Integration of nnmairix.el

Integration of nnmairix.el

[Reiner Steib]

Hi David,

now that your assignment is on file, we can install `nnmairix.el' in
Gnus.  (Sorry for the late response.)

Could you please send the current version of `nnmairix.el' to the list
for review?  I didn't look into it in detail, but at least the
following items need to be addressed:

- Documentation.  Minimally the content from
  <http://www.emacswiki.org/cgi-bin/wiki/GnusMairix> should be
  converted to texinfo to fit into `texi/gnus.texi'.  If you need help
  with texinfo, please feel free to ask.

- Key bindings:

  · Some of the summary mode bindings (`S ...') are already taken
    (S t, S f, S m).  Or are these only active in nnmairix groups?

  · The Group mode prefix `G b' is available.  But maybe there's a
    better one (I don't know).  Maybe we should arrange that the group
    mode bindings are only done if the user wants to use nnmairix?

- Doc strings:

  Some of the doc strings don't follow the Emacs conventions.  See
  `M-x checkdoc RET'.

- The defcustoms need a :version tag:
  :version "23.0" ;; No Gnus

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

Re: Integration of nnmairix.el

[David]

[-- Attachment #1: Type: text/plain, Size: 1417 bytes --]

Reiner Steib <reinersteib+gmane@imap.cc> writes:
> Could you please send the current version of `nnmairix.el' to the list
> for review?  

It's attached to this mail.

> I didn't look into it in detail, but at least the
> following items need to be addressed:
>
> - Documentation.  Minimally the content from
>   <http://www.emacswiki.org/cgi-bin/wiki/GnusMairix> should be
>   converted to texinfo to fit into `texi/gnus.texi'.

OK, I'll do that.

> - Key bindings:
>
>   · Some of the summary mode bindings (`S ...') are already taken
>     (S t, S f, S m).  Or are these only active in nnmairix groups?

No. Since these are commands which create search commands based on the
current article they should work in every group. I somehow overlooked
that these bindings are already taken. 

>   · The Group mode prefix `G b' is available.  But maybe there's a
>     better one (I don't know).  Maybe we should arrange that the group
>     mode bindings are only done if the user wants to use nnmairix?

Yes, I guess that's reasonable since those bindings are useless without
a working mairix installation, so maybe users should have to explicitly
put (require 'nnmairix) into their .gnus for activating them.

> - Doc strings:
>
>   Some of the doc strings don't follow the Emacs conventions.  See
>   `M-x checkdoc RET'.
>
> - The defcustoms need a :version tag:
>   :version "23.0" ;; No Gnus

I fixed that.

-David


[-- Attachment #2: nnmairix.el --]
[-- Type: application/emacs-lisp, Size: 51105 bytes --]

Re: Integration of nnmairix.el

[Reiner Steib]

On Tue, Jan 08 2008, David wrote:

> Reiner Steib <reinersteib+gmane@imap.cc> writes:
>> Could you please send the current version of `nnmairix.el' to the list
>> for review?  
>
> It's attached to this mail.

Thanks.  I added the version 0.5 from the emacswiki now.  Sorry for
the delay.

>> I didn't look into it in detail, but at least the
>> following items need to be addressed:
>>
>> - Documentation.  Minimally the content from
>>   <http://www.emacswiki.org/cgi-bin/wiki/GnusMairix> should be
>>   converted to texinfo to fit into `texi/gnus.texi'.
>
> OK, I'll do that.

>> - Key bindings:
>>
>>   · Some of the summary mode bindings (`S ...') are already taken
>>     (S t, S f, S m).  Or are these only active in nnmairix groups?
>
> No. Since these are commands which create search commands based on the
> current article they should work in every group. I somehow overlooked
> that these bindings are already taken. 

I don't really like the use of `$ ...' in `nnmairix-summary-mode-hook'.

>>   · The Group mode prefix `G b' is available.  But maybe there's a
>>     better one (I don't know).  Maybe we should arrange that the group
>>     mode bindings are only done if the user wants to use nnmairix?
>
> Yes, I guess that's reasonable since those bindings are useless without
> a working mairix installation, 

We also want to install nnir.el soon.  nnir.el uses `G G' (only a
single binding).  (And there's contrib/gnus-namazu.el which uses `C-c
C-n' when `gnus-namazu-insinuate' is called).  Maybe we should think
about using some two key stroke prefix for mail search facilities (I'm
not sure if it makes sense to use nnir.el and nnmairix.el at the same
time).

