParagraph so that it can cancel out a style-assigned numbering. Change-Id: I0c24af4e9bd0ea3712179a47ed3550f91c8f44b7 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107738 Tested-by: Jenkins Reviewed-by: Justin Luth <justin_luth@sil.org> (cherry picked from commit bfcd952dc7820c4cf8761c4222d29ede039391dc) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107795 (cherry picked from commit e901ffcb93d217f2cc09021a3b2fecb36f9884f6) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107796 Reviewed-by: Miklos Vajna <vmiklos@collabora.com> Tested-by: Justin Luth <justin_luth@sil.org>
2020-12-16tdf#138600 sw: fix too small print area for btlr text in nested tableMiklos Vajna3-1/+28
Regression from commit 435ab51ec8920033b7865f27f4afee8a852a0b31 (tdf#128399 sw btlr: fix clicking to lower rotated cell, 2019-10-29), the bugdoc has a btlr table cell and the row frame of the outer table has a bottom value which is very small at the point when SwFrame::GetPaintArea() is invoked for the inner btlr cell. This means the "cell should not leave its parent" mechanism kicks in and reduces the bottom of the paint area to a small value, so the text is not visible at all. Fix the problem by teaching SwFrame::GetPaintArea() that btlr cell frames are OK to leave their parent towards the bottom of the page; that parent will grow at a later phase of the layout process anyway. Change-Id: I99334bbf0116df8d8ed27f192c81c0441b6c797d Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107730 Tested-by: Jenkins Reviewed-by: Miklos Vajna <vmiklos@collabora.com> Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107787 Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107818
2020-12-11sw: add missing nullptr checkXisco Fauli1-2/+8
introduced by 0aa0fda64057647219954480ac1bab86b0f0e433 See: https://crashreport.libreoffice.org/stats/signature/SwCursor::UpDown(bool,unsigned%20short,Point%20const%20*,long,SwRootFrame%20&) Change-Id: Ifb7a86b0dfd1477d6ffa15c7d4d3289984747f87 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107561 Tested-by: Jenkins Reviewed-by: Miklos Vajna <vmiklos@collabora.com> (cherry picked from commit 0a80dcad342c1be71f467e46a0cf4f5dd1259056) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107615 Reviewed-by: Caolán McNamara <caolanm@redhat.com>
2020-12-11fix printing of 'margins in comment'Caolán McNamara1-5/+6
mangled since... commit c200aa27ee4a0f5a89af6e970c2c22580029eded Author: Caolán McNamara <caolanm@redhat.com> Date: Thu May 21 15:54:15 2020 +0100 remove Size arg from Window::Draw and depend on GetSizePixel which was a real monday-morning commit that bizarrely dropped drawing of the date/resolved widgets entirely and failed to notice that the PixelToLogic call in SwPostItMgr::DrawNotesForPage was originally called on the input pPostIt, not the output pDev Change-Id: I73b87f5889ef13018bcc43f97cd65169a3b77a65 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107534 Tested-by: Jenkins Reviewed-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>
2020-12-10tdf#138688 tdf#124646 sw: fix crash at pasting Calc dataLászló Németh4-3/+119
.. in a table before a numbered paragraph. Dispatcher calls of the workaround for tdf#124646 missed the temporary table in this case, selecting + copying nothing and after that trying to paste the same non-native Calc data again, resulting infinite recursion. Replacing FN_CHAR_LEFT with FN_LINE_UP solved the problem (FN_CHAR_LEFT selected the numbers of the numbered list instead of the temporary table). Fixing the fragile dispatcher calls, now we check the selection of the temporary table to avoid similar crashes. Unit tests are added for the fix, also for the original problem of tdf#124646 (avoid copying hidden rows from Calc, e.g. copying only visible result of a filtering). Regression from commit 0c3ac02d8a3c7ea50ae262daf134c28df5c8b343 (tdf#124646 Don't paste hidden rows of Calc sheets into Writer tables). (Note: to check/show the fix of the crash