Xphoton / gtwm crash

QNX Newsgroups (Archives) qdn.public.qnxrtp.photon
1

I run Xphoton and gtwm as:

Xphoton -nocrc &
gtwm &

and I use the Env var:
DISPLAY=127.1:0


gtwm crashes under moderate use (usually on closing a window) with the
following cryptic (encrypted?) message (messages?):


XIO: fatal IO error 254 (Connection reset by peer) on X server “127.1:0.0”
after 4112 requests (e4111r rkonro wonp epnrioncge ssseecdu)r iwtiyt hp o8 l
iecvye nftisl er e/usr/X11R6/lib/X11/xserver/SecurityPolicym
aining.


-david

2

Hello David,

I’ve talked to one of our developers and here is what he had to say:

Xphoton intercepts the pwm CLOSE message and
sends a WM_DELETE_WINDOW to the X app.
If you run an X app that ignores it, the app
won’t close. However, if you select Close
again (pwm menu or top right button) then
pwm sends a SIGHUP to the app that owns the
window. Xphoton owns the window, so it gets
the signal and exits.

Since the user is running without -once the
server resets (ie. rips down all the clients
and starts over) The message printed out is
a client getting tore down and the server starting
up again.

I hope that this helps
Rodney

David S. Alessio <david@nosysrtimespam.com> wrote:

I run Xphoton and gtwm as:

Xphoton -nocrc &
gtwm &

and I use the Env var:
DISPLAY=127.1:0



gtwm crashes under moderate use (usually on closing a window) with the
following cryptic (encrypted?) message (messages?):



XIO: fatal IO error 254 (Connection reset by peer) on X server “127.1:0.0”
after 4112 requests (e4111r rkonro wonp epnrioncge ssseecdu)r iwtiyt hp o8 l
iecvye nftisl er e/usr/X11R6/lib/X11/xserver/SecurityPolicym
aining.

-david

3

Thanks for explaining this behavior.

I should note that Xphoton often exits with Memory fault. Usually when
moving a window which overlaps an other X window. I use several X apps,
but this is most consistent with tkCVS 7.0.1 (TCL8.3.2, tk8.3.2).

-david



Gui Group wrote:

Hello David,

I’ve talked to one of our developers and here is what he had to say:

Xphoton intercepts the pwm CLOSE message and
sends a WM_DELETE_WINDOW to the X app.
If you run an X app that ignores it, the app
won’t close. However, if you select Close
again (pwm menu or top right button) then
pwm sends a SIGHUP to the app that owns the
window. Xphoton owns the window, so it gets
the signal and exits.

Since the user is running without -once the
server resets (ie. rips down all the clients
and starts over) The message printed out is
a client getting tore down and the server starting
up again.

I hope that this helps
Rodney

David S. Alessio <> david@nosysrtimespam.com> > wrote:
I run Xphoton and gtwm as:

Xphoton -nocrc &
gtwm &

and I use the Env var:
DISPLAY=127.1:0

gtwm crashes under moderate use (usually on closing a window) with the
following cryptic (encrypted?) message (messages?):

XIO: fatal IO error 254 (Connection reset by peer) on X server “127.1:0.0”
after 4112 requests (e4111r rkonro wonp epnrioncge ssseecdu)r iwtiyt hp o8 l
iecvye nftisl er e/usr/X11R6/lib/X11/xserver/SecurityPolicym
aining.

-david

4

Hey David

Could you send me a core file and a reproducable test case.

Thanks
Rodney

David Alessio <david.alessio@hsa.hitachi.com> wrote:

Thanks for explaining this behavior.

I should note that Xphoton often exits with Memory fault. Usually when
moving a window which overlaps an other X window. I use several X apps,
but this is most consistent with tkCVS 7.0.1 (TCL8.3.2, tk8.3.2).

-david



Gui Group wrote:

Hello David,

I’ve talked to one of our developers and here is what he had to say:

Xphoton intercepts the pwm CLOSE message and
sends a WM_DELETE_WINDOW to the X app.
If you run an X app that ignores it, the app
won’t close. However, if you select Close
again (pwm menu or top right button) then
pwm sends a SIGHUP to the app that owns the
window. Xphoton owns the window, so it gets
the signal and exits.

Since the user is running without -once the
server resets (ie. rips down all the clients
and starts over) The message printed out is
a client getting tore down and the server starting
up again.