For the Group mode, we could use e.g. `G b', `G B' or `G G' as prefix
for (mail) search.  In the Summary mode, we may use the same prefix or
a different one.  I don't have a strong opinion on this.  (`G b' runs
`gnus-summary-best-unread-article', but it is also on `,', so we may
steal this or move it to `G B'.)  Opinions?

> so maybe users should have to explicitly put (require 'nnmairix)
> into their .gnus for activating them.

Loading a package should not perform changes like this.  There should
be an autoloaded initialization function instead.  (If detecting a
working mairix installation is possible, we could install it from
`gnus.el'/`gnus-start.el' depending on the result.)

>> - Doc strings:
>>
>>   Some of the doc strings don't follow the Emacs conventions.  See
>>   `M-x checkdoc RET'.
>>
>> - The defcustoms need a :version tag:
>>   :version "23.0" ;; No Gnus
>
> I fixed that.

Thanks.

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

Re: Integration of nnmairix.el

[Katsumi Yamaoka]

[-- Attachment #1: Type: text/plain, Size: 1861 bytes --]

>>>>> Reiner Steib wrote:

> Thanks.  I added the version 0.5 from the emacswiki now.  Sorry for
> the delay.

Sorry for finding fault with it.  But many warnings are issued
when compiling.  Most of them look harmless, however the ones
concerning `nnmairix-request-group-with-article-number-correction'
seem to be real problems.  A patch is attached below, though I'm
not sure of it.  I would appreciate reviewing it.

Regards,

In toplevel form:
nnmairix.el:340:7:Warning: assignment to free variable `nnmairix-group-regexp'
nnmairix.el:344:7:Warning: assignment to free variable
    `nnmairix-valid-backends'
nnmairix.el:347:7:Warning: assignment to free variable `nnmairix-last-server'
nnmairix.el:350:7:Warning: assignment to free variable
    `nnmairix-current-server'

In nnmairix-open-server:
nnmairix.el:360:9:Warning: assignment to free variable
    `nnmairix-current-server'

In nnmairix-request-list:
nnmairix.el:521:37:Warning: reference to free variable `nnmairix-group-regexp'

In nnmairix-purge-old-groups:
nnmairix.el:793:41:Warning: reference to free variable `nnmairix-group-regexp'

In nnmairix-request-group-with-article-number-correction:
nnmairix.el:977:53:Warning: reference to free variable `fast'
nnmairix.el:1009:68:Warning: reference to free variable `group'

In nnmairix-get-server:
nnmairix.el:1056:52:Warning: reference to free variable `nnmairix-last-server'
nnmairix.el:1058:11:Warning: assignment to free variable
    `nnmairix-last-server'

In nnmairix-get-valid-servers:
nnmairix.el:1107:40:Warning: reference to free variable
    `nnmairix-valid-backends'

In nnmairix-update-and-clear-marks:
nnmairix.el:1262:31:Warning: reference to free variable
    `nnmairix-group-regexp'

In end of data:
nnmairix.el:1549:1:Warning: the function `gnus-registry-fetch-group' is not
    known to be defined.
Wrote gnus/lisp/nnmairix.elc


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: Type: text/x-patch, Size: 2046 bytes --]

--- nnmairix.el~	2008-02-26 21:42:09 +0000
+++ nnmairix.el	2008-02-27 02:32:03 +0000
@@ -336,18 +336,19 @@
 
 ;;; === Internal variables
 
-;; Regexp for mairix groups on back end
-(setq nnmairix-group-regexp (format "%s-\\(.*\\)-[0-9]+" nnmairix-group-prefix))
+(defvar nnmairix-group-regexp
+  (format "%s-\\(.*\\)-[0-9]+" nnmairix-group-prefix)
+  "Regexp for mairix groups on back end.")
+
+(defvar nnmairix-valid-backends '(nnimap nnml nnmaildir)
+  "Back ends (hopefully...) supported by nnmairix.
+Other backends might or might not work.")
 
