headers
Ron Isaacson | 7 Sep 2006 08:25
Favicon

IMAP hangs on Win32, XEmacs 21.5

Hi all --

I'm testing WL on a local native Win32 build of XEmacs 21.5.26, and I
find that attempts to open an IMAP connection simply hang. This works
just fine under 21.4.14 on the same machine with all of the same
configuration. I set debug-on-quit and interrupted it to get the
following stack trace:

  Debugger entered--Lisp error: (quit)
    elmo-imap4-read-response([elmo-imap4-session [0 0 0 0 0 0 0] "IMAP" "myimapserver" 7143 "myusername"
clear nil #<network connection "IMAP" (7143 . "myimapserver") state:exit> nil (17663 45269 15000) nil
nil nil nil] "elmo-imap41")
    elmo-imap4-accept-ok([elmo-imap4-session [0 0 0 0 0 0 0] "IMAP" "myimapserver" 7143 "myusername" clear
nil #<network connection "IMAP" (7143 . "myimapserver") state:exit> nil (17663 45269 15000) nil nil nil
nil] "elmo-imap41")
    elmo-imap4-send-command-wait([elmo-imap4-session [0 0 0 0 0 0 0] "IMAP" "myimapserver" 7143
"myusername" clear nil #<network connection "IMAP" (7143 . "myimapserver") state:exit> nil (17663
45269 15000) nil nil nil nil] "capability")
    #<compiled-function (session) "...(115)" [elmo-imap4-status process session 8 process-buffer
process-status (open run) -1 elmo-imap4-parse-greeting accept-process-output 1 erase-buffer
set-process-filter elmo-imap4-arrival-filter set-process-sentinel elmo-imap4-sentinel
(nonauth auth) signal elmo-open-error elmo-network-initialize-session 11 capability
elmo-imap4-send-command-wait "capability" 7 starttls (elmo-imap4-starttls-error) "starttls"
starttls-negotiate] 6>([elmo-imap4-session [0 0 0 0 0 0 0] "IMAP" "myimapserver" 7143 "myusername"
clear nil #<network connection "IMAP" (7143 . "myimapserver") state:exit> nil (17663 45269 15000) nil
nil nil nil])
    apply(#<compiled-function (session) "...(115)" [elmo-imap4-status process session 8
process-buffer process-status (open run) -1 elmo-imap4-parse-greeting accept-process-output 1
erase-buffer set-process-filter elmo-imap4-arrival-filter set-process-sentinel
elmo-imap4-sentinel (nonauth auth) signal elmo-open-error elmo-network-initialize-session 11
(Continue reading)

Permalink | Reply | 2 comments |
headers
Ron Isaacson | 8 Sep 2006 08:26
Favicon

Re: IMAP hangs on Win32, XEmacs 21.5

OK, I found the problem here. If you ever plan to use XEmacs 21.5 for
anything, read on...

I did some debugging and saw that in the trace buffer,
(elmo-imap4-find-next-line) was returning nil, meaning it didn't think
it had received a complete line. Looking at the function, I saw it was
using elmo-imap4-server-eol (\r\n) to mark the end of the line, but
the lines in my buffer appeared to end only in \n.

So first I tried setting elmo-imap4-server-eol to \n, and that let me
get a little bit further, but not very much. When reading a response
block (a single line, with a count at the end indicating how many
bytes of response follow), now elmo-imap4-find-next-line was actually
reading past the end of the block and into the following response.
Turns out the server was including the \r's in the count of bytes it
was sending, but somehow they were disappearing before the response
got into my buffer.

The problem turned out to be in XEmacs 21.5's fancy new encoding
support. Different encodings are better supported than ever before,
but somebody made what I think is a braindead decision in the
implementation of open-network-stream. From the new docs:

  The read/write coding systems used for process I/O on the process are
  determined as follows:

  1. `coding-system-for-read', `coding-system-for-write', if non-nil.
        (Intended as a temporary overriding mechanism for use by Lisp
        code.)
  2. The matching value for the service from `network-coding-system-alist',
(Continue reading)

Permalink | Reply | 0 comments |
headers
Yavor Doganov | 17 Sep 2006 22:57
X-Face

Re: Improper display of messages with no declared encoding

At Sun, 27 Aug 2006 23:57:45 +0300,
Yavor Doganov wrote:
> 
> I have some troubles with few mailing lists such as
> www-commits <at> gnu.org.  Wanderlust displays the 8bit symbols in the
> message as
> 
> Забравил съм да комитнÐ
> 
> I noticed that those mails do not have a Content-Type: header, while
> the encoding is UTF-8.

