recent imap regression

recent imap regression

[Greg Troxel]

[-- Attachment #1: Type: text/plain, Size: 2227 bytes --]


I just updated to latest git, and it looks like these two changes:

  b12f45e int, nnimap: remove to-buffer in request-head
  04ef222 gnus-int: add to-buffer parameter to gnus-request-head

which were intended to be a change and a revert, ended up doing:

diff --git a/lisp/ChangeLog b/lisp/ChangeLog
index 1bf8444..9e1cc6e 100644
--- a/lisp/ChangeLog
+++ b/lisp/ChangeLog
@@ -1,3 +1,14 @@
+2012-08-06  Julien Danjou  <julien@danjou.info>
+
+	* nnimap.el (nnimap-request-article): Remove to-buffer argument.
+
+	* gnus-int.el (gnus-request-head): Remove to-buffer argument, only
+	supported by nnimap actually. Reverts previous change.
+
+	* gnus-int.el (gnus-request-head): Add an optional to-buffer parameter
+	to mimic `gnus-request-article' and enjoy backends the nn*-request-head
+	to-buffer argument that is already supported.
+
 2012-08-05  Julien Danjou  <julien@danjou.info>
 
 	* gnus-demon.el (gnus-demon-add-handler, gnus-demon-remove-handler):
diff --git a/lisp/nnimap.el b/lisp/nnimap.el
index ebd268e..e195b42 100644
--- a/lisp/nnimap.el
+++ b/lisp/nnimap.el
@@ -597,7 +597,7 @@ textual parts.")
 		(nnheader-ms-strip-cr)))
 	    (cons group article)))))))
 
-(deffoo nnimap-request-head (article &optional group server to-buffer)
+(deffoo nnimap-request-head (article &optional group server)
   (when group
     (setq group (nnimap-decode-gnus-group group)))
   (when (nnimap-change-group group server)
@@ -610,7 +610,7 @@ textual parts.")
 	 article (format "UID FETCH %%d %s"
 			 (nnimap-header-parameters)))
 	(let ((buffer (current-buffer)))
-	  (with-current-buffer (or to-buffer nntp-server-buffer)
+	  (with-current-buffer nntp-server-buffer
 	    (erase-buffer)
 	    (insert-buffer-substring buffer)
 	    (nnheader-ms-strip-cr)


which results in the following when using delete.


nnimap-request-move-article: Wrong number of arguments: #[(article &optional group server) "\bƒ\bÅ\b!Æ\b	\"…ArÇ qˆ
;ƒ\x1fÈ\b	
#
…@É
ÊËÌ \"\"ˆpr\fqˆÍ ˆÎ\v!ˆÏ ˆ\b
*B)‡" [group server article buffer nntp-server-buffer nnimap-decode-gnus-group nnimap-change-group nnimap-buffer nnimap-find-article-by-message-id nnimap-get-whole-article ...] 5], 4


[-- Attachment #2: Type: application/pgp-signature, Size: 194 bytes --]

Re: recent imap regression

[Julien Danjou]

[-- Attachment #1: Type: text/plain, Size: 426 bytes --]

On Mon, Aug 06 2012, Greg Troxel wrote:

> I just updated to latest git, and it looks like these two changes:
>
>   b12f45e int, nnimap: remove to-buffer in request-head
>   04ef222 gnus-int: add to-buffer parameter to gnus-request-head
>
> which were intended to be a change and a revert, ended up doing:

Yep, already fixed.

-- 
Julien Danjou
-- Free Software hacker & freelance
-- http://julien.danjou.info

[-- Attachment #2: Type: application/pgp-signature, Size: 835 bytes --]

Re: recent imap regression

[Greg Troxel]

[-- Attachment #1: Type: text/plain, Size: 523 bytes --]


Julien Danjou <julien@danjou.info> writes:

> On Mon, Aug 06 2012, Greg Troxel wrote:
>
>> I just updated to latest git, and it looks like these two changes:
>>
>>   b12f45e int, nnimap: remove to-buffer in request-head
>>   04ef222 gnus-int: add to-buffer parameter to gnus-request-head
>>
>> which were intended to be a change and a revert, ended up doing:
>
> Yep, already fixed.

Thanks.  It might be good to use 'git revert' to undo things, since it's
guaranteed (absent merge conflicts) to apply inverse changesets.

[-- Attachment #2: Type: application/pgp-signature, Size: 194 bytes --]

Re: recent imap regression

[Julien Danjou]

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

On Mon, Aug 06 2012, Greg Troxel wrote:

> Thanks.  It might be good to use 'git revert' to undo things, since it's
> guaranteed (absent merge conflicts) to apply inverse changesets.

I know, but what you pointed was not a direct revert, see the changelog.
I would use git revert directly, but it seems that it's prefered to
document revert in changelog (correct me if I'm wrong).

-- 
Julien Danjou
;; Free Software hacker & freelance
;; http://julien.danjou.info

[-- Attachment #2: Type: application/pgp-signature, Size: 835 bytes --]

Re: recent imap regression

[Greg Troxel]

[-- Attachment #1: Type: text/plain, Size: 551 bytes --]


Julien Danjou <julien@danjou.info> writes:

> I know, but what you pointed was not a direct revert, see the changelog.
> I would use git revert directly, but it seems that it's prefered to
> document revert in changelog (correct me if I'm wrong).

Perhaps, but my take is that when the revert is close in time, then it's
just as well to revert the changelog entry, since then the new state is
exactly as if it never happened, which won't confuse  anyway.  I suppose
there's git revert followed by git commit --amend to destage the
changelog hunk :-)

[-- Attachment #2: Type: application/pgp-signature, Size: 194 bytes --]