Bug 455130

Summary:	can not start ksmserver, ASSERT: "d" in ... kdecore/kernel/kcomponentdata.cpp, line 158
Product:	[Fedora] Fedora	Reporter:	Rex Dieter <rdieter>
Component:	kdelibs	Assignee:	Than Ngo <than>
Status:	CLOSED RAWHIDE	QA Contact:	Fedora Extras Quality Assurance <extras-qa>
Severity:	high	Docs Contact:
Priority:	high
Version:	rawhide	CC:	kevin, ltinkl, rdieter
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:	4.0.98-2.fc10	Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2008-07-12 22:11:04 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:
Bug Blocks:	446427

Description Rex Dieter 2008-07-12 13:42:42 UTC

Just and fyi, haven't had much chance to look more, but kde-4.0.98 (built for
F-9 here), falls over with many-a-error of the form:

kdeinit4: preparing to launch /usr/libexec/kde4/klauncher
ASSERT: "d" in file
/builddir/build/BUILD/kdelibs-4.0.98/kdecore/kernel/kcomponentdata.cpp, line 158
kdeinit4: (klauncher /usr/libexec/kde4/klauncher) Pipe closed
unexpectedlykdeinit4: Pipe closed unexpectedly: No such file or director

Comment 1 Kevin Kofler 2008-07-12 16:04:50 UTC

Argh, sounds badly broken!

Comment 2 Kevin Kofler 2008-07-12 16:11:03 UTC

The context:
156 KStandardDirs *KComponentData::dirs() const
157 {
158       Q_ASSERT(d);
159       d->lazyInit(*this);
160         
161       return d->dirs;
162 }

The code of KComponentData hasn't changed since September 2007. I think we need 
a backtrace for this assertion failure.

Comment 3 Rex Dieter 2008-07-12 16:21:00 UTC

Confirmed going back to kde-4.0.85 is ok, so seems to be something particular to
4.0.98.  I'm tempted to try rebuilding everything again, to confirm it's just
not a quirk in my own builds (at least until we have any confirmation of folks
seeing this on rawhide too).

Once I confirm that, I'll get to work on getting a backtrace.

Fwiw, the 4.0.98 F-9 builds I'm using here are currently in repos at
http://apt.kde-redhat.org/apt/kde-redhat/mock/fedora-9-i386-kde/
http://apt.kde-redhat.org/apt/kde-redhat/mock/fedora-9-x86_64-kde/

Comment 4 Rex Dieter 2008-07-12 17:35:25 UTC

How most folks will experience this is that login will fail with error:
can not start ksmserver

Comment 5 Rex Dieter 2008-07-13 00:36:24 UTC

I've confirmed that a kdelibs build omitting our nocache and kstandarddirs
patches still exhibits this problem (so I'm hoping that means that those patches
don't contribute to it).

Comment 6 Kevin Kofler 2008-07-13 00:42:19 UTC

Any chance you can get a stack trace for that? It might help figuring out 
what's going wrong.

Comment 7 Kevin Kofler 2008-07-13 00:47:58 UTC

(For those who haven't followed the discussion on IRC: I have fixed this in 
4.0.98-2.fc10 by reverting a recent kinit change which triggered this bug, but 
I think the kinit change was only uncovering a latent bug and want the real 
issue to be fixed.)

Comment 8 Kevin Kofler 2008-08-29 14:54:29 UTC

I just committed a clean fix in 4.1.1-2. Details forthcoming in an upstream bug report.

Comment 9 Kevin Kofler 2008-08-29 15:23:28 UTC

Upstream bug report: https://bugs.kde.org/show_bug.cgi?id=170043