I hope that this helps
Rodney

David S. Alessio <> david@nosysrtimespam.com> > wrote:
I run Xphoton and gtwm as:

Xphoton -nocrc &
gtwm &

and I use the Env var:
DISPLAY=127.1:0

gtwm crashes under moderate use (usually on closing a window) with the
following cryptic (encrypted?) message (messages?):

XIO: fatal IO error 254 (Connection reset by peer) on X server “127.1:0.0”
after 4112 requests (e4111r rkonro wonp epnrioncge ssseecdu)r iwtiyt hp o8 l
iecvye nftisl er e/usr/X11R6/lib/X11/xserver/SecurityPolicym
aining.

-david

5

Sure! It’s a 1.3MB file (bz2’d)

I uploaded it to Hitachi’s FTP server. You can get it with anonymous
login – just don’t expect list privleges:
ftp.hsa.hitachi.com
/pub/outgoing/Xphoton.core.bz2

md5sum: 93ae3f307e47dc3b5ace4a1ad512e888 Xphoton.core.bz2

Crash:
Xphoton -nocrc &
gtwm -N &
Bring up two X apps or windows and let them overlap. Move the top of
the two around over the other. It depends on the size of the two, but
if you move at a normal to fast rate Xphoton will crash. I have no idea
what the code look like, but it seems to me that a queue of event is
being overflowed – requests to move are coming in faster than the moves
occur…

Cheers,
-david




Gui Group wrote:

Hey David

Could you send me a core file and a reproducable test case.

Thanks
Rodney

David Alessio <> david.alessio@hsa.hitachi.com> > wrote:

Thanks for explaining this behavior.

I should note that Xphoton often exits with Memory fault. Usually when
moving a window which overlaps an other X window. I use several X apps,
but this is most consistent with tkCVS 7.0.1 (TCL8.3.2, tk8.3.2).

6

Hey David

I have the file. I’ll submit a bug on this issue

Thanks
Rodney

David Alessio <david.alessio@hsa.hitachi.com> wrote:

Sure! It’s a 1.3MB file (bz2’d)

I uploaded it to Hitachi’s FTP server. You can get it with anonymous
login – just don’t expect list privleges:
ftp.hsa.hitachi.com
/pub/outgoing/Xphoton.core.bz2

md5sum: 93ae3f307e47dc3b5ace4a1ad512e888 Xphoton.core.bz2

Crash:
Xphoton -nocrc &
gtwm -N &
Bring up two X apps or windows and let them overlap. Move the top of
the two around over the other. It depends on the size of the two, but
if you move at a normal to fast rate Xphoton will crash. I have no idea
what the code look like, but it seems to me that a queue of event is
being overflowed – requests to move are coming in faster than the moves
occur…

Cheers,
-david



Gui Group wrote:

Hey David

Could you send me a core file and a reproducable test case.

Thanks
Rodney

David Alessio <> david.alessio@hsa.hitachi.com> > wrote:

Thanks for explaining this behavior.

I should note that Xphoton often exits with Memory fault. Usually when
moving a window which overlaps an other X window. I use several X apps,
but this is most consistent with tkCVS 7.0.1 (TCL8.3.2, tk8.3.2).

7

Great. Thanks.

Would you mind posting the PR?

Cheers,
-david




Gui Group wrote:

Hey David

I have the file. I’ll submit a bug on this issue

Thanks
Rodney

8

Hello David

I know that I have a core file, but when you say crash, you mean that you
get a SIGSEGV signal to the Xphoton ( It stops running all togehter), or
do you me that the windows stop drawing ( The windows are still there, but
they don’t update there contents) ?

Thanks
Rodney

David Alessio <david.alessio@hsa.hitachi.com> wrote:

Great. Thanks.

Would you mind posting the PR?

Cheers,
-david



Gui Group wrote:

Hey David

I have the file. I’ll submit a bug on this issue

Thanks
Rodney

9

SIGSEGV


Gui Group wrote:

Hello David

I know that I have a core file, but when you say crash, you mean that you
get a SIGSEGV signal to the Xphoton ( It stops running all togehter), or
do you me that the windows stop drawing ( The windows are still there, but
they don’t update there contents) ?

Thanks
Rodney

David Alessio <> david.alessio@hsa.hitachi.com> > wrote:

Great. Thanks.

Would you mind posting the PR?

Cheers,
-david

Gui Group wrote:

Hey David

I have the file. I’ll submit a bug on this issue

Thanks
Rodney