This one really puzzles me. Do you see \n in any other config file? (~/.kde[4]/share/config/*)

yes i have some in /home/test/.kde4/share/config/filetypesrc

its content is like this : 

[EmbedSettings]
embed-\napplication/ogg=true
embed-\napplication/x-nsv-vp3-mp3=true
embed-\napplication/x-ogg=true
embed-\naudio/basic=true
embed-\naudio/flac=true
embed-\naudio/midi=true
embed-\naudio/mp3=true
embed-\naudio/mpeg=true
embed-\naudio/mpeg2=true
embed-\naudio/mpeg3=true
embed-\naudio/ogg=true
embed-\naudio/x-basic=true
embed-\naudio/x-flac=true
embed-\naudio/x-mpeg=true
embed-\naudio/x-mpeg2=true
embed-\naudio/x-mpeg3=true
embed-\naudio/x-mpegurl=true
embed-\naudio/x-scpls=true
embed-\nvideo/divx=true
embed-\nvideo/fli=true
embed-\nvideo/mp4=true
embed-\nvideo/vnd.divx=true
embed-\nvideo/vnd.vivo=true
embed-\nvideo/x-fli=true
embed-\nvideo/x-mpeg=true
embed-\nvideo/x-mpeg2=true
embed-application/asx=true
embed-application/futuresplash=true
embed-application/vnd.rn-realaudio=true
embed-application/x-drm-v2=true
embed-application/x-google-vlc-plugin=true
embed-application/x-mplayer2=true
embed-application/x-ms-wmp=true
embed-application/x-ms-wmv=true
embed-application/x-quicktimeplayer=true
embed-audio/wav=true
embed-audio/x-ms-wax=true
embed-audio/x-ms-wmv=true
embed-audio/x-pn-realaudio=true
embed-audio/x-realaudio=true
embed-interface/x-winamp-skin=false
embed-text/htmlh=false
embed-unknown/mime-type=true
embed-video/divx=true
embed-video/msvideo=true
embed-video/vnd.divx=true
embed-video/x-ms-asf-plugin=true
embed-video/x-ms-wm=true
embed-video/x-ms-wmp=true
embed-video/x-ms-wvx=true
embed-video/x-quicktime=true

So this is about mimetypes again, not a KConfig bug.

Does "kmimetypefinder foo.mp3" answer
audio/mpeg
(accuracy 100)

or is there a newline (a real one this time, I think), before audio/mpeg?

i obtain this : 

$ kmimetypefinder foo.mp3
audio/mpeg
(accuracy 100)

i started to add the nonewline patch you gave me  ( the one that apply on apps/ )
now i can go on the file association part anymore ( i removed /home/test/.kde4/share/config/filetypesrc  and ~/.local/share/applications/mimeapps.list  but this is  the same.

Application : Configuration du système (systemsettings), signal SIGABRT
[Current thread is 1 (Thread 0xb5fa46d0 (LWP 13502))]