I figured that out by setting wl-mime-charset to `utf-8'.  However, in
the Summary, my Pet names are displayed with UTF-8 escaped sequences.
This is not a disaster, since I can use only us-ascii names, but the
symbols that form the threads are displayed so and this is a nuisance.

Is there any other way to handle this properly?
Permalink | Reply | 0 comments |
headers
Lloyd Zusman | 18 Sep 2006 02:11

Remote sendmail via ssh?

Is there a way to tell Wanderlust to send outgoing mail via a remote
instance of sendmail that is accessed via ssh?  By "remote", I mean that
sendmail is running on a host that is different from the one on which
Wanderlust is running.

Thanks in advance.

--

-- 
 Lloyd Zusman
 ljz <at> asfast.com
 God bless you.
Permalink | Reply | 4 comments |
headers
bremner | 18 Sep 2006 02:36
Picon

Re: Remote sendmail via ssh?

>>>>> "Lloyd" == Lloyd Zusman <ljz <at> asfast.com> writes:

    Lloyd> Is there a way to tell Wanderlust to send outgoing mail via
    Lloyd> a remote instance of sendmail that is accessed via ssh?  By
    Lloyd> "remote", I mean that sendmail is running on a host that is
    Lloyd> different from the one on which Wanderlust is running.

Yes, you should be able to adapt the following:

(defvar db-this-host (getenv "HOST") "Name of current host")

(defun db-exec-string (host cmd)
  (cond ((string-equal host db-this-host) cmd )
	(t (concat "ssh " host " " cmd))))

;; Smtp via ssh
(setq smtp-open-connection-function 'smtp-open-shell-stream)
(setq wl-smtp-posting-server "remotehost.remotedomain")
(setq wl-smtp-posting-port 25)

(defun smtp-open-shell-stream (name buffer host service)
  (start-process-shell-command
    name buffer (db-exec-string host "/usr/sbin/exim4 -bs")))
Permalink | Reply | 2 comments |
headers
Lloyd Zusman | 18 Sep 2006 02:59

Re: Remote sendmail via ssh?

bremner <at> unb.ca writes:

>>>>>> "Lloyd" == Lloyd Zusman <ljz <at> asfast.com> writes:
>
>     Lloyd> Is there a way to tell Wanderlust to send outgoing mail via
>     Lloyd> a remote instance of sendmail that is accessed via ssh?  By
>     Lloyd> "remote", I mean that sendmail is running on a host that is
>     Lloyd> different from the one on which Wanderlust is running.
>
> Yes, you should be able to adapt the following:
>
> (defvar db-this-host (getenv "HOST") "Name of current host")
>
> (defun db-exec-string (host cmd)
>   (cond ((string-equal host db-this-host) cmd )
> 	(t (concat "ssh " host " " cmd))))
>
> ;; Smtp via ssh
> (setq smtp-open-connection-function 'smtp-open-shell-stream)
> (setq wl-smtp-posting-server "remotehost.remotedomain")
> (setq wl-smtp-posting-port 25)
>
> (defun smtp-open-shell-stream (name buffer host service)
>   (start-process-shell-command
>     name buffer (db-exec-string host "/usr/sbin/exim4 -bs")))

Thank you!

I forgot to mention that I only want to do this within certain folders,
and in the others, to just send outgoing mail via my standard local smtp
(Continue reading)

Permalink | Reply | 0 comments |
headers
olin.wanderlust.7ia | 23 Sep 2006 04:11

accessing nested imap folders

------- Start of forwarded message -------
I would very much like to use wanderlust on my computers due to its
support for disconnected operation. I've installed it, along with
the necessary emacs support packages, and am using it from emacs 21.

My mail host is running the courier imap server.

I am able to create (using m a) the mailbox
  inbox:shivers/login <at> lambda.csail.mit.edu:993!
In this mailbox, I see mail.

So far, so good. What I *can't* do is work with nested folders.

I used evolution to create a folder inside my inbox. So, according to
evolution, I have the following tree:
  Inbox
    friends
  Junk
  Trash

Wanderlust doesn't show me Junk, Trash, or Inbox/friends.

