[SC-Help] Re: false multipart header, Spamcop misses URL in body

[Mike Easter]

Joe Blow wrote:
> "Mike Easter"
>> The parser uses the content type line in the headers and the content
>> type line in the body when it is separated by boundary delimitors to
>> tell it the 'content type'

> Text is text....and that is all that is being parsed.  The parser
> doesn't have to deal with the content type because it doesn't
> matter...the message is just a blob of text.

No, you are mistaken.  You don't understand how the parser uses content
type in its body parsing.

> All the parser needs to
> do is look for hyperlinks within that text.   It seems to me that
> you're implying the task is more difficult than that....why?

Because it is.  [You hate answers like that, but you present your
arguments in ways that demand those answers.]  You are going to have to
look at more good and bad body parses of various spamtypes, correctly and
incorrectly constructed, to get the full picture.  I cannot write a
treatise on that here.

The body parse does *not* simply look for the occurrence of
http://domain.tld - but instead must find it properly 'textually' [or
'contextually' textually] constructed correctly for its appropriate
content type.

>> Content type is a legitimate MIME descriptor, and the accurate
>> determination of the notify requires that the /actual/ spamvertised
>> sites get notified accurately - which requires being 'right' about
>> the content type.

Read that above again.

-- 
Mike Easter
kibitzer, not SC admin