-;; Back ends (hopefully...) supported by nnmairix.
-;; Other backends might or might not work.
-(setq nnmairix-valid-backends '(nnimap nnml nnmaildir))
+(defvar nnmairix-last-server nil
+  "Last chosen server.")
 
-;; Last chosen server
-(setq nnmairix-last-server nil)
-
-;; Current server
-(setq nnmairix-current-server nil)
+(defvar nnmairix-current-server nil
+  "Current server.")
 
 ;;; === Gnus backend functions
   
@@ -843,6 +844,9 @@
 	  (set-process-sentinel (apply 'start-process args)
 				'nnmairix-sentinel-mairix-update-finished))))))
 
+(eval-and-compile
+  (autoload 'gnus-registry-fetch-group "gnus-registry"))
+
 (defun nnmairix-goto-original-article (&optional no-registry)
   "Jump to the original group and display article.
 The original group of the article is first determined with the
@@ -974,7 +978,7 @@
   "Request FOLDER on backend for nnmairix QUALGROUP and article number correction."
   (save-excursion
     (nnmairix-call-backend
-     "request-group" folder nnmairix-backend-server fast)
+     "request-group" folder nnmairix-backend-server)
     (set-buffer nnmairix-mairix-output-buffer)
     (goto-char (point-min))
     (re-search-forward "^Matched.*messages")
@@ -1006,7 +1010,7 @@
 		(gnus-group-set-parameter
 		 qualgroup 'numcorr (list nil 0 high))))
 	    (erase-buffer)
-	    (insert (format "%d %d %d %d %s" status total low high group))
+	    (insert (format "%d %d %d %d %s" status total low high qualgroup))
 	    t)
 	(progn
 	  (nnheader-report

Re: Integration of nnmairix.el

[David]

Katsumi Yamaoka <yamaoka@jpl.org> writes:
> Sorry for finding fault with it.  But many warnings are issued
> when compiling.  Most of them look harmless, however the ones
> concerning `nnmairix-request-group-with-article-number-correction'
> seem to be real problems.  A patch is attached below, though I'm
> not sure of it.  I would appreciate reviewing it.

Thanks for looking into this. Patch looks good but in the last hunk it
should better read

+	    (insert (format "%d %d %d %d %s" status total low high (gnus-group-real-name qualgroup)))

I applied the other things and will release a new version soon.

Thanks again,
David

Re: Integration of nnmairix.el

[Wes Hardaker]

>>>>> "D" == David  <de_bb@arcor.de> writes:

D> I applied the other things and will release a new version soon.

So...  I installed mairix because of this and decided to give it a shot
and wow, it looks cool.

However...  nnmarix (0.5) itself causes big issues with my gnus setup.
Even just loading it causes problems regardless of whether I use it.

I have a number of different sort types set up in my group parameters
and upon loading nnmarix they all get shot and my sorting is shown
incorrectly for many of my folders for some reason.

About every other time I try to enter a group it says there is some
issue with the keybindings (but the second time works).
-- 
"In the bathtub of history the truth is harder to hold than the soap,
 and much more difficult to find."  -- Terry Pratchett

Re: Integration of nnmairix.el

[David]

Wes Hardaker <wes@hardakers.net> writes:
> However...  nnmarix (0.5) itself causes big issues with my gnus setup.
> Even just loading it causes problems regardless of whether I use it.
>
> I have a number of different sort types set up in my group parameters
> and upon loading nnmarix they all get shot and my sorting is shown
> incorrectly for many of my folders for some reason.

Could you please post the group parameters for a group where this
happens? Just do a 'G p' on the group and copy&paste. Nnmairix uses
quite a few group parameters, but it should never mess with the existing
ones.

> About every other time I try to enter a group it says there is some
> issue with the keybindings (but the second time works).

Please post the error message that is given. If it's an error, please do
(setq debug-on-error t) and post the resulting backtrace.

-David

Re: Integration of nnmairix.el

[Wes Hardaker]

>>>>> "D" == David  <de_bb@arcor.de> writes:

D> Wes Hardaker <wes@hardakers.net> writes:
>> However...  nnmarix (0.5) itself causes big issues with my gnus setup.
>> Even just loading it causes problems regardless of whether I use it.
>> 
>> I have a number of different sort types set up in my group parameters
>> and upon loading nnmarix they all get shot and my sorting is shown
>> incorrectly for many of my folders for some reason.

