Merge issues in `tls.el' (EMACS_22_BASE vs. trunk)

Merge issues in `tls.el' (EMACS_22_BASE vs. trunk)

[Reiner Steib]

Hi,

it seems that something went wrong when syncing Emacs_22 and the trunk
in `tls.el':

- EMACS_22_BASE should not use `rx'.  (The log of revision 1.27 says
  "Revert merge that re-required rx.").

- In the trunk, `tls-end-of-info' is defcustomed, but not used at all.
  It is used in `open-tls-stream' in EMACS_22_BASE.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: Merge issues in `tls.el' (EMACS_22_BASE vs. trunk)

[Glenn Morris]

Reiner Steib wrote:

> it seems that something went wrong when syncing Emacs_22 and the trunk
> in `tls.el':

No, something went wrong syncing gnus trunk to emacs trunk. :)
Anyway, well spotted.

> - EMACS_22_BASE should not use `rx'.  (The log of revision 1.27 says
>   "Revert merge that re-required rx.").

There's no problem there.

> - In the trunk, `tls-end-of-info' is defcustomed, but not used at all.
>   It is used in `open-tls-stream' in EMACS_22_BASE.

The 2007-12-05 changes of Elias Oltmanns to open-tls-stream seem to
have clobbered the 2007-11-04 changes of Riccardo Murri. Does this
patch look ok?

*** tls.el.~1.30.~	2008-02-28 20:00:45.000000000 -0800
--- tls.el	2008-03-01 15:35:48.000000000 -0800
***************
*** 244,259 ****
  	      (sit-for 1)))
  	  (message "Opening TLS connection with `%s'...%s" cmd
  		   (if done "done" "failed"))
! 	  (if done
! 	      (setq done process)
! 	    (delete-process process))))
!       (when done
  	(save-excursion
! 	  (set-buffer buffer)
! 	  (when
  	      (or
  	       (and tls-checktrust
! 		    (progn
  		      (goto-char (point-min))
  		      (re-search-forward tls-untrusted nil t))
  		    (or
--- 244,272 ----
  	      (sit-for 1)))
  	  (message "Opening TLS connection with `%s'...%s" cmd
  		   (if done "done" "failed"))
! 	  (if (not done)
! 	      (delete-process process)
! 	    ;; advance point to after all informational messages that
! 	    ;; `openssl s_client' and `gnutls' print
! 	    (let ((start-of-data nil))
! 	      (while
! 		   (not (setq start-of-data
! 			      ;; the string matching `tls-end-of-info'
! 			      ;; might come in separate chunks from
! 			      ;; `accept-process-output', so start the
! 			      ;; search where `tls-success' ended
  			      (save-excursion
! 				(if (re-search-forward tls-end-of-info nil t)
! 				    (match-end 0)))))
! 		(accept-process-output process 1))
! 	      (if start-of-data
! 		  ;; move point to start of client data
! 		  (goto-char start-of-data)))
! 	  (setq done process))))
!       (when (and done
  		 (or
  		  (and tls-checktrust
! 		       (save-excursion
  			 (goto-char (point-min))
  			 (re-search-forward tls-untrusted nil t))
  		       (or
***************
*** 262,276 ****
  		     (not (yes-or-no-p
  			   (format "The certificate presented by `%s' is NOT trusted. Accept anyway? " host)))))
  	       (and tls-hostmismatch
! 		    (progn
  		      (goto-char (point-min))
  		      (re-search-forward tls-hostmismatch nil t))
  		    (not (yes-or-no-p
! 			  (format "Host name in certificate doesn't match `%s'. Connect anyway? " host)))))
  	    (setq done nil)
! 	    (delete-process process))))
        (message "Opening TLS connection to `%s'...%s"
! 	       host (if done "done" "failed")))
      (when use-temp-buffer
        (if done (set-process-buffer process nil))
        (kill-buffer buffer))
--- 275,289 ----
  			(not (yes-or-no-p
  			      (format "The certificate presented by `%s' is NOT trusted. Accept anyway? " host)))))
  		  (and tls-hostmismatch
! 		       (save-excursion
  			 (goto-char (point-min))
  			 (re-search-forward tls-hostmismatch nil t))
  		       (not (yes-or-no-p
! 			     (format "Host name in certificate doesn't match `%s'. Connect anyway? " host))))))
  	(setq done nil)
! 	(delete-process process)))
      (message "Opening TLS connection to `%s'...%s"
! 	     host (if done "done" "failed"))
      (when use-temp-buffer
        (if done (set-process-buffer process nil))
        (kill-buffer buffer))

Re: Merge issues in `tls.el' (EMACS_22_BASE vs. trunk)

[Reiner Steib]

On Sun, Mar 02 2008, Glenn Morris wrote:

> Reiner Steib wrote:
>> it seems that something went wrong when syncing Emacs_22 and the trunk
>> in `tls.el':
>
> No, something went wrong syncing gnus trunk to emacs trunk. :)
> Anyway, well spotted.

