Re: RE: MORE Checker Samples

[Jock Dempsey <guru@anvilfire.com>]

Gunnar asked,
>But if the checker actually looked for certain links, and let some sites
>without any Ringlink HTML code pass, it's really strange. I can't see how
this
>could happen.
Its not.  I thought it was until I realized we were testing for NULL.  Any
page that is successfuly found in my rings will pass since there is no link
being compared.  HOWEVER, if I use option #2 and look for specific links
every page will FAIL that has the Javascript code.
>What's the contents of your htmlcode.txt file?
<!-- Begin The Blacksmiths Ring code fragment -->
<SCRIPT LANGUAGE= "JavaScript">var BlacksmithsRingID='$::siteid'; </SCRIPT>
<SCRIPT LANGUAGE= "JavaScript"
SRC="http://rings.anvilfire.com/data/smithring/smithring.js"></SCRIPT>
<!-- End  Blacksmiths Ring -->
There is no HTML (other than the SCRIPT statements). 
The two lines of Javascript call a Javascript source file with
document.write lines and produce the HTML.  The first line contains the
siteid variable, the second calls the SRC (source code).  The following is
one of 19 lines in the  'smithring.js' file that create the HTML ring
navigation code.  I've added hard returns to make it more readable.
document.write 
('<A HREF="http://rings.anvilfire.com/cgi-bin/prev.pl?
ringid=smithring;siteid=' + BlacksmithsRingID + '" TARGET="_top">
<FONT COLOR="#FE0000"><< Previous</FONT></A> | ');
As you can see this is the PREV link.  The variable BlacksmithsRingID
inserts the siteid string (or number).  Similar to your htmlcode.txt the
.js file is a template that is the same for all sites.  The example above
uses only one variable but I have another ring that uses three (siteID,
e-mail, name).
The htmlcode.txt file uses the $::siteid variable to automaticaly generate
the Javascript the same as it would the HTML.  I'll bet you never thought
anyone would do THAT with Ringlink!
YES, I know the FONT color in the anchor code is non-standard.  However, it
works in both IE and NS, and it prevents the font color set on the page OR
via CSS to change the link text color.
--------------------------------------------
A third checker option to check for a string provided by the Ringmaster
would let me check for 'BlacksmithsRingID' (or any other line in the
Javascript) and the checker would work for my rings and others that are
using my jsNAVBAR.
---------------------------------------------
>> 1. Is it possible that these are spurious errors that should be
>> waited out?
>> 
>> 2. Is it possible that because the checker didn't find the links it
>> returned an error?
>Sorry, I don't understand 1. and 2.
1) It seems that some of the "free" hosts that have database generated code
are interupting the expected HTML response.  While the system is processing
the request the server may be breifly responding with an error mesage.
Could it be that browsers use more "retries" (not just waiting) than the
checker and eventualy get through?
2) Sometimes programs that run into unexpected errors return the last error
message or whatever happens to be in memory.  I was just wondering if
because I have created a condition where the string being looked for was
NULL that this may have created the wrong errors to be returned. . .
Hey, I am HARD on software.
Jock D.