IE cannot open the Internet site ... Operation aborted

[Mika]

Hello, we understand you guys may be able to help.

We have a page which has been working great for over a year and gets many
hits. However recently something got changed that we cannot seem to find,
and now *sometimes* if you refresh the page (generally while it is still
loading) in IE7, we get the popup window error:

Internet Explorer cannot open the Internet site...
Operation aborted

Here is an example of the page in question:
http://tinyurl.com/35mwxr [broadband recommended]

We appreciate you may have other comments on the site such as the size of
the pages, however we are not looking to change that at present, with faster
broadband becoming more abundant. Search engine listings alert the user
that broadband is recommended for this technology.

As we said, the site worked great, but for this unknown reason now needs
some sort of a tweak to fix. The changes we made before it happened were
related to the body tag but they have been completely undone yet the issue
remains :-S

Thanks in advance to anyone able to find the cure for us to test.

Regards,
Mika @ SHS.com

 

[Stevo]

Internet Explorer cannot open the Internet site...
Operation aborted

That happens on IE when you try to add items to the DOM before the page
is complete. Wait for onload event to fire before doing it, or poll the
document.readyState waiting for "complete" (although "interactive" may
sometimes be good enough).

 

[Bone Ur]

Well bust mah britches and call me cheeky, on Fri, 16 Nov 2007 18:12:37
GMT Mika scribed:

Hello, we understand you guys may be able to help.

We have a page which has been working great for over a year and gets
many hits. However recently something got changed that we cannot seem
to find, and now *sometimes* if you refresh the page (generally while
it is still loading) in IE7, we get the popup window error:

Internet Explorer cannot open the Internet site...
Operation aborted

Here is an example of the page in question:
http://tinyurl.com/35mwxr [broadband recommended]

We appreciate you may have other comments on the site such as the size
of the pages, however we are not looking to change that at present,
with faster broadband becoming more abundant. Search engine listings
alert the user that broadband is recommended for this technology.

As we said, the site worked great, but for this unknown reason now
needs some sort of a tweak to fix. The changes we made before it
happened were related to the body tag but they have been completely
undone yet the issue remains :-S

Thanks in advance to anyone able to find the cure for us to test.

I did a refresh-while-loading in Opera and got tons of errors, probably
related to the javascript (...apparently being interrupted). Oth, a
refresh-post-load produced no errors in the same browser. Ergo, I suspect
that the abundant j/s is causing the problem. Suggest you try ie7 without
the j/s on.

 

[Mika]

Well bust mah britches and call me cheeky, on Fri, 16 Nov 2007 18:12:37
GMT Mika scribed:

Hello, we understand you guys may be able to help.

We have a page which has been working great for over a year and gets
many hits. However recently something got changed that we cannot seem
to find, and now *sometimes* if you refresh the page (generally while
it is still loading) in IE7, we get the popup window error:

Internet Explorer cannot open the Internet site...
Operation aborted

Here is an example of the page in question:
http://tinyurl.com/35mwxr [broadband recommended]

We appreciate you may have other comments on the site such as the size
of the pages, however we are not looking to change that at present,
with faster broadband becoming more abundant. Search engine listings
alert the user that broadband is recommended for this technology.

As we said, the site worked great, but for this unknown reason now
needs some sort of a tweak to fix. The changes we made before it
happened were related to the body tag but they have been completely
undone yet the issue remains :-S

Thanks in advance to anyone able to find the cure for us to test.

I did a refresh-while-loading in Opera and got tons of errors, probably
related to the javascript (...apparently being interrupted). Oth, a
refresh-post-load produced no errors in the same browser. Ergo, I suspect
that the abundant j/s is causing the problem. Suggest you try ie7 without
the j/s on.

Thanks Bone, however as j/s in crucial we wonder if there is a workaround
you can think of to this?

 

[Mika]

That happens on IE when you try to add items to the DOM before the page is
complete. Wait for onload event to fire before doing it, or poll the
document.readyState waiting for "complete" (although "interactive" may
sometimes be good enough).

Thanks Stevo.

Can you elaborate on how we could make the Document Object Model only happen
after onload, or better still implement your polling suggestion.

I am not one of the programmers myself, but would appreciate it if you could
briefly explain in layman's terms what should be done. I.e. "Find and
change <body onload=... to xxxx", so we can try it here. Many thanks.

 

[Mika]

Well bust mah britches and call me cheeky, on Fri, 16 Nov 2007 18:12:37
GMT Mika scribed:

Hello, we understand you guys may be able to help.

We have a page which has been working great for over a year and gets
many hits. However recently something got changed that we cannot seem
to find, and now *sometimes* if you refresh the page (generally while
it is still loading) in IE7, we get the popup window error:

Internet Explorer cannot open the Internet site...
Operation aborted

Here is an example of the page in question:
http://tinyurl.com/35mwxr [broadband recommended]

We appreciate you may have other comments on the site such as the size
of the pages, however we are not looking to change that at present,
with faster broadband becoming more abundant. Search engine listings
alert the user that broadband is recommended for this technology.