(In fact, I found it when comparing v5_10 and EMACS_22_BASE.)

>> - EMACS_22_BASE should not use `rx'.  (The log of revision 1.27 says
>>   "Revert merge that re-required rx.").
>
> There's no problem there.

The "problem" is loading rx and (maybe more important) unnecessary
differences between branch and trunk.

I'd suggest to replace the former one.  But Yidong or Stefan should
decide if it should be before or after the release of Emacs 22.2.

>> - In the trunk, `tls-end-of-info' is defcustomed, but not used at all.
>>   It is used in `open-tls-stream' in EMACS_22_BASE.
>
> The 2007-12-05 changes of Elias Oltmanns to open-tls-stream seem to
> have clobbered the 2007-11-04 changes of Riccardo Murri. Does this
> patch look ok?

Looks ok for me (cursory inspection only).

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: Merge issues in `tls.el' (EMACS_22_BASE vs. trunk)

[Glenn Morris]

Reiner Steib wrote:

> The "problem" is loading rx and (maybe more important) unnecessary
> differences between branch and trunk.

As I remember it, Riccardo Murri submitted a bug fix which I
installed. Due to my personal preference, I re-wrote it to not use rx.
I only did this in the trunk because this part was not a bug fix.
I don't use tls and haven't tested my re-write in practice.

Re: Merge issues in `tls.el' (EMACS_22_BASE vs. trunk)

[Katsumi Yamaoka]

> Reiner Steib wrote:

>> The "problem" is loading rx and (maybe more important) unnecessary
>> differences between branch and trunk.

In the Gnus v5-10 branch, which is being synch'd with the Emacs
EMACS_22_BASE branch, the version of tls.el that uses rx has been
merged yesterday.  And tls.el came not to be able to byte compile
with XEmacs, because rx.el is not available in XEmacs.  Could it
be rewritten not using rx.el?

(info "(gnus)Emacsen")
,----
| Gnus should work on:
|
|    * Emacs 21.1 and up.
|
|    * XEmacs 21.4 and up.
`----

Regards,

Re: Merge issues in `tls.el' (EMACS_22_BASE vs. trunk)

[Reiner Steib]

On Mon, Mar 10 2008, Katsumi Yamaoka wrote:

>> Reiner Steib wrote:
>>> The "problem" is loading rx and (maybe more important) unnecessary
>>> differences between branch and trunk.
>
> In the Gnus v5-10 branch, which is being synch'd with the Emacs
> EMACS_22_BASE branch, the version of tls.el that uses rx has been
> merged yesterday.  And tls.el came not to be able to byte compile
> with XEmacs, because rx.el is not available in XEmacs.  Could it
> be rewritten not using rx.el?

Yes, the version in Emacs trunk and Gnus trunk, the defcustom of
`tls-end-of-info' has been rewritten not to use rx.  Please use this
in v5-10.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: Merge issues in `tls.el' (EMACS_22_BASE vs. trunk)

[Katsumi Yamaoka]

>>>>> Reiner Steib wrote:
> On Mon, Mar 10 2008, Katsumi Yamaoka wrote:

>> In the Gnus v5-10 branch, which is being synch'd with the Emacs
>> EMACS_22_BASE branch, the version of tls.el that uses rx has been
>> merged yesterday.  And tls.el came not to be able to byte compile
>> with XEmacs, because rx.el is not available in XEmacs.  Could it
>> be rewritten not using rx.el?

> Yes, the version in Emacs trunk and Gnus trunk, the defcustom of
> `tls-end-of-info' has been rewritten not to use rx.  Please use this
> in v5-10.

Thanks.  I've imported the following change made in the Emacs trunk
(i.e., rdiff -r1.20 -r1.21) to the Gnus v5-10 branch.

2007-11-04  Glenn Morris  <rgm@gnu.org>

	* net/tls.el: Don't require rx when compiling.
	(tls-end-of-info): Rewrite without using rx.
	(open-tls-stream): Use with-current-buffer.

Regards,