Thread 1 (Thread 0xb5fa46d0 (LWP 13502)):
[KCrash Handler]
#6  0xffffe424 in __kernel_vsyscall ()
#7  0xb68c3d90 in raise (sig=6) at ../nptl/sysdeps/unix/sysv/linux/raise.c:64
#8  0xb68c57f8 in abort () at abort.c:88
#9  0xb744b445 in qt_message_output (msgType=QtFatalMsg, 
    buf=0xbfd40280 "ASSERT: \"!mimeName.contains('\\n')\" in file /home/neoclust/rpm/BUILD/kdebase-4.1.85/apps/konqueror/settings/filetypes/mimetypedata.cpp, line 53") at global/qglobal.cpp:2102
#10 0xb744b535 in qFatal (msg=0xb759eb78 "ASSERT: \"%s\" in file %s, line %d") at global/qglobal.cpp:2303
#11 0xb744b5e5 in qt_assert (assertion=0xb47c70ba "!mimeName.contains('\\n')", file=0xb47c6e10 "/home/neoclust/rpm/BUILD/kdebase-4.1.85/apps/konqueror/settings/filetypes/mimetypedata.cpp", line=53)
    at global/qglobal.cpp:1872
#12 0xb47c1bf4 in MimeTypeData (this=0x9077380, mime={d = 0xbfd42318}) at /usr/src/debug/kdebase-4.1.85/apps/konqueror/settings/filetypes/mimetypedata.cpp:53
#13 0xb47bf1b6 in TypesListItem (this=0x9077360, parent=0x90789a8, mimetype={d = 0xbfd42374}) at /usr/src/debug/kdebase-4.1.85/apps/konqueror/settings/filetypes/typeslistitem.cpp:39
#14 0xb47b3d1e in FileTypesView::readFileTypes (this=0x8f2fd48) at /usr/src/debug/kdebase-4.1.85/apps/konqueror/settings/filetypes/filetypesview.cpp:216
#15 0xb7ad0ae3 in KCModule::qt_metacall (this=0x8f2fd48, _c=QMetaObject::InvokeMetaMethod, _id=2, _a=0x8ea7a38) at /usr/src/debug/kdelibs-4.1.85/build/kdeui/kcmodule.moc:77
#16 0xb47b4372 in FileTypesView::qt_metacall (this=0x8f2fd48, _c=QMetaObject::InvokeMetaMethod, _id=29, _a=0x8ea7a38)
    at /usr/src/debug/kdebase-4.1.85/build/apps/konqueror/settings/filetypes/filetypesview.moc:73
#17 0xb755ab8a in QMetaCallEvent::placeMetaCall (this=0x8fe8af8, object=0x8f2fd48) at kernel/qobject.cpp:529
#18 0xb755c782 in QObject::event (this=0x8f2fd48, e=0x8fe8af8) at kernel/qobject.cpp:1155
#19 0xb6c5a05f in QWidget::event (this=0x8f2fd48, event=0x8fe8af8) at kernel/qwidget.cpp:7551
#20 0xb6bfffac in QApplicationPrivate::notify_helper (this=0x8d82620, receiver=0x8f2fd48, e=0x8fe8af8) at kernel/qapplication.cpp:3803
#21 0xb6c082ad in QApplication::notify (this=0xbfd42b94, receiver=0x8f2fd48, e=0x8fe8af8) at kernel/qapplication.cpp:3768
#22 0xb7a269b0 in KApplication::notify (this=0xbfd42b94, receiver=0x8f2fd48, event=0x8fe8af8) at /usr/src/debug/kdelibs-4.1.85/kdeui/kernel/kapplication.cpp:307
#23 0xb754c853 in QCoreApplication::notifyInternal (this=0xbfd42b94, receiver=0x8f2fd48, event=0x8fe8af8) at kernel/qcoreapplication.cpp:587
#24 0xb754d535 in QCoreApplicationPrivate::sendPostedEvents (receiver=0x0, event_type=0, data=0x8d6d310) at kernel/qcoreapplication.h:209
#25 0xb754d71c in QCoreApplication::sendPostedEvents (receiver=0x0, event_type=0) at kernel/qcoreapplication.cpp:1091
#26 0xb757872d in postEventSourceDispatch (s=0x8d85168) at kernel/qcoreapplication.h:214
#27 0xb62c1cba in IA__g_main_context_dispatch (context=0x8d84838) at gmain.c:1814
#28 0xb62c53d8 in g_main_context_iterate (context=0x8d84838, block=1, dispatch=1, self=0x8d82758) at gmain.c:2448
#29 0xb62c5598 in IA__g_main_context_iteration (context=0x8d84838, may_block=1) at gmain.c:2511
#30 0xb757834a in QEventDispatcherGlib::processEvents (this=0x8d8fa18, flags={i = -1076614536}) at kernel/qeventdispatcher_glib.cpp:319
#31 0xb6c9c92a in QGuiEventDispatcherGlib::processEvents (this=0x8d8fa18, flags={i = -1076614488}) at kernel/qguieventdispatcher_glib.cpp:198
#32 0xb754aec3 in QEventLoop::processEvents (this=0xbfd42b20, flags={i = -1076614424}) at kernel/qeventloop.cpp:143
#33 0xb754b081 in QEventLoop::exec (this=0xbfd42b20, flags={i = -1076614360}) at kernel/qeventloop.cpp:194
#34 0xb754d7f2 in QCoreApplication::exec () at kernel/qcoreapplication.cpp:845
#35 0xb6bffe14 in QApplication::exec () at kernel/qapplication.cpp:3331
#36 0x08056a7b in main (argc=-1076613932, argv=0xb7f437d4) at /usr/src/debug/kdebase-workspace-4.1.85/systemsettings/main.cpp:67

the patches applied for the debuging are : 
http://web.davidfaure.fr/tmp/nonewline.diff
http://web.davidfaure.fr/tmp/kopenwithdialog-nonewline.diff

Good, getting somewhere.
I made a separate test program to understand where this comes from - please unpack, then
cmake . && make && ./kallmimes

If it asserts, then the bug is in kbuildsycoca4. Uncomment this debug output in kded/kbuildmimetypefactory.cpp:
     kDebug() << "Creating mimetype" << name << "from file" << file << "path" << fullPath;
and add this line after it:
    Q_ASSERT(!name.contains('\n'));
Rebuild, then run kbuildsycoca4 --noincremental, and tell me if it asserts.

If kallmimes works fine then, hmm, I'm even more puzzled and the bug is in the filetypes editor.

Created attachment 29603 [details]
kallmimes.tar.bz2

ok it assert with it

ASSERT: "!mimetype.contains('\n')" in file /home/neoclust/rpm/BUILD/testprogs/kallmimes/kallmimes.cpp, line 39
Abandon


i am rebuilding my kdelibs with the kdebug and the assert you asked.

this is how this is in the konsole  :
"video/x-flic"
"audio/x-ape"
"application/sieve"
"
video/x-mpeg2"
ASSERT: "!mimetype.contains('\n')" in file /home/neoclust/rpm/BUILD/testprogs/kallmimes/kallmimes.cpp, line 39
Abandon

nspluginscan bug, please test  http://www.davidfaure.fr/tmp/pluginscan.cpp.diff

\o/

now i obtain :

[Added Associations]
audio/ogg=kde4-amarok.desktop;

OK, for the record, the bug was in the NP_GetMIMEDescription implementation of /usr/lib/mozilla/plugins/mplayerplug-in.so, which returns mimetypes separated by ";\n" instead of just ";" as specified by https://developer.mozilla.org/En/NP_GetMIMEDescription

Let's have the workaround, anyway, can't hurt.

SVN commit 901122 by dfaure:

Don't generate mimetypes like "\nvideo/x-mpeg2" -- the bug was in the NP_GetMIMEDescription implementation of
/usr/lib/mozilla/plugins/mplayerplug-in.so, which returns mimetypes separated by ";\n" instead of just ";" as specified by
https://developer.mozilla.org/En/NP_GetMIMEDescription... Nicolas will report that bug, but let's have the workaround
anyway, can't hurt.
BUG: 178562


 M  +1 -1      pluginscan.cpp  


WebSVN link: http://websvn.kde.org/?view=rev&revision=901122