Bug 83366

Summary: EDITING: Copy and Paste crashes Calc when original doc closed
Product: LibreOffice Reporter: Luke <lukebenes>
Component: SpreadsheetAssignee: Not Assigned <libreoffice-bugs>
Status: NEW --- QA Contact:
Severity: normal    
Priority: highest CC: jmadero.dev, libreoffice, lo_bugs, ogldelphi, raul.malea
Version: 4.3.1.2 releaseKeywords: bisected, have-backtrace, regression
Hardware: Other   
OS: All   
Whiteboard: bibisected
i915 platform: i915 features:
Bug Depends on:    
Bug Blocks: 75025    
Attachments: gdb on the core file

Description Luke 2014-09-01 19:13:57 UTC
This is an issue with the latest 4.4 master, and cannot be reproduced on 4.2.x. If you skip step 4, it's also not reproducable. 

Steps to Reproduce:
1. Open  Attachment 103075 [details] in Calc 4.4
2. Highlight I5:I24
3. Copy
4. Close the document (Ctrl-W)
5. Create a new Spreadsheet
6. Paste

Expected Results:
Contents of I5:I24 in the new spreadsheet

Actual Results:
Calc Crashes
Comment 1 Joel Madero 2014-09-03 08:10:59 UTC
Confirmed:
Bodhi 2.x
LibreOffice 4.3.1.2 release - updating version to reflect this

I don't see any crash in bibisect package so probably a recent regression. 


Marking as:
New - confirmed
Critical - this is pretty nasty, easy to repro and pretty common thing that someone might do. Serious data loss possibility
Highest - MAB (regression + data loss + common set of steps)
Comment 2 Terrence Enger 2014-09-03 11:49:23 UTC
I have make a segfault by copying from a single cell and closing the
workbook before pasting into a new worksheet.

This observation is on master commit f74a633, fetched 2014-08-23 1728
UTC, configured:
    --enable-option-checking=fatal --enable-dbgutil --enable-crashdump
    --without-system-postgresql --without-myspell-dicts
    --with-extra-buildid --without-doxygen
    --with-external-tar=/home/terry/lo_hacking/git/src
    --enable-online-update
built and running on debian-wheezy 64-bit.

However, running in a chroot enviornment to debian-sid, that same
build does not crash, and neither do three versions from the daily
dbgutil bibisect: oldest, 2014-09-02, and 2014-08-23.  The chroot sid
environment has newer system libraries, of course.


Joel,

Did you have your crash and your attempt to bibisect on the same
platform?

Thanks,
Terry.
Comment 3 Joel Madero 2014-09-03 11:50:44 UTC
Yes same system. Very strange
Comment 4 Terrence Enger 2014-09-03 14:37:51 UTC
Created attachment 105687 [details]
gdb on the core file

Even stranger than that.  My crash in master happens only sometimes.
I am almost reduced to thinking that the crash happens only when I
grow careless about what I am doing.

Meanwhile this backtrace from commit 754a633 shows a segfault in
ScFormatShell::GetStyleState.
Comment 5 Björn Michaelsen 2014-11-28 09:45:22 UTC
(This is an automated message.)

Setting priority to highest as this is a MAB. This is part of an effort to make the importance of MAB reflected in priority too.
Comment 6 R.Yu. 2014-12-30 13:58:38 UTC
git bisect start
# bad: [0777cd085a7633a48e03d25948cc67fce87b7ac7] source-hash-b800d0b6ad74ce4a9adb23b865dd174d1eefa47b
git bisect bad 0777cd085a7633a48e03d25948cc67fce87b7ac7
# good: [812c4a492375ac47b3557fbb32f5637fc89d60d9] source-hash-dea4a3b9d7182700abeb4dc756a24a9e8dea8474
git bisect good 812c4a492375ac47b3557fbb32f5637fc89d60d9
# bad: [8677ba6e74a774fb44ec7831f14e53d8663f59ed] source-hash-eb213e490d9a366477b921d1a408d85c4638499e
git bisect bad 8677ba6e74a774fb44ec7831f14e53d8663f59ed
# bad: [42455e7bcd46db5523f7f8e1931bbd0bd4b51cfc] source-hash-b5608fd429790a3d1153341b2c86303b7090b15a
git bisect bad 42455e7bcd46db5523f7f8e1931bbd0bd4b51cfc
# good: [253e662b25e57508dbc46753f0f28af36d5f4e25] source-hash-f7d51f43deda5e28df63f1b8e168e84838d0d0b4
git bisect good 253e662b25e57508dbc46753f0f28af36d5f4e25
# bad: [8461c942c2060278f64283368705e0183709c4d7] source-hash-c7d390bf21623c148ff5c3955561b903d9581da8
git bisect bad 8461c942c2060278f64283368705e0183709c4d7
# bad: [a59adcf09aa0f7900332f3e0e48e5b0fbeffe82d] source-hash-c0cb909ba7f01e3d644e8555289be3ab0785d15d
git bisect bad a59adcf09aa0f7900332f3e0e48e5b0fbeffe82d
# skip: [30e36f1e3a8d687c8ee493dda6d8fcd9b651fdb2] source-hash-ea93714e9427e6802e843c4585871aeef0a5e28d
git bisect skip 30e36f1e3a8d687c8ee493dda6d8fcd9b651fdb2
# bad: [998586070e97205086dca415df199a23465ef6be] source-hash-e8e22fcefd569bdf6c3fbe67bb64206e67c3e6b7
git bisect bad 998586070e97205086dca415df199a23465ef6be
# good: [0219641f7ec1bee5654f7b8c02d246fa27b03ab1] source-hash-9835fa87ac6ffe43bd9cc85839b2fea1fca2fcad
git bisect good 0219641f7ec1bee5654f7b8c02d246fa27b03ab1
# only skipped commits left to test
# possible first bad commit: [998586070e97205086dca415df199a23465ef6be] source-hash-e8e22fcefd569bdf6c3fbe67bb64206e67c3e6b7
# possible first bad commit: [30e36f1e3a8d687c8ee493dda6d8fcd9b651fdb2] source-hash-ea93714e9427e6802e843c4585871aeef0a5e28d

There are only 'skip'ped commits left to test.
The first bad commit could be any of:
30e36f1e3a8d687c8ee493dda6d8fcd9b651fdb2
998586070e97205086dca415df199a23465ef6be
We cannot bisect more!
Comment 7 Luke 2015-01-04 22:06:05 UTC
I have manually bisected this down to:
fdo#79676: Initialize with a default chart only from the UI.
http://cgit.freedesktop.org/libreoffice/core/commit/?id=c5539bb585370f095cb6c09e38a7dd005d237295

The previous commit, 178550923533b7c015e29d957a2fc3857ba73eb3 was GOOD. 

Kohei, could you please look how your fix for Bug 79676 caused this regression? Thanks for everything!
Comment 8 Joel Madero 2015-01-05 02:06:17 UTC
Kohei does not follow bugzilla anymore. I'll try to ping him personally to see if he will investigate

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.