D> Could you please post the group parameters for a group where this
D> happens? Just do a 'G p' on the group and copy&paste. Nnmairix uses
D> quite a few group parameters, but it should never mess with the existing
D> ones.

I'll also admit here that I have a very "tricked out" gnus setup.
Including a plugin I wrote a while back to display the summary buffer as
it builds (as opposed to really-long-wait times; gives a bit of visual
feedback that it's doing something).

Here's the sorting parameters set on a topic for some of the groups with
issues.  Note the thread-sort functions setting:

  ((total-expire)
   (gnus-thread-hide-subtree nil)
   (gnus-large-newsgroup 20000)
   (gnus-use-scoring nil)
   (gnus-use-adaptive-scoring nil)
   (expiry-wait . 21)
   (display . [not expire])
   (gnus-summary-line-format "%uW%d%U%R%ub%I%(%[ %-15,15n%]%2t%) %-40s %uS\n")
   (wes-brokestuff "%z%d")
   (gnus-thread-sort-functions '(gnus-thread-sort-by-most-recent-number)))

(I'm using most-recent-number since it's faster than other sort methods
generally).

Here's a summary buffer of my 40 messages fedora nnimap folder with
nnmatrix loaded.  Note how the dates jump around a bit (there are some
25-Feb's in the middle).  (truncated to 75 columns for readability)

  25-Feb     221 [ bugzilla@redhat] 2 [Bug 434840] New: perl-Crypt-OpenSSL-
  25-FebE     50 [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-X509-0.6-
  25-FebE     50 [ Koji Build Syst] 1 dnssec-tools-1.3.2-1.fc8 successfully
  25-FebE     50 [ Koji Build Syst] 1 dnssec-tools-1.3.2-1.fc7 successfully
  19-FebE     50 [ Koji Build Syst] 1 Package: perl-Digest-SHA-5.45-2.fc9 T
  19-FebE     50 [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-Random-0.
  19-FebE     50 [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-X509-0.5-
  19-FebE     50 [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-RSA-0.25-
  19-FebE     50 [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-PKCS10-0.
  19-FebE     50 [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-Bignum-0.
  19-FebE     50 [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-DSA-0.13-
  19-FebE     50 [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-AES-0.02-
  18-FebE     50 [ Koji Build Syst] 1 dnssec-tools-1.3.2-1.fc7 successfully
  18-FebE     50 [ Koji Build Syst] 1 dnssec-tools-1.3.2-1.fc8 successfully
  15-FebE     50 [ Koji Build Syst] 1 dnssec-tools-1.3-4.fc7 successfully m
  15-FebE     50 [ Koji Build Syst] 1 Package: dnssec-tools-1.3.2-1.fc7 Tag
  15-FebE     50 [ Koji Build Syst] 1 Package: dnssec-tools-1.3.2-1.fc8 Tag
  15-FebE     50 [ Koji Build Syst] 1 Package: dnssec-tools-1.3.2-1.fc9 Tag
  15-FebE     50 [ Koji Build Syst] 1 Package: dnssec-tools-1.3.1-2.fc9 Tag
  15-FebE     50 [ Koji Build Syst] 1 Package: dnssec-tools-1.3.1-1.fc9 Tag
  25-FebE +    0 [ Wes Hardaker   ] 1 rpms/perl-Crypt-OpenSSL-X509/devel .c
  25-FebE      0 [ fedora-hams-req] 1 Welcome to the "Fedora-hams" mailing
  25-FebEA     0 [ fedora-hams-req] 1 confirm 8e76f8ff28fd9422a68d9cfc40a8b
  19-FebE      0 [ Flickner Pabst ] 1 talaria
  18-FebE      0 [ Jesse Keating  ] 1 rpms/perl-Digest-SHA/devel perl-Diges
  18-FebE      0 [ Jesse Keating  ] 1 rpms/perl-Crypt-OpenSSL-X509/devel pe
  18-FebE      0 [ Jesse Keating  ] 1 rpms/perl-Crypt-OpenSSL-Random/devel
  18-FebE      0 [ Jesse Keating  ] 1 rpms/perl-Crypt-OpenSSL-RSA/devel per
  18-FebE      0 [ Jesse Keating  ] 1 rpms/perl-Crypt-OpenSSL-PKCS10/devel
  18-FebE      0 [ Jesse Keating  ] 1 rpms/perl-Crypt-OpenSSL-DSA/devel per
  18-FebE      0 [ Jesse Keating  ] 1 rpms/perl-Crypt-OpenSSL-Bignum/devel
  18-FebE      0 [ Jesse Keating  ] 1 rpms/perl-Crypt-OpenSSL-AES/devel per
  15-FebE +    0 [ Wes Hardaker   ] 1 rpms/dnssec-tools/F-7 .cvsignore, 1.4
  15-FebE +    0 [ Wes Hardaker   ] 1 rpms/dnssec-tools/F-8 .cvsignore, 1.4
  15-FebE +    0 [ Wes Hardaker   ] 1 rpms/dnssec-tools/devel dnssec-tools.
  15-FebE +    0 [ Wes Hardaker   ] 1 rpms/dnssec-tools/devel sources,1.5,1
  15-FebE +    0 [ Wes Hardaker   ] 1 rpms/dnssec-tools/devel sources,1.4,1
  15-FebE +    0 [ Wes Hardaker   ] 1 rpms/dnssec-tools/devel dnssec-tools-
  15-FebE +    0 [ Wes Hardaker   ] 1 rpms/dnssec-tools/F-8 dnssec-tools.sp

