Description: Crash after undo they modifications made by track & changes (document comparison) Steps to Reproduce: 1. Download attachment 166847 [details] 2. Open the attached file (export of attachment 166847 [details] to DOCX) 3. Edit -> Track changes -> Compare Document -> Select the ODT 4. Accept all or reject all (maybe 2 different bugs?) 5. CTRL+Z everything -> crash Probably to many undo-steps Actual Results: Crash Expected Results: No crash Reproducible: Always User Profile Reset: No Additional Info: Version: 7.1.0.0.alpha0+ (x64) Build ID: 94e6e140491de31c0788c91af855a75a3bb12709 CPU threads: 4; OS: Windows 6.3 Build 9600; UI render: Skia/Raster; VCL: win Locale: nl-NL (nl_NL); UI: en-US Calc: CL
No crash in Version: 5.4.0.0.alpha1+ Build ID: 9feb7f7039a3b59974cbf266922177e961a52dd1 CPU threads: 4; OS: Windows 6.29; UI render: GL; Locale: nl-NL (nl_NL); Calc: CL
no attached file
Created attachment 166870 [details] Example file
(In reply to Xisco Faulí from comment #2) > no attached file It's always the same :-(
Created attachment 166955 [details] backtrace from assertion At the end of step 3., my LO crashes with message (rewrapped): soffice.bin: /home/terry/lo_hacking/git/libo6/sw/source/core/bastyp/index.cxx:234: virtual SwIndexReg::~SwIndexReg(): Assertion `!m_pFirst && !m_pLast && "There are still indices registered"' failed. This is in a local build of commit fbd6d6cc (2020-10-30), configured with --enable-debug , built and running on debian-buster. I suspect that my LO is crashing earlier in STR because of --enable-debug. I have found two bugs already reporting the same assertion: - bug 126539, upon reject changes. - bug 130699, upon other undo/redo. Five or six of frames leading up to the assertion in the backtraces attached to those bugs match function names with my backtrace.
No crash for me on Fedora 32, either build id 94e6e140491de31c0788c91af855a75a3bb12709 or today's master.
Hello Telesto, Could you please try to reproduce it with a master build from http://dev-builds.libreoffice.org/daily/master/ ? You can install it alongside the standard version. I have set the bug's status to 'NEEDINFO'. Please change it back to 'UNCONFIRMED' if the bug is still present in the master build
Created attachment 173421 [details] DOCX
Created attachment 173422 [details] ODT 1. Open the DOCX attachment 2. Edit -> Track Changes -> Compare document 3. Pick the ODT 4. Edit -> Track Changes -> Manage 5. Reject all 6. Press and hold CTRL+Z -> Crash or freeze
Version: 7.3.0.0.alpha0+ (x64) / LibreOffice Community Build ID: bd2f2273d83dcca43eb6b465308707efd45e7adf CPU threads: 4; OS: Windows 6.3 Build 9600; UI render: Skia/Raster; VCL: win Locale: nl-NL (nl_NL); UI: en-US Calc: CL
Created attachment 173478 [details] backtrace on linux The interesting part of the backtrace starts at frame #18, where idx==778, m_nSize==778. In a local --enable-debug --enable-dbgutil build of commit d95e400b (2021-06-18), built and running on debian-buster, using the STR from the bug description, I see a crash in step 5 after the second <ctrl>+Z. Final terminal messages are (rewrapped): warn:legacy.osl:22417:22417:sw/source/core/access/acccontext.cxx:445: fire event for disposed frame? warn:legacy.osl:22417:22417:sw/source/core/layout/anchoredobject.cxx:569: <SwAnchoredObject::GetObjRectWithSpaces> - cache for object rectangle inclusive spaces marked as valid, but it couldn't be. Missing invalidation of cache. warn:sw.core:22417:22417:sw/source/core/docnode/node.cxx:1980: Wrong cond collection, skipping check of Cond Colls. warn:sw.core:22417:22417:sw/source/core/docnode/node.cxx:1980: Wrong cond collection, skipping check of Cond Colls. warn:legacy.osl:22417:22417:sw/source/core/access/acccontext.cxx:445: fire event for disposed frame? warn:legacy.osl:22417:22417:sw/source/core/access/acccontext.cxx:445: fire event for disposed frame? soffice.bin: /home/terry/lo_hacking/git/libo6/sw/source/core/bastyp/bparr.cxx:84: BigPtrEntry* BigPtrArray::operator[](sal_uLong) const: Assertion `idx < m_nSize' failed. Unspecified Application Error I am setting status NEW.
Created attachment 173479 [details] bibisect in linux-64-6.1, tail of terminal output Working in bibisect-linux-64-6.1 on debian buster, I see: commit s-h date -------- -------- ------------------- good e3c3da3d 6aa9f098 2019-01-15 17:13:04 bad 8e8eab51 736a3841 2019-01-15 18:01:57 and commit message starting : commit 736a38419f49d4b96bd845f5767a028df59e4cff Author: Serge Krot <Serge.Krot@cib.de> AuthorDate: Thu Dec 6 19:11:58 2018 +0100 Commit: Michael Stahl <Michael.Stahl@cib.de> CommitDate: Tue Jan 15 19:01:57 2019 +0100 sw: DOCX: recognize TOC title during import I am removing keyword bibsectRequest and add keywords bibisected, bisected.
Still crashing Version: 7.4.0.0.alpha0+ (x64) / LibreOffice Community Build ID: 4a388f5e01ebb5a512931d11e48c4380382239c8 CPU threads: 4; OS: Windows 6.3 Build 9600; UI render: Skia/Raster; VCL: win Locale: nl-NL (nl_NL); UI: en-US Calc: CL
Adding CC: to Serge Krot
I followed the exact steps in Comment #9. I was able to reproduce CRASH: - https://crashreport.libreoffice.org/stats/crash_details/ac96378d-583e-452a-906d-afa95049754f using: Version: 7.6.4.1 (X86_64) / LibreOffice Community Build ID: e19e193f88cd6c0525a17fb7a176ed8e6a3e2aa1 CPU threads: 8; OS: Windows 10.0 Build 19045; UI render: Skia/Raster; VCL: win Locale: en-US (en_US); UI: en-US Calc: CL threaded