hashcash: duplicate payments

hashcash: duplicate payments

[Andreas Seltenreich]

[-- Attachment #1: Type: text/plain, Size: 636 bytes --]

Hi,

I just noticed that mail-add-payment (called from message-send-message)
doesn't recognise asynchronously generated payments for addresses that
contain uppercase characters. Attached is a patch that makes the test
case-insensitive.

An alternative solution would be to supply the "-C" option to the
external hashcash program, but IMHO this would unnecessarily restrict
the user in his customisation needs (hashcash-extra-generate-parameters).

Thanks,
andreas

2006-03-26  Andreas Seltenreich  <uwi7@rz.uni-karlsruhe.de>

	* hashcash.el (hashcash-already-paid-p): Bind case-fold-search
	when searching for already-paid recipients.


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: hashcash.el.patch --]
[-- Type: text/x-patch, Size: 824 bytes --]

Index: hashcash.el
===================================================================
RCS file: /usr/local/cvsroot/gnus/lisp/hashcash.el,v
retrieving revision 7.14
diff -c -r7.14 hashcash.el
*** hashcash.el	6 Jul 2005 06:42:01 -0000	7.14
--- hashcash.el	26 Mar 2006 16:49:17 -0000
***************
*** 210,216 ****
    (save-excursion
      (save-restriction
        (message-narrow-to-headers-or-head)
!       (let ((token (message-fetch-field "x-hashcash")))
  	(and (stringp token)
  	     (string-match (regexp-quote recipient) token))))))
  
--- 210,217 ----
    (save-excursion
      (save-restriction
        (message-narrow-to-headers-or-head)
!       (let ((token (message-fetch-field "x-hashcash"))
! 	    (case-fold-search t))
  	(and (stringp token)
  	     (string-match (regexp-quote recipient) token))))))
  

Re: hashcash: duplicate payments

[Andreas Seltenreich]

[-- Attachment #1: Type: text/plain, Size: 478 bytes --]

Andreas Seltenreich schrob:

> X-Hashcash: 1:24:060326:ding@gnus.org::S1H1LHoTY/UosfJN:0000ymPT
> X-Hashcash: 1:24:060326:ding@gnus.org::bZhXJcbr0RqrGqAV:00008EL9

Apparently, the hashcash code is also run on temporary buffers during
resending. Is binding message-generate-hashcash at those places an
acceptable solution?

regards,
andreas

2006-03-26  Andreas Seltenreich  <uwi7@rz.uni-karlsruhe.de>

	* message.el (message-resend): Bind message-generate-hashcash to
	to nil.


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: message.el.patch --]
[-- Type: text/x-patch, Size: 878 bytes --]

Index: message.el
===================================================================
RCS file: /usr/local/cvsroot/gnus/lisp/message.el,v
retrieving revision 7.135
diff -c -r7.135 message.el
*** message.el	15 Mar 2006 18:03:48 -0000	7.135
--- message.el	26 Mar 2006 19:57:40 -0000
***************
*** 6507,6512 ****
--- 6507,6513 ----
  	(set-buffer (get-buffer-create " *message resend*"))
  	(erase-buffer))
        (let ((message-this-is-mail t)
+ 	    message-generate-hashcash
  	    message-setup-hook)
  	(message-setup `((To . ,address))))
        ;; Insert our usual headers.
***************
*** 6544,6549 ****
--- 6545,6551 ----
        ;; Send it.
        (let ((message-inhibit-body-encoding t)
  	    message-required-mail-headers
+ 	    message-generate-hashcash
  	    rfc2047-encode-encoded-words)
  	(message-send-mail))
        (kill-buffer (current-buffer)))

Re: hashcash: duplicate payments

[Simon Josefsson]

Andreas Seltenreich <andreas+ding@gate450.dyndns.org> writes:

> Hi,
>
> I just noticed that mail-add-payment (called from message-send-message)
> doesn't recognise asynchronously generated payments for addresses that
> contain uppercase characters. Attached is a patch that makes the test
> case-insensitive.
>
> An alternative solution would be to supply the "-C" option to the
> external hashcash program, but IMHO this would unnecessarily restrict
> the user in his customisation needs (hashcash-extra-generate-parameters).
>
> Thanks,
> andreas
>
> 2006-03-26  Andreas Seltenreich  <uwi7@rz.uni-karlsruhe.de>
>
> 	* hashcash.el (hashcash-already-paid-p): Bind case-fold-search
> 	when searching for already-paid recipients.

Applied, thanks!

