Re: [Evolution] (no subject)

[Ronald Kuetemeier <ronald kuetemeier com>]

On Wed, 2002-10-02 at 07:31, Dan Winship wrote:
> On Wed, 2002-10-02 at 03:10, Jason Grimsman wrote:
> > I recently decided to run the 2.5.x kernel, and after installing it, 
> > evolution would freeze whenever i hit the reply buttons, and usually when i 
> > hit the "new" button...It also happened when i would try to view the 
> > calendar or contacts.
>
> That sounds like a problem with oaf activation. (When you open the
> composer it needs to activate the HTML editor control, and when you open
> calendar or contacts it needs to start the calendar/contact backend.)
>
> It's impossible to guess what the exact problem is though. In general,
> people have not had much luck running evolution with unstable kernels in
> the past. Unless you want to sit down with gdb and find the problem
> yourself, your best bet is to wait until it's closer to being stable.

Actually it seems to be a problem with glib. I did a gdb and strace both
pointing to glib-1.2. I don't know if it even makes sense to fix 1.2, I
use 2.0 already just evolution is using 1.2. So we will know when
evolution is ported to 2.0 if the problem still exists and fix it in
2.0.  Or if linux-2.6 comes out before evolution is ported to 2.0, well
somebody has to dig deeper.
Here is the strace section with kernel 2.4.19:

open("/dev/urandom", O_RDONLY)          = 19
read(19, "\231\235\342\206\351\r\241\"", 8) = 8
close(19)                               = 0
open("/dev/urandom", O_RDONLY)          = 19
read(19, "\t.\207\354\3168\f\261", 8)   = 8
close(19)                               = 0
open("/dev/urandom", O_RDONLY)          = 19
read(19, "ee\325i\301\267\202\277", 8)  = 8
close(19)                               = 0
socket(PF_UNIX, SOCK_STREAM, 0)         = 19
connect(19, {sin_family=AF_UNIX,
path="/tmp/orbit-ronald/orb-19562820742091404245"}, 44) = 0
fcntl64(19, F_SETFD, FD_CLOEXEC)        = 0
fcntl64(19, F_GETFL)                    = 0x2 (flags O_RDWR)
fcntl64(19, F_SETFL, O_RDWR|O_NONBLOCK) = 0
writev(19, [{"GIOP\1\0\1", 7}, {"\0<\0\0\0", 5}, {"\0\0\0\0", 4},
{"\200\364\377\277\1", 5}, {"\0\0\0", 3},
{"\33\0\0\0\1\0\0\0ee\325i\301\267\202\277\365\377\377\377"..., 31},
{"\0", 1}, {"\10\0\0\0getType\0", 12}, {"\0\0\0\0", 4}], 9) = 72
select(20, [8 11 12 13 16 17 18 19], NULL, [8 11 12 13 16 17 18 19],
NULL) = 1 (in [8])
accept(8, {sin_family=AF_UNIX, path="
Xxë øóÿ¿0±&@0±&@<80>!ôÿ¿R<83>&@<80>!^"}, [2]) = 20
 


Here is the strace section with kernel 2.5.40:

open("/dev/urandom", O_RDONLY)          = 19
read(19, "\205\22\311\202\6\356\207\353", 8) = 8
close(19)                               = 0
open("/dev/urandom", O_RDONLY)          = 19
read(19, "\0067\243\274\374\207\226r", 8) = 8
close(19)                               = 0
open("/dev/urandom", O_RDONLY)          = 19
read(19, "\251\242\34K\23\316\301\350", 8) = 8
close(19)                               = 0
writev(18, [{"GIOP\1\0\1", 7}, {"\0<\0\0\0", 5}, {"\0\0\0\0", 4},
{"\200\364\377\277\1", 5}, {"\0\0\0", 3},
{"\33\0\0\0\1\0\0\0\251\242\34K\23\316\301\350\365\377\377"..., 31},
{"\0", 1}, {"\10\0\0\0getType\0", 12}, {"\0\0\0\0", 4}], 9) = 72
select(19, [8 11 12 13 16 17 18], NULL, [8 11 12 13 16 17 18], NULL
<unfinished ...>

Notice there is a socket missing, and therefore evolution waits on
select:

socket(PF_UNIX, SOCK_STREAM, 0)         = 19
connect(19, {sin_family=AF_UNIX,
path="/tmp/orbit-ronald/orb-19562820742091404245"}, 44) = 0

So somewhere after reading /dev/urandom something goes wrong. I know
there where changes to urandom, but have no idea how they can effect
user programs.

Happy debug,
Ronald



> -- Dan
>
>
> _______________________________________________
> evolution maillist  -  evolution ximian com
> http://lists.ximian.com/mailman/listinfo/evolution