Re: Pure Python implementation of hashcash v.1

"Eric S. Johansson" <[email protected]>
Newsgroups gmane.mail.spam.hashcash
Message-ID <[email protected]>
David Mertz, Ph.D. wrote:
> On Sep 24, 2004, at 2:36 PM, Eric S. Johansson wrote:
> 
>>> I also check and write email on my Powerbook, with OSX and using 
>>> Mail.app.  Anyone know if anyone has figured out how to interface 
>>> with Mail.app, it would certainly be nice to get hashcash integration 
>>> there.
> 
> 
>> well, you could take my hacks on smtpd.py and the postfix stamping 
>> filter interface code and build yourself a stamping proxy.  On the 
>> other hand, I have been intending to create a version of Camram that 
>> runs on client side machines and the mac is the most likely candidate.
> 
> 
> Yeah, a proxy is a possibility.  But it just seems so *fugly*!  Adding a 
> X-Hashcash: header doesn't have anything to do with the 
> SMTP/IMAP/sendmail/whatever interface; it's just a text processing task 
> on an email message.

well, you've just described a whole 'nother mission from God.  Fix all 
MUA's to have a rational and consistent interface for processing 
messages at different points in the handling.

it's more than just a simple "plug in the header" filter.  You really 
need to process things in background which means you have a queue and 
retry logic and a delivery agent etc. etc..  Without these things, you 
end up torturing the end-user, forcing them to wait for stamp generation 
to complete.

> Ideally, a mailer should just be able to post-process a message before 
> sending it (using all the network interface code already built into the 
> mailer.  Many *nix mailer let you specify an arbitrary filter for 
> outgoing messages... but MacOSX isn't targeted this way, unfortunately 
> (I understand novices don't want to be confused by that, but a buried 
> preference wouldn't hurt them).  Likewise, MUAs should be able to filter 
> using an arbitrary external utility.

like I said.  Mission from God project.  The really annoying thing is 
that MUAs really aren't suited for this kind of filtering and that it is 
only because of the filters we're talking about here is there any need. 
  Therefore, maybe a side project should be creating this kind of filter 
interface in Thunderbird.  If anyone has any contacts in Eudora land, 
see if one could get them to put the same interface there.

> FWIW, I will write a utility 'add_hashcash_stamps.py' that will simply 
> take an rfc2822 message on STDIN, and write one with added headers on 
> STDOUT.  I reckon this utility will optionally use an external compiled 
> minter (e.g. hashcash.exe) for speed, but use my hashcash.py if one is 
> not available.  Not sure the best way to configure the name/switches  of 
> the external tool, but I'll decide on something.

simple modification to my hashcash module.

> I guess while I'm at it, I might also do 'check_hashcash_stamp.py' that 
> takes an rfc2822.  This is a little tricker in a way though.  How do you 
> know whether a resource in an X-Hashcash: header is one you want to 
> accept?  For example, I use many different email addresses within my 
> 'gnosis.cx' domain (for various contacts).  In principle, I might accept 
> a stamp with any of these resources, since I'll accept email to any of 
> the addresses.  But there are actually infinitely many resources I might 
> accept inside a stamp (subject to it really having the right bits, date, 
> etc).  Then again, I don't want to accept just any resource... for 
> example, what if a spammer CC:'s herself, and produces a valid stamp 
> only for her own email address?

solved that problem in Camram as well.  I force the user to enumerate 
all valid e-mail addresses and I iterate over the list of valid e-mail 
addresses whenever I find a stamp in a message.  Again, check the 
hashcash module.

it would just take a simple wrapper to make them command line friendly.

---eric
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.