Re: [htdig-dev] Re: ssl and 3.2.0b4-20021110

[Gilles D. <gr...@sc...>]

According to Budd, Sinclair:
> Hello
>   After much scrambling, I found the script  "handler.pl"  by Geoffrey
> Hutchison <ghu...@ws...> which is used with
> "external_protocols" to fetch https  pages. 
>  
>    IT WORKS A WONDER.
> 
> There is a small documentation issue .  In  attrs.html  under
> "external_protocols" the statement  "The first field is a single character
> that specifies the record type." appears.  In the handler.pl  the first
> field is shown as "character:"   i.e. there are two characters.
>      Also from the documentation it is not clear whether the  u  record is
> mandatory or it is only needed for a redirect.  In the handler.pl  script
> the u record is only reported if the Location header is returned from the
> web server.

The 'u' record is indeed only used for redirects, and is by no means
mandatory, as far as I can tell by looking at the code.  I don't think
it would hurt anything if an external transport did issue 'u' records
without also issuing an 's' (status) record that triggered a redirect,
but there's no point in doing so, let alone a need to do so.

As for the colon, it seems to be totally optional as far as the transport
record processing is concerned.  Indeed, the tab character is absolutely
essential, but the parsing code seems to ignore any characters between
the first letter of the record and the tab.

I don't know what the problem is with the built-in HTTPS handling code
in 3.2.0b4, but with all the problem reports we've had on it it's safe
to say this code is broken, and you should use an external transport
for HTTPS until the problem is resolved.  Anyone out there know HTTPS
enough to take a crack at this?

-- 
Gilles R. Detillieux              E-mail: <gr...@sc...>
Spinal Cord Research Centre       WWW:    http://www.scrc.umanitoba.ca/
Dept. Physiology, U. of Manitoba  Winnipeg, MB  R3E 3J7  (Canada)