kazu-yamamoto / mew Goto Github PK

mew-6.5/info/mew.info-4's Variable Index has reference to mew-from.
But destination page has no mew-from.

Please remove or change link destination.

Hi, I'm using "Mew version 6.6" on Ubuntu 12.04.

Today I encountered the title's error after I started M-x mew command.
It seems the mew issued the error during mew-status-update.
I made sure this error at emacs-version 24.3.1(from apt-get emacs24) and 25.0.50.1(from git
and emacs-24 branch).
(I did "apt-get update && apt-get upgrade" before I make sure it.
So the emacs 24.3.1 is probably latest.)

I found mew-called-interactively-p definition in mew-env.el (line 318 - 325)

(cond
 ((>= emacs-major-version 24)
  ;; this must be macro. If implemented as a function, its behavior
  ;; is changed.
  (defmacro mew-called-interactively-p ()
    '(called-interactively-p 'interactive)))
 (t
  (defalias 'mew-called-interactively-p 'called-interactively-p)))

Somehow, I suppressed this error at myself by below code:

(defalias 'mew-called-interactively-p 'called-interactively-p)

I checked Emacs's modifications by (view-emacs-news 24), but I
couldn't find about the called-interactively-p. So I'm not sure my
modification is correct or not. Anyway I report you this problem.

Regards

Since important features have been merged and documentation was radically improved, I guess that we should release v6.10.

@tats Could you get #172 done?

With Google Apps, an SMTP password is cached incorrectly, reported in
[mew-dist 29609] http://www.mew.org/pipermail/mew-dist/2013-June/029298.html.

If mew-smtp-user is "[email protected]", mew-smtp-passtag() returns
"[email protected]". If mew-smtp-user is "[email protected]" (same
user, different domain), the passtag is same, which causes
incorrect caching.

In mew-smtp-passtag(), replacing mew-smtp-get-user with
mew-smtp-get-auth-user seems to fix this problem.

(Using mew-smtp-get-user seems intentional, I don't know why.)

OS: 64bit Windows 10 version 1909
Emacs: 27.1 (self build with MSYS2)
Mew: HEAD of master branch of Git repository

Let's assume user mistyped password. If it is either IMAP or POP then message is displayed in minibuffer that password is wrong. On the other hand if it is SMTP then message is displayed with popped up window. It is because message-box function is used to display message. On Window the popped up window often disappears before user clicks it (for example focus accidentally moved to other window). Then following message is displayed in minibuffer.

"error in process sentinel: Quit"

And after that user can't send mail with SMTP until he restarts Emacs.

Need to update Japanese text for #89.

It would be nice if VCALENDAR attachments were displayed in a nice form (possibly inline, at the bottom of the mail) and if it was possible to insert these events into the calendar with one keystroke.

Some ideas for Gnus: http://lists.gnu.org/archive/html/info-gnus-english/2011-10/msg00046.html

I use Mew 6.5 from released tar ball.
Reaply-To of e-mail is ignored when mew-config-alist has mail-domain setting in ~/.mew.

My ~/.emacs is here.
(setq load-path
(append (list
(expand-file-name "/opt/elisp/mew-6.5")
)
load-path))

(autoload 'mew "mew" nil t)
(autoload 'mew-send "mew" nil t)

And my ~/.mew is here.
(setq mew-config-alist
'(
(default
(user "ryo_on")
(name "Ryo ONODERA")
(mail-domain "yk.rim.or.jp")
)
)
)

Could you please push v6.7rc2 and v6.7 to GitHub?

Hi,

I don't really know if this is a bug...But Mew just does not support non-ascii characters while searching in the folder.
It seems that Mew uses grep to grep the email files. Since non-ascii characters are encoded in some manner, grep just can't search them....

Any solutions?

Hi,

when the mewencode binary is corrupt, mew sends the error message it gets from calling mewencode as mail.

In my example mew sent mails containing only this line (using mew 6.7):

emacs: /home/roklein/.emacs.d/mew/bin/mewencode: Exec format error

I got it working using the following diff (though I don't know currently which of the patched code paths is used):

