voyager.server crash

I’ve been having trouble with voyager.server crashing after
about 60,000 page reloads (QNX 4.25, Voyager 2.02). It seems
to slow down after about 30,000 reloads, and keeps slowing
until it finally crashes, leaving the voyager client locked

The memory usage also grows steadily (about 500KB/hour) before it
crashes. But a memory leak on an embedded system? Nah… :slight_smile:

Has anyone else had similar trouble? (Or does anyone know
of another browser available for QNX 4.25?)

We’ve created a small set of test pages that show the problem
after about 5 hours (see below). Put all three files in a
directory and point voyager at “crash.htm”.

I’d certainly appreciate any feedback anyone can provide about
what might be happening, or how we might avoid it. Our only
work-around right now is to slay/restart voyager.

  • Larry


    — crash.htm —

— crashform.htm –

V'ger Crash test:


Outer loop:
Inner loop:
( Inner loop triggers Outer loop every: Max: )

— crashloop.htm —

timeout

I would have to say that you are quite lucky. I find the QNX 4
voyager locks up a lot. If I’m running both Vmail and Voyager
the likelyhood increases dramatically. It all seems data dependent
however. Some web sites seem to have stuff that throws it for
a loop. I don’t think there’s much point in dicussing it since
there seems to be ZERO support for this product now.


Previously, Larry Martin wrote in qdn.public.qnx4.photon:

I’ve been having trouble with voyager.server crashing after
about 60,000 page reloads (QNX 4.25, Voyager 2.02). It seems
to slow down after about 30,000 reloads, and keeps slowing
until it finally crashes, leaving the voyager client locked

The memory usage also grows steadily (about 500KB/hour) before it
crashes. But a memory leak on an embedded system? Nah… > :slight_smile:

Has anyone else had similar trouble? (Or does anyone know
of another browser available for QNX 4.25?)

We’ve created a small set of test pages that show the problem
after about 5 hours (see below). Put all three files in a
directory and point voyager at “crash.htm”.

I’d certainly appreciate any feedback anyone can provide about
what might be happening, or how we might avoid it. Our only
work-around right now is to slay/restart voyager.

  • Larry


    — crash.htm —
    html
    head></head
    frameset frameborder=‘1’ framespacing=‘0’ rows=‘64,*’
    frame scrolling=‘no’ name=‘loop’ src=‘crashloop.htm’
    frame scrolling=‘yes’ name=‘disp’ src=‘crashform.htm’
    /frameset
    /html

— crashform.htm –
HTML
head
title></title
/head
body
h3
V’ger Crash test:
/h3
form name=‘myform’ method=‘get’
br
Outer loop: <input name=‘counterouter’ type=‘text’ size=10 maxlength=10
value=’’
br
Inner loop: <input name=‘counterinner’ type=‘text’ size=10 maxlength=10
value=’’
br
( Inner loop triggers Outer loop every:
Max: <input name=‘countermax’ type=‘text’ size=10 maxlength=10
value=‘1000000’
)
/form
/body
/HTML

— crashloop.htm —
html
head
title></title
/head
body
timeout
script type=‘text/javascript’ language=‘JavaScript’><!–
// ------------------------------------------------------
var timeout = 1000; // every second
timeout = 50;
function myreload() {
// make sure the form is there:
if( parent.disp.document )
if( parent.disp.document.myform )
if( parent.disp.document.myform.countermax )
{
var cn = parent.disp.document.myform.countermax.value;
var ci = parent.disp.document.myform.counterinner.value;
if( ++ci == cn )
{
var cj =
parent.disp.document.myform.counterouter.value;
parent.disp.document.myform.counterouter.value = ++cj;
ci = 0;
}
parent.disp.document.myform.counterinner.value = ci;
}
location.replace( ‘crashloop.htm’ );
}
setTimeout( ‘myreload()’, timeout );
// ------------------------------------------------------
//–></script
/body
/HTML


Mitchell Schoenbrun --------- maschoen@pobox.com

Mitchell Schoenbrun <maschoen@pobox.com> wrote in
news:Voyager.020109164158.4232C@node1:

I would have to say that you are quite lucky. I find the QNX 4
voyager locks up a lot. If I’m running both Vmail and Voyager
the likelyhood increases dramatically. It all seems data dependent
however. Some web sites seem to have stuff that throws it for
a loop. I don’t think there’s much point in dicussing it since
there seems to be ZERO support for this product now.

Well, thanks for the insight into QNX support for this product.
We just replicated the crash on the RTP 6 download, so maybe I
can find out something from that angle.

  • Larry

I believe the support staff is currently looking into this problem and I
have been looking at voyager. We will get back to you ASAP about this
issue.

Brian

Larry Martin wrote:

Mitchell Schoenbrun <> maschoen@pobox.com> > wrote in
news:Voyager.020109164158.4232C@node1:


I would have to say that you are quite lucky. I find the QNX 4
voyager locks up a lot. If I’m running both Vmail and Voyager
the likelyhood increases dramatically. It all seems data dependent
however. Some web sites seem to have stuff that throws it for
a loop. I don’t think there’s much point in dicussing it since
there seems to be ZERO support for this product now.


Well, thanks for the insight into QNX support for this product.
We just replicated the crash on the RTP 6 download, so maybe I
can find out something from that angle.

  • Larry

Great, and thanks for your help!

  • Larry

Brian <briane@qnx.com> wrote in news:3C446502.7060309@qnx.com:

I believe the support staff is currently looking into this problem and I
have been looking at voyager. We will get back to you ASAP about this
issue.

Brian

Larry Martin wrote:

Mitchell Schoenbrun <> maschoen@pobox.com> > wrote in
news:Voyager.020109164158.4232C@node1:


I would have to say that you are quite lucky. I find the QNX 4
voyager locks up a lot. If I’m running both Vmail and Voyager
the likelyhood increases dramatically. It all seems data dependent
however. Some web sites seem to have stuff that throws it for
a loop. I don’t think there’s much point in dicussing it since
there seems to be ZERO support for this product now.


Well, thanks for the insight into QNX support for this product.
We just replicated the crash on the RTP 6 download, so maybe I
can find out something from that angle.

  • Larry

    \