Re: [Evolution-hackers] Bonobo activation fails... gconf/schema related?

[Jules Colding <colding omesc com>]

OK, top-posting but... this has me completely baffled:

I could naturally not start Evolution 2.4 from  the command line so I
though that I'd try the script from go-evolution, even though it should
do absolutely nothing differently from what I did from the terminal.

Lo and behold, Evolution started.

In a perplexed mood I started trimming the script to spot the
difference, if any. I ended up with an empty script which just invoked
"/home/evo/opt/bin/evolution-2.4" and still managed to start Evolution.
Mystified I tried the command line again.

Lo and behold, Evolution started.

I have no idea what is, or has been, going on. It is really a big
mystery to me and it does leave me somewhat anxious I must say. 

Best regards,
  jules



On Thu, 2005-09-01 at 09:49 +0200, Jules Colding wrote:
> hi Michael,
> 
> On Thu, 2005-09-01 at 11:27 +0800, Not Zed wrote:
> > No, the schema warnings - like gconf in general - are just rubbish.  It
> > never checks the schemas anyway, they are just used for some stupid
> > defaults which should be in the code instead, and the ones it will be
> > finding from the system install will work fine enough (and nothing
> > important is looked up at this stage either).
> 
> OK. Good to know.
> 
> > All i can think of is some vesion mismatch somewhere.  Make sure you
> > clear out any eds/camel related libraries from your install directory -
> > the versioning is so broken you can't have previous installs lying
> > around.
> 
> Yes, I am right now deleting ~/opt and rebuilding everything from
> scratch to be sure nothing is left.
> 
> > Try using strace perhaps to see what it's trying to load, use "strace -f
> > -o trace evolutio-2.4" then quit and use a viewer to look at 'trace',
> > since it will generate gobs of stuff.  Grep for things like opening
> > libevolution-mail.so, 
> 
> It doesn't do that. I am only seeing libevolution-a11y.so.0 and
> libevolution-widgets-a11y.so.0 being successfully opened. No other
> libevolution*.so.* are attempted to be opened.
> 
> > ".server" files, etc.
> 
> No ".server" files are opened. It doesn't even try. Actually, I can't
> see that any open() calls fails except for some X-related theme files,
> cursors and such.
> 
> But it does, funny enough (or not...), try to kill
> evolution-alarm-notify from Evolution 2.2. This is naturally not
> permitted as the process is owned by another user ("colding"). I am
> trying to start Evolution as user "evo".
> 
> Evolution eventually gets a SIGPIPE which I think is what kills it,
> although it doesn't seem related to the attempted kill on my alarm
> process. But what raises the signal?
> 
> Why does it try to kill the wrong alarm process and do you think this
> could be why it can't start?
> 
> Thanks a lot,
>   jules
> 
> 
> Full trace at <http://www.omesc.com/content/downloads/evo-trace>.
> 
> Excerpts from trace (bonobo-slay has been run first):
> 
> 
> 16458 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2aaaaaac1000
> 16458 read(3, "7785 (evolution-alarm) S 1 7785 "..., 1024) = 243
> .
> .
> .
> 16464 execve("/usr/bin/killall", ["/usr/bin/killall", "-9", "lt-evolution-alarm-notify"], [/* 61 vars */]) = 0
> 16464 uname({sys="Linux", node="omc-2", ...}) = 0
> 16464 brk(0)                            = 0x504000
> .
> .
> .
> 16464 write(2, "lt-evolution-alarm-notify: no pr"..., 45) = 45
> 16464 exit_group(1)                     = ?
> 16463 <... wait4 resumed> [{WIFEXITED(s) && WEXITSTATUS(s) == 1}], 0, NULL) = 16464
> 16463 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
> 16463 --- SIGCHLD (Child exited) @ 0 (0) ---
> 16463 wait4(-1, 0x7fffffffd624, WNOHANG, NULL) = -1 ECHILD (No child processes)
> 16463 rt_sigreturn(0xffffffffffffffff)  = 0
> 16463 rt_sigaction(SIGINT, {SIG_DFL}, {0x4335d0, [], SA_RESTORER, 0x2aaaaacf3890}, 8) = 0
> 16463 exit_group(1)                     = ?
> 16453 <... wait4 resumed> [{WIFEXITED(s) && WEXITSTATUS(s) == 1}], 0, NULL) = 16463
> 16453 rt_sigaction(SIGINT, {SIG_DFL}, NULL, 8) = 0
> 16453 rt_sigaction(SIGQUIT, {SIG_DFL}, NULL, 8) = 0
> 16453 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
> 16453 --- SIGCHLD (Child exited) @ 0 (0) ---
> 16453 rt_sigreturn(0x561200)            = 0
> 16453 rt_sigaction(SIGINT, {SIG_IGN}, {SIG_DFL}, 8) = 0
> 16453 rt_sigaction(SIGQUIT, {SIG_IGN}, {SIG_DFL}, 8) = 0
> 16453 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
> 16453 clone(child_stack=0, flags=CLONE_PARENT_SETTID|SIGCHLD, parent_tidptr=0x7fffffffd83c) = 16465
> 16453 wait4(16465,  <unfinished ...>
> 16465 rt_sigaction(SIGINT, {SIG_DFL}, NULL, 8) = 0
> 16465 rt_sigaction(SIGQUIT, {SIG_DFL}, NULL, 8) = 0
> 16465 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
> 16465 execve("/bin/sh", ["sh", "-c", "/usr/bin/killall -9 evolution-al"...], [/* 62 vars */]) = 0
> 16465 uname({sys="Linux", node="omc-2", ...}) = 0
> 16465 brk(0)                            = 0x5c4000
> .
> .
> .
> 16466 kill(7785, SIGKILL)               = -1 EPERM (Operation not permitted)
> 16466 write(2, "evolution-alarm-notify(7785): Op"..., 54) = 54
> .
> .
> .
> 
> 16466 write(2, "evolution-alarm-notify: no proce"..., 42) = 42
> 16466 exit_group(1)                     = ?
> 16465 <... wait4 resumed> [{WIFEXITED(s) && WEXITSTATUS(s) == 1}], 0, NULL) = 16466
> 16465 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
> 16465 --- SIGCHLD (Child exited) @ 0 (0) ---
> 16465 wait4(-1, 0x7fffffffd624, WNOHANG, NULL) = -1 ECHILD (No child processes)
> 16465 rt_sigreturn(0xffffffffffffffff)  = 0
> 16465 rt_sigaction(SIGINT, {SIG_DFL}, {0x4335d0, [], SA_RESTORER, 0x2aaaaacf3890}, 8) = 0
> 16465 exit_group(1)                     = ?
> 16453 <... wait4 resumed> [{WIFEXITED(s) && WEXITSTATUS(s) == 1}], 0, NULL) = 16465
> 16453 rt_sigaction(SIGINT, {SIG_DFL}, NULL, 8) = 0
> 16453 rt_sigaction(SIGQUIT, {SIG_DFL}, NULL, 8) = 0
> 16453 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
> 16453 --- SIGCHLD (Child exited) @ 0 (0) ---
> 16453 rt_sigreturn(0x561200)            = 0
> 16453 writev(18, [{"GIOP\1\2\1\5\0\0\0\0", 12}], 1) = -1 EPIPE (Broken pipe)
> 16453 --- SIGPIPE (Broken pipe) @ 0 (0) ---
> 16453 close(18)                         = 0
> 
> 
> 
> 
> _______________________________________________
> Evolution-hackers mailing list
> Evolution-hackers gnome org
> http://mail.gnome.org/mailman/listinfo/evolution-hackers