Here is 40 from without nnmatrix loaded.  Note that in this one the
messages are the proper order (sorted by date, most recent first).

  25-Feb   [ bugzilla@redhat] 2 [Bug 434840] New: perl-Crypt-OpenSSL-X509-0
  26-FebE      [ bugzilla@redhat] 1 [Bug 434840] perl-Crypt-OpenSSL-X509-0.
  25-FebE  [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-X509-0.6-1.fc9
  25-FebE  [ Koji Build Syst] 1 dnssec-tools-1.3.2-1.fc8 successfully moved
  25-FebE +[ Wes Hardaker   ] 1 rpms/perl-Crypt-OpenSSL-X509/devel .cvsigno
  25-FebE  [ Koji Build Syst] 1 dnssec-tools-1.3.2-1.fc7 successfully moved
  25-FebE  [ fedora-hams-req] 1 Welcome to the "Fedora-hams" mailing list n
  25-FebEA [ fedora-hams-req] 1 confirm 8e76f8ff28fd9422a68d9cfc40a8bc1b782
  19-FebE  [ Koji Build Syst] 1 Package: perl-Digest-SHA-5.45-2.fc9 Tag: di
  19-FebE  [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-Random-0.04-4.f
  19-FebE  [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-X509-0.5-4.fc9
  19-FebE  [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-RSA-0.25-4.fc9
  19-FebE  [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-PKCS10-0.06-8.f
  19-FebE  [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-Bignum-0.04-3.f
  19-FebE  [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-DSA-0.13-6.fc9
  19-FebE  [ Koji Build Syst] 1 Package: perl-Crypt-OpenSSL-AES-0.02-4.fc9
  19-FebE  [ Flickner Pabst ] 1 talaria                                  no
  18-FebE  [ Jesse Keating  ] 1 rpms/perl-Digest-SHA/devel perl-Digest-SHA.
  18-FebE  [ Jesse Keating  ] 1 rpms/perl-Crypt-OpenSSL-X509/devel perl-Cry
  18-FebE  [ Jesse Keating  ] 1 rpms/perl-Crypt-OpenSSL-Random/devel perl-C
  18-FebE  [ Jesse Keating  ] 1 rpms/perl-Crypt-OpenSSL-RSA/devel perl-Cryp
  18-FebE  [ Jesse Keating  ] 1 rpms/perl-Crypt-OpenSSL-PKCS10/devel perl-C
  18-FebE  [ Jesse Keating  ] 1 rpms/perl-Crypt-OpenSSL-DSA/devel perl-Cryp
  18-FebE  [ Jesse Keating  ] 1 rpms/perl-Crypt-OpenSSL-Bignum/devel perl-C
  18-FebE  [ Jesse Keating  ] 1 rpms/perl-Crypt-OpenSSL-AES/devel perl-Cryp
  18-FebE  [ Koji Build Syst] 1 dnssec-tools-1.3.2-1.fc7 successfully moved
  18-FebE  [ Koji Build Syst] 1 dnssec-tools-1.3.2-1.fc8 successfully moved
  15-FebE  [ Koji Build Syst] 1 dnssec-tools-1.3-4.fc7 successfully moved f
  15-FebE  [ Koji Build Syst] 1 Package: dnssec-tools-1.3.2-1.fc7 Tag: dist
  15-FebE  [ Koji Build Syst] 1 Package: dnssec-tools-1.3.2-1.fc8 Tag: dist
  15-FebE  [ Koji Build Syst] 1 Package: dnssec-tools-1.3.2-1.fc9 Tag: dist
  15-FebE +[ Wes Hardaker   ] 1 rpms/dnssec-tools/F-7 .cvsignore, 1.4, 1.5
  15-FebE +[ Wes Hardaker   ] 1 rpms/dnssec-tools/F-8 .cvsignore, 1.4, 1.5
  15-FebE +[ Wes Hardaker   ] 1 rpms/dnssec-tools/devel dnssec-tools.spec,
  15-FebE +[ Wes Hardaker   ] 1 rpms/dnssec-tools/devel sources,1.5,1.6  ha
  15-FebE +[ Wes Hardaker   ] 1 rpms/dnssec-tools/devel sources,1.4,1.5 .cv
  15-FebE  [ Koji Build Syst] 1 Package: dnssec-tools-1.3.1-2.fc9 Tag: dist
  15-FebE +[ Wes Hardaker   ] 1 rpms/dnssec-tools/devel dnssec-tools-1.3.1-
  15-FebE  [ Koji Build Syst] 1 Package: dnssec-tools-1.3.1-1.fc9 Tag: dist
  15-FebE +[ Wes Hardaker   ] 1 rpms/dnssec-tools/F-8 dnssec-tools.spec,1.1


