Bug 83555 - Crash in presentation by entering/leaving specific slide with page-down
Summary: Crash in presentation by entering/leaving specific slide with page-down
Alias: None
Product: LibreOffice
Classification: Unclassified
Component: Impress (show other bugs)
(earliest affected) Master
Hardware: Other Linux (All)
: medium normal
Assignee: Michael Stahl (allotropia)
Keywords: bisected
Depends on:
Reported: 2014-09-06 09:12 UTC by Cor Nouws
Modified: 2014-09-14 14:11 UTC (History)
3 users (show)

See Also:
Crash report or crash signature:

crashing presentation (2.84 MB, application/vnd.oasis.opendocument.presentation)
2014-09-06 09:13 UTC, Cor Nouws

Note You need to log in before you can comment on or make changes to this bug.
Description Cor Nouws 2014-09-06 09:12:14 UTC
I'll attach a presentation.
Focus on slide one in design mode
4 slides.
Pg-Dwn, Pg-Dwn ...
From slide #3 to #4 impress crashes
Comment 1 Cor Nouws 2014-09-06 09:13:57 UTC
Created attachment 105832 [details]
crashing presentation
Comment 2 A (Andy) 2014-09-06 10:58:27 UTC
For me not reproducible with LO (Win 8.1).
Therefore, this issue maybe hasn't occurred in earlier LO releases? 

Steps done:
1. Opened the attached presentation
2. Went to VIEW -> TOOLBARS and selected FORM CONTROLS
3. Switched ON the DESIGN MODE (icon at the top right)
4. Used Page Down keys to scroll through the presenation
Comment 3 Jean-Baptiste Faure 2014-09-06 12:15:46 UTC
I do not reproduce the crash with following versions build at home under Ubuntu 14.04 x86-64 :
- Version: Build ID: 10817252c56f48b0026158fc37a55005098a9d78
- Version: Build ID: dbc24b1ff71a8da7e23e9c1657492fb3c8a66de8

To be sure: what do you mean by "design mode"? I understood that as edit mode.

Best regards. JBF
Comment 4 Cor Nouws 2014-09-06 17:00:56 UTC

Sorry for not being clear. I mean the 'Normal' mode.
And my Version:
Build ID: 1298c1d13572c6fbfbabb813b2d6843368c6df1f
Working on Ubuntu Linux 32 bits

It also crashes with a clean user profile.
Comment 5 Jean-Baptiste Faure 2014-09-06 17:19:11 UTC
Hmmm, your build is old (395 commits older than mine), I suggest you to try again with a current daily build.

Best regards. JBF
Comment 6 Terrence Enger 2014-09-06 22:02:34 UTC
Following the instructions provided, I have made a segfault with
PC=0x0; frame 1 of the backtrace is in function

I notice in at least one version, if I start by typing <End> to
display Slide 4, there is no crash, not even when I type <Home> and
then successive <Page Down> to leave and return to Slide 4.  Maybe
this says something about how the segfault happens.

The crash appears to have been introduced and removed from master
within the range covered by the daily dbgutil bibisect repo.
Specifically, it first appears in 369f344... 2014-08-10 and it is
first gone in 8c0ffaa... 2014-09-01.

So, this bug report is obviously headed for FESOLVED WORKSFORME.  But
I wonder if there may be implications for other branches.  I shall
leave it for somebody with better git-foo than mine to resolve the

For completeness, as I worked from the oldest version in the repo to
find the introduction of the crash, I see `git bisect bad` says:

    369f34427476c645cac930bdf29234638a1b41ae is the first bad commit
    commit 369f34427476c645cac930bdf29234638a1b41ae
    Author: Miklos Vajna <vmiklos@collabora.co.uk>
    Date:   Sun Aug 10 11:05:46 2014 +0000


    :100644 100644 8c5b1d442bd7d00aa095491efff03c497abf0714 60ecce52baca1650da7c57df652a27469fdb75d5 M	build-info.txt
    :040000 040000 e6b8fb5a60898792ecfb8245b51e0abe14f768de 13fc7876272e8fbcbe2683b53338aeb72061869e M	opt

and `git bisect log`:

    # bad: [8c0ffaa3f2ace1b3a8e8e1db44b11ba083637e4d] 2014-09-01
    # good: [b3130c846de5cf1b4be48b48dfc780bb369549fa] 2014-05-21
    git bisect start '8c0ffaa' 'oldest'
    # good: [eb37477c7906b4c4b7eb11433c7566f669a040d1] 2014-07-11
    git bisect good eb37477c7906b4c4b7eb11433c7566f669a040d1
    # good: [3b906f385ccc5e3bda8892bf4ff72744876172af] 2014-08-06
    git bisect good 3b906f385ccc5e3bda8892bf4ff72744876172af
    # bad: [e0f6f58d1ff3c05c404dbdc21dbe1ccbc94d4276] 2014-08-19
    git bisect bad e0f6f58d1ff3c05c404dbdc21dbe1ccbc94d4276
    # bad: [59a8d89d4255a1cbcab4fcfeff9270dd56c62aed] 2014-08-12
    git bisect bad 59a8d89d4255a1cbcab4fcfeff9270dd56c62aed
    # good: [ccf7a47fb02d2909d2ad6d03d6218d3cf07b3b2a] 2014-08-09
    git bisect good ccf7a47fb02d2909d2ad6d03d6218d3cf07b3b2a
    # bad: [6b0c7213477d9c4ebf74bdf74871a684f1a9b6dd] 2014-08-11
    git bisect bad 6b0c7213477d9c4ebf74bdf74871a684f1a9b6dd
    # bad: [369f34427476c645cac930bdf29234638a1b41ae] 2014-08-10
    git bisect bad 369f34427476c645cac930bdf29234638a1b41ae

and working from the version of 2014-08-29 to find the disappearance
of the crash, I see `git bisect bad`:

    8c0ffaa3f2ace1b3a8e8e1db44b11ba083637e4d is the first bad commit
    commit 8c0ffaa3f2ace1b3a8e8e1db44b11ba083637e4d
    Author: Miklos Vajna <vmiklos@collabora.co.uk>
    Date:   Mon Sep 1 09:48:23 2014 +0000


    :100644 100644 de56afe3d5f5e122723947b24311bd00acad1f70 60248ccd1189423ee8b6cafc100cfe547e335431 M	build-info.txt
    :040000 040000 f22263c2481b8cdc146b58a23b60c9ec2d6f56f0 7e57ec88799dc9cff16206c4351dec8337982f7a M	opt

and `git bisect log:

    # bad: [827afd643c75bdea4680bfee7b131286c0bdedba] 2014-09-06
    # good: [a8232b4c392ff3508dc1a7b6ab5dd9a6d2bfdd99] 2014-08-29
    git bisect start 'origin/master' 'a8232b4c'
    # bad: [75759d2694646b0451a091448da51bf098722c35] 2014-09-02
    git bisect bad 75759d2694646b0451a091448da51bf098722c35
    # good: [db4bc4671fbec1a556703a0fb0046aed7bb45d90] 2014-08-31
    git bisect good db4bc4671fbec1a556703a0fb0046aed7bb45d90
    # bad: [8c0ffaa3f2ace1b3a8e8e1db44b11ba083637e4d] 2014-09-01
    git bisect bad 8c0ffaa3f2ace1b3a8e8e1db44b11ba083637e4d
    # first bad commit: [8c0ffaa3f2ace1b3a8e8e1db44b11ba083637e4d] 2014-09-01
Comment 7 Caolán McNamara 2014-09-14 14:11:14 UTC
This was fixed by 790920befb242d4843d7a933c8a3b7af31d404f9