manually, run the test with $ (cd sw && make -srj8 UITest_sw_table UITEST_TEST_NAME="sheetToTable.sheetToTable.test_tdf138688" SAL_USE_VCLPLUGIN=gen) adding import time time.sleep(5) to the called function of sheetToTable.py) Change-Id: I7b90af8219d6fd00b75d91f7c92fff5744373cc6 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107508 Tested-by: László Németh <nemeth@numbertext.org> Reviewed-by: László Németh <nemeth@numbertext.org> (cherry picked from commit 7720f8cf22718415adb3db2304916581f864f884) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107487 Tested-by: Jenkins Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
2020-12-10uninit_memberCaolán McNamara1-1/+3
Change-Id: Idd42a139dabf74ee20afa8b6fb3c9ec25c7b0b19 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107390 Tested-by: Jenkins Reviewed-by: Michael Stahl <michael.stahl@cib.de>
2020-12-08sw: layout: SwFlyNotify should invalidate a content or tab frameMichael Stahl1-1/+8
The problem happens during layout for a temporary mail merge document, but that is probably incidental. (The SwRootFrame is the 2nd one out of 3 being createed, the one for the target document.) The problem is that a text frame with a fly moves backward, and ~SwFlyNotify() invalidates the position of the next frame, but the next frame is a section frame, and this does not encourage the text frame inside the section frame to move backward at a later time, which leads to general user unhappiness. So invalidate the lower of the section frame too. Change-Id: I14fa20a279979c029c468f5c6f2e5ecad4ccd240 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107360 Tested-by: Jenkins Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de> (cherry picked from commit c797fc768152a57f6cc7fb9dd447138e4b0629ee) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107311 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2020-12-07tdf#138612 DOCX import: fix lost part of split table cellLászló Németh2-0/+17
by removing unnecessary rewriting of cell properties during import of w:tblPrEx. Regression from commit da8ea444b004a0be36964ae9a778f73e752b2673 (tdf#133455 DOCX import: fix table border regression) Change-Id: I93ce36991437644db439c8cf02e1a8503fbdfba7 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107239 Tested-by: Jenkins Reviewed-by: László Németh <nemeth@numbertext.org> (cherry picked from commit f319d6b543c2367546bc80d138e56ed03731e265) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107304 (cherry picked from commit 2c4806d6b8d79c1b95a41b72d295888bca01d5d8) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107305 Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
2020-12-02tdf#138471 wrong adjustment used for spinbuttonCaolán McNamara1-1/+1
resulting in one adjustment used by two spinbuttons Change-Id: I939b00301699288fd8516cb7d6b8cd11c53c9573 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106949 Tested-by: Jenkins Reviewed-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>
2020-12-01sw_fieldmarkhide: fix getFieldMarkFor() to return innermost fieldMichael Stahl1-3/+4
... like it's supposed to; if it finds the outer field and then the next field is inside but ends *before* the rPos then the loop is erroneously terminated. Change-Id: Ic5e469d1c05cecd7feca275dc1689af2c40567f3 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106710 Tested-by: Jenkins Reviewed-by: Michael Stahl <michael.stahl@cib.de> (cherry picked from commit 6a11f94ab366e1d736e97e31e908e86aa5d8b533) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106689 (cherry picked from commit fd596591b385a8b03459a7dd3f79e65a59f530f4) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106939 Reviewed-by: Caolán McNamara <caolanm@redhat.com>
2020-12-01tdf#134626 assertion in ~SwTextNodeCaolán McNamara1-7/+4