>> About every other time I try to enter a group it says there is some
>> issue with the keybindings (but the second time works).

D> Please post the error message that is given. If it's an error, please do

I admittedly didn't want to do that earlier, since once loading it I
need to restart XEmacs (which is what I'm using btw) to fix the issues.
(IE, nnmatrix needs to be unloaded).

Having gotten over my other pressing needs, I took the time this time:

Debugger entered--Lisp error: (error "Invalid prefix keys in sequence" ?$ [?$ ?t])
  define-key(#<keymap size 406 0x115b> [?$ ?t] nnmairix-search-thread-this-article)
  nnmairix-summary-mode-hook()
  run-hooks(nnmairix-summary-mode-hook)
  apply(run-hooks nnmairix-summary-mode-hook)
  run-mode-hooks(gnus-summary-mode-hook)
  apply(run-mode-hooks gnus-summary-mode-hook)
  gnus-run-mode-hooks(gnus-summary-mode-hook)
  gnus-summary-mode("nnimap+hardakers:mlists.ding")
  gnus-summary-setup-buffer("nnimap+hardakers:mlists.ding")
  gnus-summary-read-group-1("nnimap+hardakers:mlists.ding" 5 t nil nil nil)
  gnus-summary-read-group("nnimap+hardakers:mlists.ding" 5 t nil nil nil nil)
  gnus-group-read-group(5 t)
  gnus-group-select-group(5)
  gnus-topic-select-group(5)
  call-interactively(gnus-topic-select-group)

-- 
"In the bathtub of history the truth is harder to hold than the soap,
 and much more difficult to find."  -- Terry Pratchett

Re: Integration of nnmairix.el

[David]

Wes Hardaker <wes@hardakers.net> writes:
> D> Could you please post the group parameters for a group where this
> D> happens? Just do a 'G p' on the group and copy&paste. Nnmairix uses
> D> quite a few group parameters, but it should never mess with the existing
> D> ones.

