Hi Simon

On 20 March 2012 19:35, Simon King <[email protected]> wrote:

> Do I understand correctly: You say that probably the crash is related with
> a
> general problem of Java on Debian/Ubuntu?
>

No, sorry. I was speculating that in future the move from sun-java to
openjdk
on Debian/Ubuntu will cause even more problems with jmol.

However, I just now installed openjdk-7-plugin on Ubuntu 12.04 and the
applet
does similar to natty: it says in the jmol box: Loading jmol applet, and
counts up
the seconds (at 130s now). So it is not worse, it is the same. I'm not
going to
try and reload it several times to reproduce the crash right now.


My bad though; I logged into the natty box that I used earlier today (where
I got
two crashes and two delays like above) and I see that both sun-java and
openjdk is installed, so now I'm no longer sure that was indeed sun-java.
I will check tomorrow as this link is to slow for X forwarding.

That doesn't convince me, because of the following:


> The example does work with sage-4.6.2 and with sage-5.0.beta8 (built with
> my
> system-wide gcc) on the *same* machine where the crash has occurred. Also
> it
> does work with sage-5.0.beta7 on an openSUSE 12.1 laptop (again with
> system-wide
> gcc).
>
> So, I think the only difference between crash and non-crash is whether the
> gcc
> spkg or a system-wide gcc was used.
>
> I'm afraid that right now I can not test whether the crash is
> reproducible, but
> I will try tomorrow, and then post on #12369.
>

The first sage today on natty was built with a system-wide gcc, the default
version on Ubuntu 11.04. So your experience and mine with system-wide gcc
is opposites: yours works, mine crashes.

The other test I am doing now, is the prebuilt binary from sagemath.org for
10.04 worked into a deb
for 12.04. I don't know which gcc was used for that.

All my sages are system-wide gcc, and I get the crash and/or the delay,
I think randomly. I.e. the same natty machine this morning, 4 times in a
row,
gave different results.

Do you use openjdk or sun-java or another java?


UPDATE: now this is odd. On the Ubuntu 12.04 box with the prebuilt binary
from sagemath.org for 10.04 packaged into a deb:

Jmol was at 610seconds still loading: I clicked evaluate again, and I got
an image: it worked.
A few times. Then I closed and opened the firefox tab, and firefox crashed.
Then I reopened firefox, and the image opened properly again. Then I closed
the firefox tab, restarted sage, and clicked on the worksheet, and the jmol
box
opened and the image showed. Now every time I click evaluate it loads.


In short, random, bad results. I guess we need some better test cases for
this?
And I think you need to run it several times, evaluating, closeing and
opening
worksheet, firefox tab, and restarting sage, to get a sample of the real
behaviour.

I should also do these tests only on the sage-from-source on ubuntu11.04 at
work,
where I can test different versions of sage and gcc and java, but one
version of firefox/ubuntu.
Not my development versions of both sage+ubuntu12.04 at home.

Regards,
Jan
-- 
  .~.
  /V\     Jan Groenewald
 /( )\    www.aims.ac.za
 ^^-^^

-- 
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/sage-support
URL: http://www.sagemath.org

Reply via email to