As we said, the site worked great, but for this unknown reason now
needs some sort of a tweak to fix. The changes we made before it
happened were related to the body tag but they have been completely
undone yet the issue remains :-S

Thanks in advance to anyone able to find the cure for us to test.

I did a refresh-while-loading in Opera and got tons of errors, probably
related to the javascript (...apparently being interrupted). Oth, a
refresh-post-load produced no errors in the same browser. Ergo, I suspect
that the abundant j/s is causing the problem. Suggest you try ie7 without
the j/s on.

PS: Due to Opera functionality limitations (partly on the width of
streetscape pages) that browser is not supported by our site, although what
you found out is interesting.

 

[Bone Ur]

Well bust mah britches and call me cheeky, on Fri, 16 Nov 2007 22:07:09
GMT Mika scribed:

PS: Due to Opera functionality limitations (partly on the width of
streetscape pages) that browser is not supported by our site, although
what you found out is interesting.

I notice you're using an "shtml" extension, but...

Here's something to try. I emphasize it's just a test. Change your
extension to php and see what happens. (I hope your server supports it.)
The idea relates to caching; who knows, maybe it'll make a difference.

 

[Mika]

Well bust mah britches and call me cheeky, on Fri, 16 Nov 2007 22:07:09
GMT Mika scribed:


I notice you're using an "shtml" extension, but...

Here's something to try. I emphasize it's just a test. Change your
extension to php and see what happens. (I hope your server supports it.)
The idea relates to caching; who knows, maybe it'll make a difference.

It seems to work! Can anyone replicate the old bug at the new php page?
http://tinyurl.com/324enb

 

[Mika]

Well bust mah britches and call me cheeky, on Fri, 16 Nov 2007 22:07:09
GMT Mika scribed:


I notice you're using an "shtml" extension, but...

Here's something to try. I emphasize it's just a test. Change your
extension to php and see what happens. (I hope your server supports it.)
The idea relates to caching; who knows, maybe it'll make a difference.

That does work http://tinyurl.com/324enb so the theory is sound BUT the
problem is php does to work correctly with our Google Maps integration, and
also we use SSI such as <!--#include virtual="/_borders/top.htm"--> which is
why we need shtml.

 

[Jonathan N. Little]

Mika wrote:
idea relates to caching; who knows, maybe it'll make a difference.

That does work http://tinyurl.com/324enb so the theory is sound BUT the
problem is php does to work correctly with our Google Maps integration, and
also we use SSI such as <!--#include virtual="/_borders/top.htm"--> which is
why we need shtml.

Ah but you can easily convert that to php like this:

<?php include_once("/_borders/top.htm"); ?>


and forego the SSI

 

[André Gillibert]

Mika wrote:

That does work http://tinyurl.com/324enb so the theory is sound BUT the
problem is php does to work correctly with our Google Maps integration,
and
also we use SSI such as <!--#include virtual="/_borders/top.htm"-->
which is
why we need shtml.

If you use Apache, you can use rewrite_mod.
A URI doesn't need to hold the name of the resource on the server side.

Anyway, changing the URI terminal characters shouldn't change the behavior
of user agents.
The fact is: Opera and probably some others RFC-2616 conforming user
agents doesn't cache, by default, resources the URI of which contains a
request string following a question mark.
Please, don't rely on this ridiculous hack. Use proper HTTP headers to
prevent caching.

 

[Mika]

Mika wrote:



If you use Apache, you can use rewrite_mod.
A URI doesn't need to hold the name of the resource on the server side.

Anyway, changing the URI terminal characters shouldn't change the behavior
of user agents.
The fact is: Opera and probably some others RFC-2616 conforming user
agents doesn't cache, by default, resources the URI of which contains a
request string following a question mark.
Please, don't rely on this ridiculous hack. Use proper HTTP headers to
prevent caching.

I agree. What headers should we try?

Is it as simple as changing the doctype?

Thanks for all the help so far.

 

[Mika]

Mika wrote:
idea relates to caching; who knows, maybe it'll make a difference.

Ah but you can easily convert that to php like this:

<?php include_once("/_borders/top.htm"); ?>


and forego the SSI

That would be great, but our map code is still thrown completely off by
using php. If you try to walk along the street, the map walks the wrong
way, the man does not appear, etc. It looks like we need a solution that
works with shtml. Perhaps changing the doctype?

 

[Jonathan N. Little]

That would be great, but our map code is still thrown completely off by
using php. If you try to walk along the street, the map walks the wrong
way, the man does not appear, etc. It looks like we need a solution that
works with shtml. Perhaps changing the doctype?

Oh that site...I remember a what was it a couple of year ago...

