"Alan J Robertson" wrote in message

....
OK - follow-up. I've disabled the friendly HTTP error messages but they
won't disappear! (the option is definitely unchecked) - even tried
restarting the PC, but they still persist!

Some other things I've noticed -

the "s:" link in the error message - I realised that when I hovered the
mouse over this link it wasn't actually pointing to "s:" but instead to
"res://shdoclc/syntax.htm#" - I think the s came from the end of res.


Ah, in that case I agree--your idea about it being an internal problem
could be just as valid.

That URL is a self-reference, which you can see just by entering it
in the Address bar. I don't know if there is a way to request a different
link to be specified with it. Notice though that it perfectly validates
Frank's suggestion that scripting may be involved. Use View, Source
on that page and do a find for script.

Hmm... there are several comments in there to give some clues about how
the Homepage() function is supposed to work, including some instructions
for testing it. The comments were obviously copied from another source
file and there are some typos in it to make us work harder eg but here is
an example which I got to work first:

example
res://shdoclc.dll/syntax.htm#https://www.microsoft.com/bar.htm
/example

That is just an obvious example from following the comments
and adjusting them to this particular page. From there we could
experiment to see how closely to a valid URL we could come
but still result in your symptom. Clearly there are a lot of possible
cases which will not be handled by this function. Depending on
how its inputs are filtered by its callers it may be valid that it is
not coded more carefully to catch such cases.

The first thing it looks for is ://
Therefore any URL passed to it which is not prefixed by
a conventional protocol prefix will cause the symptom.

Then strangely it assumes that there must be a slash
also present. However (this is the really strange part)
if the slash is missing but the URL is otherwise fine
the end result is somehow the whole expression!

example
res://shdoclc.dll/syntax.htm#https://www.microsoft.com
/example


This doesn't even get into the question of why this routine
is being called in the first place but could explain why you
are getting that s: symptom.


PS You mentioned 'taking a trace' of the problem - what does this involve?
Do you mean just a tracert?

I meant do at minimum an HTTP packet capture
so you can see both the actual requests and responses
the same as IE is seeing them. To do that you could use
a proxy tracer such as FiddlerTool or you could capture
all TCP/IP packets using netcap (from the XP Pro Support
Tools package) and format them with Ethereal (freeware).


If I copy and past the URL it shows absolutely
fine in IE, it's just when I CLICK on a link the OE that it sometimes
occurs. Do you mean some way of getting windows scripting to follow what's
happening?



Let's examine your examples in more detail then.

Are the examples you gave from HTML E-mail?
How did you extract them? E.g. with right-click Copy Shortcut
or by capturing the text portion of the anchor?
(I'm thinking that HTML source may be the problem
in a similar way that phishing sites exploit it.)
Also a common problem with HTML links is for posters
to try pasting in a partial URL and then editing it,
not realizing that they are only editing an anchor's displayed
text, not the actual HREF itself. In order to do that
in OE they should use the Edit menu's Remove Hyperlink
command, edit the text and then when it is converted
to a link the HREF and the text will be the same.



Cheers & thanks for your ongoing advice!

Alan



HTH

Robert
---