I poked around the mail archives for wl-en, and found some similar problems.
None of the solutions work for me.

- - I tried to make the folder
      :shivers/login <at> lambda.csail.mit.edu:993!
  This died with the standard failure where wl says "Folder does not exist,
  create it? (y or n)" and then dying when I say y with "IMAP error: Invalid
  mailbox name."
(Continue reading)

Permalink | Reply | 16 comments |
headers
Yoichi NAKAYAMA | 23 Sep 2006 23:55
X-Face
Gravatar

Re: accessing nested imap folders

At Fri, 22 Sep 2006 19:11:14 -0700 (PDT),
olin.wanderlust.7ia <at> shivers.mail0.org wrote:
> My mail host is running the courier imap server.
> 
> I am able to create (using m a) the mailbox
>   inbox:shivers/login <at> lambda.csail.mit.edu:993!
> In this mailbox, I see mail.
> 
> So far, so good. What I *can't* do is work with nested folders.
> 
> I used evolution to create a folder inside my inbox. So, according to
> evolution, I have the following tree:
>   Inbox
>     friends
>   Junk
>   Trash
> 
> Wanderlust doesn't show me Junk, Trash, or Inbox/friends.
> 
> I poked around the mail archives for wl-en, and found some similar problems.
> None of the solutions work for me.
> 
> - - I tried to make the folder
>       :shivers/login <at> lambda.csail.mit.edu:993!
>   This died with the standard failure where wl says "Folder does not exist,
>   create it? (y or n)" and then dying when I say y with "IMAP error: Invalid
>   mailbox name."

To create a group with all folders under certain place, use "m A" instead of "m a".
(Continue reading)

Permalink | Reply | 14 comments |
headers
olin.wanderlust.7ia | 24 Sep 2006 19:14

Re: accessing nested imap folders

    From: Yoichi NAKAYAMA <yoichi <at> geiin.org>
    To create a group with all folders under certain place, use "m A" instead of "m a".

    How does the following work? (using "." as separator)
	 inbox.friends:shivers/login <at> lambda.csail.mit.edu:993!

Fantastic. It works! Thanks, Yoichi.

I have some other questions, in my project to upgrade my email technology. 
Here's my situation.

First, I'm a die-hard emacs user. I've been working in emacs for > 25 years,
and wrote some of its code. It is my primary interface to text; I want my
mail system to live inside emacs. That rules out evolution & thunderbird.

So I've used rmail up to now. I have about 1.7 Gb of email in babyl format.
(Well, I have more, but the really old mail is archived separately.)
I keep my main mailbox in one file, and have a distinct directory where I
file mail once I've read it. My main rmail file is about 130Mb, and the rest
is filed in about 1200 different files in my mail directory. My workflow is
that once I've read and acted on a message, I move it to a file named after
the author of the message.

There are some nice features about this setup -- it's "neater" and simpler
to keep all the messages for one mailbox in a single file, rather than a
complicated maildir directory setup. But it's breaking down from age and
overload.

- First, rmail is dying of age & neglect. Xemacs's implementation is broken --
  it incorrectly implements the babyl format and will die on legal rmail
(Continue reading)

Permalink | Reply | 12 comments |
headers
olin.wanderlust.7ia | 24 Sep 2006 19:54

Wanderlust & courier co-managing a maildir folder tree

I have a maildir-format folder tree in ~/Maildir. 
- Postfix delivers to the inbox.
- I remotely access the entire folder tree with courier imap.

I tried to locally access the folder tree with wanderlust, specifying a
wanderlust folder of just "." This works, but wanderlust gets confused 
by all the random file structure courier uses. So rather than showing
me the folder tree, which (for my test setup) is basically this

     Inbox
     Morefriends
     friends
     rmail-test
     tbird
       tb-a

I instead see that *plus* all the little management files courier keeps around,
spuriously shown as folders:

  [-].:3/6559/6581
     .:3/3/5
     ..Morefriends:0/0/1
     ..Morefriends/courierimapkeywords:*/*/*
     ..friends:*/*/*
     ..friends/courierimapkeywords:*/*/*
     ..rmail-test:0/6556/6575
     ..rmail-test/courierimapkeywords:*/*/*
     ..tbird:*/*/*
     ..tbird/courierimapkeywords:*/*/*
     ..tbird.tb-a:*/*/*
(Continue reading)

Permalink | Reply | 2 comments |

Gmane