1 |
Thanks for the reply. |
2 |
|
3 |
strace showed very detailed the last syscalls before the SIGSEV: |
4 |
|
5 |
[pid 21801] sendmsg(9, {msg_name(0)=NULL, |
6 |
msg_iov(2)=[{"l\4\1\1\0\0\0\0\3\0\0\0X\0\0\0\1\1o\0\23\0\0\0/StatusNotifierItem\0\0\0\0\0\2\1s\0\32\0\0\0org.kde.StatusNotifierItem\0\0\0 |
7 |
\0\0\0\3\1s\0\7\0\0\0NewIcon\0", 104}, {"", 0}], msg_controllen=0, |
8 |
msg_flags=0}, MSG_NOSIGNAL) = 104 |
9 |
[pid 21801] --- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_MAPERR, si_addr=0x8} |
10 |
--- |
11 |
|
12 |
and file descriptor 9 was connected to dbus. Well, but this did not show WHERE |
13 |
it crashed. |
14 |
|
15 |
Next try was ltrace, but ltrace cannot be applied on python scripts (no ELF). |
16 |
|
17 |
Finally, I checked the dependencies, and starting with pykde4 seemed to be a |
18 |
good idea. Recompiled it with "USE=debug" and "CXXFLAGS=-O2 -g", and voila, |
19 |
synaptiks worked without a problem. Hmmm. |
20 |
|
21 |
So, recompiled pykde with the normal options, and it still worked. |
22 |
Very strange, since I ran revdep-rebuild and pyhton-updater after the world |
23 |
upgrade ... |
24 |
Note@self: Maybe updating/re-emerging PyQt4 and pykde4 is worth the time, even |
25 |
if its not neccessary. |
26 |
|
27 |
However, it seems to work now. |
28 |
|
29 |
Best regards |
30 |
Alex |
31 |
|
32 |
|
33 |
|
34 |
|
35 |
|
36 |
Am Samstag, 22. März 2014, 17:19:14 schrieb Tom Wijsman: |
37 |
> On Sat, 22 Mar 2014 16:47:23 +0100 |
38 |
> |
39 |
> Alexander Puchmayr <alexander.puchmayr@×××××××.at> wrote: |
40 |
> > Symptom: |
41 |
> > Start synaptiks configuration module, get immedeately DrKonqi crash |
42 |
> > report (trace attached, likely not much useful since I disabled debug |
43 |
> > flags) Start either from kde systemsettings or konsole; the latter |
44 |
> > gives |
45 |
> > |
46 |
> > alex@prometheus /datadisk/home/alex $ synaptiks |
47 |
> > Fontconfig warning: "/etc/fonts/conf.d/50-user.conf", line 14: |
48 |
> > reading configurations from ~/.fonts.conf is deprecated. |
49 |
> > KCrash: Application 'synaptiks' crashing... |
50 |
> > KCrash: Attempting to start /usr/lib64/kde4/libexec/drkonqi from |
51 |
> > kdeinit sock_file=/homedisk/alex/.kde4/socket-prometheus/kdeinit4__0 |
52 |
> > unnamed app(17763): Communication problem with "synaptiks" , it |
53 |
> > probably crashed. |
54 |
> > Error message was: "org.freedesktop.DBus.Error.NoReply" : " "Message |
55 |
> > did not receive a reply (timeout by message bus)" " |
56 |
> > |
57 |
> > Any ideas what could go wrong? |
58 |
> |
59 |
> The crash would reveal what is wrong; therefore, I suggest you to |
60 |
> recompile the libraries of those packages with debug flags. |
61 |
> |
62 |
> See http://wiki.gentoo.org/wiki/Project:Quality_Assurance/Backtraces |
63 |
> |
64 |
> Besides that, you might be able to use strace and/or ltrace (with -f |
65 |
> parameter; if needed, increase the -s parameter) to see what happens |
66 |
> under the hood. |
67 |
> |
68 |
> I'm suspecting you'll find the cause by inspecting both; because as |
69 |
> the trace mixes in evaluation of a Python script, you might need to use |
70 |
> strace and/or ltrace to discover which Python script this is. |
71 |
> |
72 |
> A next step could be to use a debugger on that Python script if the |
73 |
> trace, strace and/or ltrace are insufficient to reveal the cause. |
74 |
> |
75 |
> When you have found the cause, can you file the details in a bug at |
76 |
> https://bugs.gentoo.org such that the maintainers are aware of this? |
77 |
> |
78 |
> Thank you very much in advance. |
79 |
-- |
80 |
Dipl. Ing. Alexander Puchmayr |
81 |
Grundbachstraße 36 |
82 |
4060 Leonding, Österreich |