For one user, the use of sieve filter delete is making kontact segfault with a sieve error.
We need to have information about how to reproduce.
Description
Description
Details
Details
- Ticket Type
- Task
Revisions and Commits
Revisions and Commits
| Restricted Differential Revision |
Related Objects
Related Objects
- Duplicates Merged Here
- T1184: Kontact crashes with some sieve setups
Event Timeline
Comment Actions
That is being observed for some users. Kontact just fails to start. Disabling Sieve in account configuration makes it work.
Thread 1 (Thread 0x7f593e81f800 (LWP 25383)): [KCrash Handler] #5 0x00007f593c9672b0 in KUrl::setFileName(QString const&) () from /usr/lib/libkdecore.so.5 #6 0x00007f58c62350c9 in ?? () from /usr/lib/libksieveui.so.4 #7 0x00007f58c6235113 in ?? () from /usr/lib/libksieveui.so.4 #8 0x00007f58c62354c1 in ?? () from /usr/lib/libksieveui.so.4 #9 0x00007f593c28187a in QMetaObject::activate(QObject*, QMetaObject const*, int, void**) () from /usr/ #10 0x00007f58c433b69b in KManageSieve::SieveJob::gotScript(KManageSieve::SieveJob*, bool, QString const #11 0x00007f58c433a9b7 in KManageSieve::SieveJob::Private::handleResponse(KManageSieve::Response const&, #12 0x00007f58c43371c1 in ?? () from /usr/lib/libkmanagesieve.so.4 #13 0x00007f58c4338063 in ?? () from /usr/lib/libkmanagesieve.so.4 #14 0x00007f58c4338185 in ?? () from /usr/lib/libkmanagesieve.so.4 #15 0x00007f593c28187a in QMetaObject::activate(QObject*, QMetaObject const*, int, void**) () from /usr/ #16 0x00007f593c9f9402 in ?? () from /usr/lib/libkdecore.so.5
Comment Actions
A better backtrace with debug symbols installed for libksieveui4 and kdelibs5 would be very useful.
Comment Actions
@pfr debug symbols are available as a separate package.
Please install:
- kdepimlibs-dbg
- kdepim-dbg
Comment Actions
Ok. my fault. I forgot to mention that we are not the original reporter (adomaitis). We use Kontact on Windows, and get Kontact directly from Kolab. Current installer is Kontact-E14-2016-04-15-06-11.exe
Comment Actions
Oh, I see. In that case it's not as easy. I'll have to look into how I can provide debug symbols with the windows installer.
Comment Actions
I tried no severl times with our current test instance and I can't trigger the crash,
-> That means that it is dependen from the sieve scripts or the way the scritps are activated.
Can you please give me the sieve scripts?
Is KEP:14 used for the sieve scripts?
Is there an alternate URL added for the sieve server?
managesieve also gives out many debug statements, can you give me this output?
Comment Actions
I think we now found a way to force Kontact into the problematic state: please see https://issues.kolab.org/show_bug.cgi?id=5349
Comment Actions
We only have server-side Sieve scripts, and they are not activated (see attached screenshot).
Here is the content:
kontact:
require "fileinto";
#SCRIPTNAME: xyz
if header :contains "Von" "xing"
{
fileinto "Spam";
}
roundcube:
require ["date","relational","vacation-seconds"];
- EDITOR Roundcube (Managesieve)
- EDITOR_VERSION 8.4
- rule:[Urlaub]
if false # allof (currentdate :zone "+0200" :value "ge" "date" "2016-04-02", currentdate :zone "+0200" :value "le" "date" "2016-04-08")
{
vacation :seconds 125 :subject "fghcgfh" "xcfgjxdxdgk";
}
Comment Actions
please open kdebugdialog and activate kio_sieve and kontact and than restart kontact and give use the debuglog for this session.
Comment Actions
If KEP:14 support is used is hidden from the user inside roundcube and kontact, so we need the raw filelisting. If there is a script name USER than it is KEP:14. But this we see in the debuglog.