> Index: hashcash.el
> ===================================================================
> RCS file: /usr/local/cvsroot/gnus/lisp/hashcash.el,v
> retrieving revision 7.14
> diff -c -r7.14 hashcash.el
> *** hashcash.el	6 Jul 2005 06:42:01 -0000	7.14
> --- hashcash.el	26 Mar 2006 16:49:17 -0000
> ***************
> *** 210,216 ****
>     (save-excursion
>       (save-restriction
>         (message-narrow-to-headers-or-head)
> !       (let ((token (message-fetch-field "x-hashcash")))
>   	(and (stringp token)
>   	     (string-match (regexp-quote recipient) token))))))
>   
> --- 210,217 ----
>     (save-excursion
>       (save-restriction
>         (message-narrow-to-headers-or-head)
> !       (let ((token (message-fetch-field "x-hashcash"))
> ! 	    (case-fold-search t))
>   	(and (stringp token)
>   	     (string-match (regexp-quote recipient) token))))))
>   

Re: hashcash: duplicate payments

[Simon Josefsson]

Andreas Seltenreich <andreas+ding@gate450.dyndns.org> writes:

> Andreas Seltenreich schrob:
>
>> X-Hashcash: 1:24:060326:ding@gnus.org::S1H1LHoTY/UosfJN:0000ymPT
>> X-Hashcash: 1:24:060326:ding@gnus.org::bZhXJcbr0RqrGqAV:00008EL9
>
> Apparently, the hashcash code is also run on temporary buffers during
> resending. Is binding message-generate-hashcash at those places an
> acceptable solution?

It seems slightly out of place, but I can't think of a better
solution, so I installed it.  Thanks!

> regards,
> andreas
>
> 2006-03-26  Andreas Seltenreich  <uwi7@rz.uni-karlsruhe.de>
>
> 	* message.el (message-resend): Bind message-generate-hashcash to
> 	to nil.
>
> Index: message.el
> ===================================================================
> RCS file: /usr/local/cvsroot/gnus/lisp/message.el,v
> retrieving revision 7.135
> diff -c -r7.135 message.el
> *** message.el	15 Mar 2006 18:03:48 -0000	7.135
> --- message.el	26 Mar 2006 19:57:40 -0000
> ***************
> *** 6507,6512 ****
> --- 6507,6513 ----
>   	(set-buffer (get-buffer-create " *message resend*"))
>   	(erase-buffer))
>         (let ((message-this-is-mail t)
> + 	    message-generate-hashcash
>   	    message-setup-hook)
>   	(message-setup `((To . ,address))))
>         ;; Insert our usual headers.
> ***************
> *** 6544,6549 ****
> --- 6545,6551 ----
>         ;; Send it.
>         (let ((message-inhibit-body-encoding t)
>   	    message-required-mail-headers
> + 	    message-generate-hashcash
>   	    rfc2047-encode-encoded-words)
>   	(message-send-mail))
>         (kill-buffer (current-buffer)))

Re: hashcash: duplicate payments

[Andreas Seltenreich]

Simon Josefsson schrob:

> Andreas Seltenreich <andreas+ding@gate450.dyndns.org> writes:
>
>> Apparently, the hashcash code is also run on temporary buffers during
>> resending. Is binding message-generate-hashcash at those places an
>> acceptable solution?
>
> It seems slightly out of place, but I can't think of a better
> solution, so I installed it.

Yeah, I had an uneasy feeling when I first saw that the hashcash stuff
was hardcoded into message.el, too.  And it went even uneasier now
that I realised that the draft code also uses scratch message buffers.

It does bind the message hooks to nil when it does this, as does the
resend code.  So I'm wondering if it wouldn't be more appropriate to
resort to using hooks for hashcash.el, which would IMHO result in a
more deterministic behaviour.

regards,
andreas

Re: hashcash: duplicate payments

[Simon Josefsson]

Andreas Seltenreich <andreas+ding@gate450.dyndns.org> writes:

> Simon Josefsson schrob:
>
>> Andreas Seltenreich <andreas+ding@gate450.dyndns.org> writes:
>>
>>> Apparently, the hashcash code is also run on temporary buffers during
>>> resending. Is binding message-generate-hashcash at those places an
>>> acceptable solution?
>>
>> It seems slightly out of place, but I can't think of a better
>> solution, so I installed it.
>
> Yeah, I had an uneasy feeling when I first saw that the hashcash stuff
> was hardcoded into message.el, too.  And it went even uneasier now
> that I realised that the draft code also uses scratch message buffers.
>
> It does bind the message hooks to nil when it does this, as does the
> resend code.  So I'm wondering if it wouldn't be more appropriate to
> resort to using hooks for hashcash.el, which would IMHO result in a
> more deterministic behaviour.

That seems better, yes.  Would you like to work on it?