> Here's a summary buffer of my 40 messages fedora nnimap folder with
> nnmatrix loaded.  Note how the dates jump around a bit (there are some
> 25-Feb's in the middle).  (truncated to 75 columns for readability)

And this happens solely because of (require 'nnmairix)? All this require
does is to put some key definitions into the summary/group-hook and to
define the new back end. I have no idea why this would mess with the
article numbers or sort functions in other back ends...

> Having gotten over my other pressing needs, I took the time this time:
>
> Debugger entered--Lisp error: (error "Invalid prefix keys in sequence" ?$ [?$ ?t])

Xemacs doesn't seem to like the '$' prefix. This will be changed anyway,
but until then you can simply edit the nnmairix-summary-mode-hook in
nnmairix.el to use something else.

-David

Re: Integration of nnmairix.el

[Wes Hardaker]

>>>>> "D" == David  <de_bb@arcor.de> writes:

>> Here's a summary buffer of my 40 messages fedora nnimap folder with
>> nnmatrix loaded.  Note how the dates jump around a bit (there are some
>> 25-Feb's in the middle).  (truncated to 75 columns for readability)

D> And this happens solely because of (require 'nnmairix)? All this require
D> does is to put some key definitions into the summary/group-hook and to
D> define the new back end. I have no idea why this would mess with the
D> article numbers or sort functions in other back ends...

Yes.  It blew me away too that just loading it could cause such a problem.

D> Xemacs doesn't seem to like the '$' prefix. This will be changed
D> anyway, but until then you can simply edit the
D> nnmairix-summary-mode-hook in nnmairix.el to use something else.

Thanks...  (G B isn't taken)
-- 
"In the bathtub of history the truth is harder to hold than the soap,
 and much more difficult to find."  -- Terry Pratchett

Re: Integration of nnmairix.el

[Reiner Steib]

On Wed, Feb 27 2008, David wrote:
> Katsumi Yamaoka <yamaoka@jpl.org> writes:
>> A patch is attached below, though I'm not sure of it.  I would
>> appreciate reviewing it.
>
> Thanks for looking into this. Patch looks good but in the last hunk it
> should better read [...]

Katsumi, please commit these changes.

> I applied the other things and will release a new version soon.

Please post a diff (-u) against the version in CVS.  Thus we won't
lose changes made in CVS.  If you'd like to maintain `nnmairix.el' in
Gnus CVS, feel free to ask Lars for write access.

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

Re: Integration of nnmairix.el

[Katsumi Yamaoka]

>>>>> Reiner Steib wrote:
> On Wed, Feb 27 2008, David wrote:
>> Katsumi Yamaoka <yamaoka@jpl.org> writes:
>>> A patch is attached below, though I'm not sure of it.  I would
>>> appreciate reviewing it.
>>
>> Thanks for looking into this. Patch looks good but in the last hunk it
>> should better read [...]

> Katsumi, please commit these changes.

Done.

>>>>> David wrote:
> Thanks for looking into this. Patch looks good but in the last hunk it
> should better read

> +	    (insert (format "%d %d %d %d %s" status total low high (gnus-group-real-name qualgroup)))

I did so.  Thanks.  Now all the versions of (X)Emacsen that Gnus
supports issue no warning against nnmairix.el.

Regards,

Re: Integration of nnmairix.el

[Reiner Steib]

On Wed, Feb 27 2008, Reiner Steib wrote:

> Please post a diff (-u) against the version in CVS.  Thus we won't
> lose changes made in CVS.  

Additionally, the Emacs developers made several changes:

,----[ <http://mid.gmane.org/E1JUIRm-0005OD-MB@cvs.savannah.gnu.org> ]
|  2008-02-27  Glenn Morris  <rgm@gnu.org>
| 
|  	* nnmairix.el: Change defcustom :version from 23.0 to 23.1.
| +	(nnmairix-group-regexp, nnmairix-valid-backends): Convert from free
| +	variables to defconsts.  Convert comments to doc-strings.
| +	(nnmairix-last-server, nnmairix-current-server): Convert from free
| +	variables to defvars.  Convert comments to doc-strings.
| +	(gnus-registry-fetch-group): Autoload.
| +	(nnmairix-replace-group-and-numbers): Use mapc rather than mapcar.
| +	(nnmairix-widget-get-values, nnmairix-widget-make-query-from-widgets)
| +	(nnmairix-widget-build-editable-fields): Use car cddr rather than caddr.
| +	(nnmairix-request-group): Bind nnmairix-fast and nnmairix-group around
| +	nnmairix-request-group-with-article-number-correction call.
| +	(nnmairix-fast, nnmairix-group): New, less general names, for free
| +	variables passed from nnmairix-request-group to
| +	nnmairix-request-group-with-article-number-correction.  Declare.
| +	(nnmairix-request-group-with-article-number-correction):
| +	Use nnmairix-fast, nnmairix-group rather than fast, group.
`----