taking the simplest example of... http://bugs.documentfoundation.org/attachment.cgi?id=167361 a) on ctrl+v paste into the new document: DocumentContentOperationsManager::CopyImplImpl is called and that (at "Also copy all format templates") causes SwTextNode::ChgTextCollUpdateNum to call rDoc.GetNodes().UpdateOutlineNode(*this) so the new paragraph is now set as an outline node of the document DocNodes that seems perfectly reasonable b) ctrl+v again no new paragraph, text appended to node of a) c) ctrl+z SwUndoSaveContent::MoveToUndoNds is called and rNodes.MakeTextNode is called to make a new TextNode to move the content into. The SwNodes the TextNode is created in are rNds = rDoc.GetUndoManager().GetUndoNodes(), i.e. the undo nodes. MakeTextNode calls... SwTextNode *pNode = new SwTextNode( rWhere, pColl, nullptr ) and the SwTextNode ctor unconditionally does... GetNodes().UpdateOutlineNode(*this); so its set as an outline node in those undo SwNodes In SwNodes::MakeTextNode right after the SwTextNode *pNode = new SwTextNode( rWhere, pColl, nullptr ) call we have... // call method <UpdateOutlineNode(..)> only for the document nodes array if ( IsDocNodes() ) UpdateOutlineNode(*pNode); That seems odd, because we have *already* unconditionally called UpdateOutlineNode in the SwTextNode ctor. d) ctrl+z The SwTextFormatColl is destroyed and SwTextNode::TriggerNodeUpdate is called for the TextNode of a) and is updated. It is also called for the TextNode of c) but due to... // #125329# - assure that text node is in document nodes array if ( !rDoc.IsInDtor() && &rDoc.GetNodes() == &GetNodes() ) c is *not* updated, so on eventual dtor of c the assert fires. Change-Id: I56f9b00dd4665fd44359a01baef66c82937c2781 --- If the conditional UpdateOutlineNode call in SwNodes::MakeTextNode is the right thing to do where only TextNodes in the DocNodes should be updated, then presumably the equivalent call in the SwTextNode ctor should be also conditional, making that MakeTextNode call redundant in any case. ChgTextCollUpdateNum always calls rDoc.GetNodes().UpdateOutlineNode(*this); which suggests that the SwTextNode::ChgFormatColl function should rely on that to call UpdateOutlineNode and not call it directly to also avoid a UpdateOutlineNode in the undo nodes Change-Id: Iedc3fb4d41c24dbbe41fbac259abe41d3757be09 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106011 Tested-by: Jenkins Tested-by: Xisco Fauli <xiscofauli@libreoffice.org> Reviewed-by: Michael Stahl <michael.stahl@cib.de> (cherry picked from commit 138cf950dcdc8a1240452cad6867e197279f42b0) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106937 Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
2020-11-26tdf#138130: sw_uiwriter: Add unittestXisco Fauli2-0/+30
Change-Id: If7eca097e83003d657cc39ae617eaccabcf62a5a Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106610 Tested-by: Jenkins Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org> Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106633
2020-11-25tdf#138130, tdf#123621 sw: fix textbox position according to DOCXTibor Nagy4-3/+22
i.e. when spAutoFit is present in DOCX, and size of the textbox is relative. See also commit cab956c480eb4f619580285c7b9a15b9e6d9b780 (tdf#112312 DOCX legacy shape export: keep fixed size). Co-authored-by: Attila Szűcs (NISZ) Change-Id: I29af97001954ad353a386164b68cd22f6230e3e5 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/98401 Tested-by: László Németh <nemeth@numbertext.org> Reviewed-by: László Németh <nemeth@numbertext.org> Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org> This commit also includes: Clean-up of commit I29af97001954ad353a386164b68cd22f6230e3e5 Change-Id: I1963a5d899da2d122e60c4d87ca32aba4fb1f32a Reviewed-on: https://gerrit.libreoffice.org/c/core/+/102449 Tested-by: Jenkins Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com> Signed-off-by: Xisco Fauli <xiscofauli@libreoffice.org> Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106589
2020-11-25tdf#138253 sw textbox: fix content of draw format on undoMiklos Vajna3-0/+41