Well if there was ever a concept not conducive to html, this is it! What
you really want is Flash, (Yes Travis, if he really want to do this
thing Flash would be far better) sand progressively load sections...

 

[Jonathan N. Little]

Oh that site...I remember a what was it a couple of year ago...

Not sure how my "s" wandered all the way from.............^

Well if there was ever a concept not conducive to html, this is it! What
you really want is Flash, (Yes Travis, if he really want to do this
thing Flash would be far better) sand progressively load sections...

all the way down here!.............^

Damn gypsy "s"!

 

[Mika]

Oh that site...I remember a what was it a couple of year ago...

Well if there was ever a concept not conducive to html, this is it! What
you really want is Flash, (Yes Travis, if he really want to do this thing
Flash would be far better) sand progressively load sections...

Whilst we appreciate the feedback, the site works perfectly well - aside
from this new bug which has crept in. That is really all we are looking for
your advice with, and understood someone here might have a neat simple
solution. In some of the other discussions we are almost there.

I hope you appreciate that with some much work done and the site working, we
are not about to completely rewrite it in Flash because of one small glitch.
Simpler to fix the glitch.

Thanks in advance to anyone able to help with that.

 

[VK]

Accidentally I have read this thread and I noticed that the KB article
I linked in "Active time for HTTP connection" thread may be giving
another possible hint:

http://support.microsoft.com/kb/813827

"If you set the KeepAliveTimeout value to less than 60,000 (one
minute), you may have problems communicating with Web servers that
require persistent HTTP connections. For example, you may receive a
"Page cannot be displayed" error message."

It may be possible that some visitors have KeepAliveTimeout manually
set to too short, or that for some requests the inactivity period goes
beyond the default allowed 60,000ms

I cannot reproduce your problem neither in IE6 nor IE7 so just
guessing. If you have a machine stable reproducing the problem, then I
would check one by one:
1) HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion
\InternetSettings
Try to set KeepAliveTimeout DWORD value to 180000 and add new key
ServerInfoTimeout with DWORD value to 180000

2) Tools :: Internet Options :: Advanced :: HTTP 1.1 settings :: check
on "Use HTTP 1.1 through proxy connections"
(default is off)

 

[Jonathan N. Little]

Accidentally I have read this thread and I noticed that the KB article
I linked in "Active time for HTTP connection" thread may be giving
another possible hint:

http://support.microsoft.com/kb/813827

"If you set the KeepAliveTimeout value to less than 60,000 (one
minute), you may have problems communicating with Web servers that
require persistent HTTP connections. For example, you may receive a
"Page cannot be displayed" error message."

It may be possible that some visitors have KeepAliveTimeout manually
set to too short, or that for some requests the inactivity period goes
beyond the default allowed 60,000ms

I cannot reproduce your problem neither in IE6 nor IE7 so just
guessing. If you have a machine stable reproducing the problem, then I
would check one by one:
1) HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion
\InternetSettings
Try to set KeepAliveTimeout DWORD value to 180000 and add new key
ServerInfoTimeout with DWORD value to 180000

2) Tools :: Internet Options :: Advanced :: HTTP 1.1 settings :: check
on "Use HTTP 1.1 through proxy connections"
(default is off)

If this is the case because the OP's "web app" requires such delays in
order process and transmit the site do you really expect visitors to
hack their default Windows' registry just to see the site? And we have
long-winded arguments whether or not the average user understands the
browser back button!

 

[VK]

If this is the case because the OP's "web app" requires such delays in
order process and transmit the site do you really expect visitors to
hack their default Windows' registry just to see the site?

Of course not! We are on the stage of identifying the problem, not
solving it. If my guess is correct - but it may be very well wrong -
then OP has to think of a server-side adjustment to ensure that the
inactivity period will never exceed the default KeepAliveTimeout.
There is a number of such workarounds.

P.S. Also if this is a problem then the question also gets OT for this
NG and should be transferred to some server scripting group, depending
on used server-side software.

 

[Mika]

Accidentally I have read this thread and I noticed that the KB article
I linked in "Active time for HTTP connection" thread may be giving
another possible hint:

http://support.microsoft.com/kb/813827

"If you set the KeepAliveTimeout value to less than 60,000 (one
minute), you may have problems communicating with Web servers that
require persistent HTTP connections. For example, you may receive a
"Page cannot be displayed" error message."

Thank you however that is not the error message we receive. 'Operation
aborted' is the error, and only when refreshing the page whilst it is still
loading. I do not feel this error is likely to relate, but thank you
anyway...

It may be possible that some visitors have KeepAliveTimeout manually
set to too short, or that for some requests the inactivity period goes
beyond the default allowed 60,000ms

I cannot reproduce your problem neither in IE6 nor IE7 so just
guessing. If you have a machine stable reproducing the problem, then I
would check one by one:

Have you tried refreshing the page multiple times whilst it is still
loading? The error does not occur every time.