BTW, do you read this list/newsgroup or would you like to be cc-ed?

> If you'd like to maintain `nnmairix.el' in Gnus CVS, feel free to
> ask Lars for write access.

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

Re: Integration of nnmairix.el

[Miles Bader]

Reiner Steib <reinersteib+gmane@imap.cc> writes:
>> Please post a diff (-u) against the version in CVS.  Thus we won't
>> lose changes made in CVS.  
>
> Additionally, the Emacs developers made several changes:
>
> |  2008-02-27  Glenn Morris  <rgm@gnu.org>
> | 
> |  	* nnmairix.el: Change defcustom :version from 23.0 to 23.1.

Note that I reverted some of those changes in favor of cleaner changes
made by Kasumi et al.

All should be merged into Gnu CVS, in any case.

-Miles

-- 
Saa, shall we dance?  (from a dance-class advertisement)

Re: Integration of nnmairix.el

[David]

Reiner Steib <reinersteib+gmane@imap.cc> writes:
> Additionally, the Emacs developers made several changes:

Yes, thanks for mentioning that.

> BTW, do you read this list/newsgroup or would you like to be cc-ed?

I read it through gmane, so no need to CC me.

>> If you'd like to maintain `nnmairix.el' in Gnus CVS, feel free to
>> ask Lars for write access.

I will do that.

-David

Re: Integration of nnmairix.el

[David]

Reiner Steib <reinersteib+gmane@imap.cc> writes:
>>> - Documentation.  Minimally the content from
>>>   <http://www.emacswiki.org/cgi-bin/wiki/GnusMairix> should be
>>>   converted to texinfo to fit into `texi/gnus.texi'.
>>
>> OK, I'll do that.

... and I'm working on it. :-) I will release a new version soon which
more or less contains all the features which I had in mind when I
started with nnmairix, especially the possibility to propagate marks
from articles in nnmairix groups to the original ones. This will make it
possible to actually use nnmairix groups for incoming mail (as an
alternative to splitting). When that's finished I'll write up the
documentation. 

> We also want to install nnir.el soon.  nnir.el uses `G G' (only a
> single binding).  (And there's contrib/gnus-namazu.el which uses `C-c
> C-n' when `gnus-namazu-insinuate' is called).  Maybe we should think
> about using some two key stroke prefix for mail search facilities (I'm
> not sure if it makes sense to use nnir.el and nnmairix.el at the same
> time).

Some people might want to use nnir.el e.g. for IMAP and nnmairix.el for
locally stored archive groups, so it should be possible to use both.

> For the Group mode, we could use e.g. `G b', `G B' or `G G' as prefix
> for (mail) search.  In the Summary mode, we may use the same prefix or
> a different one.  I don't have a strong opinion on this.  (`G b' runs
> `gnus-summary-best-unread-article', but it is also on `,', so we may
> steal this or move it to `G B'.)  Opinions?

Using the same prefix in group and summary mode is a good idea.

>> so maybe users should have to explicitly put (require 'nnmairix)
>> into their .gnus for activating them.
>
> Loading a package should not perform changes like this.  There should
> be an autoloaded initialization function instead.  (If detecting a
> working mairix installation is possible, we could install it from
> `gnus.el'/`gnus-start.el' depending on the result.)

We could simply check for the existence of the configuration file
~/.mairixrc, although that's of course configurable and also wouldn't
catch the case where mairix is called remotely. Maybe we could just
enable 'G b c' for creating a nnmairix server with the default search
group (one could put this in the menu under "Foreign Groups"). The rest
would then only be enabled at start up if a nnmairix server exists.

-David