Regression from commit da4f9b77a6cd39b1ae5babdd476d1575c8b9371c (tdf#135149 sw: fix deleting textbox of as-char shapes, 2020-09-07), the crash was caused by a user-after-free triggered from the X11 clipboard code. This could happen beause the clipboard document had a draw frame format, which has an SwNode pointer, but the SwNode instance was outside that clipboard document, and the owning document is already gone. So by the time the clipboard document would be deleted, the SwNodeIndex can't de-register itself. The root cause was that the doc model was corrupted after a cut of a textbox + undo: the textbox pointers of the fly/draw formats were OK, but not the SwFormatContent of the draw format. (cherry picked from commit 42e8e16cf93dcf944e5c1106f76aaa32057c0397) Conflicts: sw/source/core/layout/atrfrm.cxx Change-Id: I5761b72948caca397320aed801559e8493c33e1b Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106539 Tested-by: Jenkins Reviewed-by: Caolán McNamara <caolanm@redhat.com>
2020-11-23tdf#138210 check if CustomShapeGeometry exist first.Mark Hung2-0/+9
CustomShapeGeometry does not exist for a text frame. Getting the property throws an Exception and cause a general IO error. Change-Id: I0e31780292d45211bfd1250d0d359c72def50583 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/105834 Tested-by: Jenkins Reviewed-by: Mark Hung <marklh9@gmail.com> (cherry picked from commit 8c14adfa76956e76bac98330ce67f080c90af184) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106237 Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
2020-11-20tdf#138108 terminate on uncaught exception in a11y callbackCaolán McNamara1-1/+1
#1 0x00007ffff5d3fafe in accessibility::AccessibleContextBase::ThrowIfDisposed() (this=0x85de440) at editeng/source/accessibility/AccessibleContextBase.cxx:513 #2 0x00007ffff5d3fbee in accessibility::AccessibleContextBase::getAccessibleParent() (this=0x85de440) at editeng/source/accessibility/AccessibleContextBase.cxx:203 #3 0x00007fff999cad24 in SwAccessibleMap::DoInvalidateShapeSelection(bool) (this=0x687e610, bInvalidateFocusMode=false) at sw/source/core/access/accmap.cxx:1647 #4 0x00007fff999cc5a5 in SwAccessibleMap::InvalidateShapeSelection() (this=0x687e610) at sw/source/core/access/accmap.cxx:1152 #5 0x00007fff999d78b0 in SwAccessibleMap::InvalidateCursorPosition(SwFrame const*) (this=0x687e610, pFrame=0x5a3faa0) at sw/source/core/access/accmap.cxx:2744 #6 0x00007fff9aae5c32 in SwViewShellImp::InvalidateAccessibleCursorPosition(SwFrame const*) (this=0x6594330, pFrame=0x5a3faa0) at sw/source/core/view/viewimp.cxx:345 #7 0x00007fff99b01fcd in SwCursorShell::UpdateCursor(unsigned short, bool) (this=0x6594840, eFlags=7, bIdleEnd=false) at sw/source/core/crsr/crsrsh.cxx:2020 #8 0x00007fff99b02db4 in SwCursorShell::EndAction(bool, bool) (this=0x6594840, bIdleEnd=false, DoSetPosX=false) at sw/source/core/crsr/crsrsh.cxx:279 #9 0x00007fff9a2249b2 in SwFEShell::SelectObj(Point const&, unsigned char, SdrObject*) (this=0x6594840, rPt=Point = {...}, nFlag=0 '\000', pObj=0x0) at sw/source/core/frmedt/feshview.cxx:323 #10 0x00007fff9b30eacb in SwWrtShell::UnSelectFrame() (this=0x6594840) at sw/source/uibase/wrtsh/select.cxx:326 Change-Id: I122d087b67c3615afb0c34aa714bc1066946c8c1 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106139 Reviewed-by: Michael Stahl <michael.stahl@cib.de> Tested-by: Jenkins
2020-11-19sw_redlinehide: fix assert on Replace with nothingMichael Stahl1-2/+5
SwTextNode::ReplaceText() doesn't need to notify in that case, avoids assert in UpdateMergedParaForInsert(), seen in testDateFormFieldContentOperations. Change-Id: Ie80c14bc903d91a289f470f8a732932342a283be Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106062 Tested-by: Jenkins Reviewed-by: Michael Stahl <michael.stahl@cib.de> (cherry picked from commit 53f17e2a0f2a113a8b677940a314b5934c49acc3) Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106007 Reviewed-by: Caolán McNamara <caolanm@redhat.com>
generated by cgit v1.2.3 (git 2.39.1) at 2024-05-06 05:56:23 +0000