diff --git a/mew-encode.el b/mew-encode.el
index 6dd3ad1..8da33d7 100644
--- a/mew-encode.el
+++ b/mew-encode.el
@@ -703,7 +703,13 @@
            (write-region beg (point-max) file nil 'no-msg))
          (delete-region beg (point-max)))
        (mew-piolet mew-cs-text-for-read mew-cs-dummy
-         (apply 'call-process mew-prog-mime-encode file t nil opt))
+          (let ((return-code
+                 (apply 'call-process mew-prog-mime-encode file t nil opt)))
+            (when (not (and (integerp return-code)
+                            (eql 0 return-code)))
+              (mew-encode-error
+               (concat mew-prog-mime-encode " returns invalid results")))
+            return-code))
        (if textp (mew-delete-file file))))
      (t
       (mew-encode-error (concat mew-prog-mime-encode " does not exist"))))
@@ -1150,7 +1156,13 @@
            (write-region (point-min) (point-max) file nil 'no-msg)
            (delete-region (point-min) (point-max)))
          (mew-piolet mew-cs-text-for-read mew-cs-dummy
-           (apply 'call-process mew-prog-mime-encode file t nil opt))
+            (let ((return-code
+                   (apply 'call-process mew-prog-mime-encode file t nil opt)))
+              (when (not (and (integerp return-code)
+                              (eql 0 return-code)))
+                (mew-encode-error
+                 (concat mew-prog-mime-encode " returns invalid results")))
+              return-code))
          (mew-delete-file file)))
        (t
        (mew-encode-error (concat mew-prog-mime-encode " does not exist")))))))

PS: patching the mew-prog-mime-decode calls in mew-decode.el and mew-summary4.el did not work; I get a "MIME decoding error: " instead for the same corrupt mewencode binary.

Best regards
Robert

Hi,

I'd like to report these 2 bugs related to Hyper Estraier search.

regex in mew-search-virtual-with-est

In mew-search-virtual-with-est the list of Hyper Estraier results is not parsed correctly when mew-mail-path ends in a directory name that also appears on the IMAP host, for example when mew-mail-path is "~/Mail", and the IMAP server contains a directory "%Mail/Github".

The bug is due to the regex used:

 (format "file://.*/%s/\\(.*\\)/\\([0-9]+\\)\\(%s\\)?$"
            (file-name-nondirectory mew-mail-path))

In the case, when mew-search-est results contain messages from %Mail/Github, the regex will eat the complete path until "Github". and these results are not displayed in the Mew virtual folder.

As a workaround, I renamed my local folder to "~/Maillocal"

Replacing the .* for the first part of the path with the non-greedy .*? seems to solve the problem (I've made a pull request).

string formatting in mew-search-est

In mew-search-est, the following parameter is passed to estcmd search.

 (format "@uri STRINC %s" (mew-q-encode-string path ?%) )

If path contains _ character, these are encoded as %5F, but it looks like estcmd does not expect this encoding (using Hyper Estraier 1.4.13). For example, if mew-mail-path is "~/Mail_local", the generated Hyper Estraier command in mew-search-est is

 estcmd search -attr "@uri STRINC Mail%5Flocal/" -vu -max -1 -ord "@cdate NUMA" /home/username/Mail_local/casket/ "pattern"

which gives 0 results. Using the unencoded path instead, like

 estcmd search -attr "@uri STRINC Mail_local/" -vu -max -1 -ord "@cdate NUMA" /home/username/Mail_local/casket/ "pattern"

gives the correct results. Perhaps mew-q-encode-string is not required there?

Thank you very much,

Thomas

If the next message in my summary buffer is signed with S/MIME, the verifying process hangs forever, e.g.

gpgsm --verify --include-certs 3 --status-fd 1 /tmp/shTE9rJr/mew8mlHIc.sig /tmp/shTE9rJr/mew8mlHIc

Is there a way to avoid these verifications? Or better: how to fix the gpgsm (from GnuPG 2.2.12) call?

Emacs ver: This is GNU Emacs 27.1 (build 1, x86_64-apple-darwin19.6.0, NS appkit-1894.60 Version 10.15.7 (Build 19H114))
of 2020-12-27
Mew ver: from Github on Jan 11, 2021
Server: Gmail IMAP server via stunnel

Situation:

• Emacs27.1 on OSX (10.15.7) with latest Mew.
• .emacs, .mew.el and other configuration files are working configuration on the other machine.
• Folder initilization run C-uZ to get IMAP folder information from the server.
• When I run C-uZ, the process complete without any error but nothing happen.
• IMAP communication includes auth worked well which was checked by sending email (tcpdump and log)
  (confirmed successful copy the email in Sent folder via IMAP port which means stunnel is working over IMAP)
• Looks some process doesn't work or skip in mew.el between line 241 and 262.
• There is no message when running with mew-debug flag
• I guess the process returned before accessing IMAP since there is no packet confirmed by C-uZ

Please let me know if anything I should check.

Thank you!

Hi,

I've been using Mew for almost two years, but Now I found a functional bug when using Mew with GPG 2.1.4, i.e. "Master password is wrong".

Because it's not actually a bug, so it cannot be backtraced.
That happens when mew-passwd-load or mew-passwd-save tries to start-process, but process filter cannot get any result from GPG, so Mew would not prompt user for input password. Therefore, master password is never set.

The reason is, from GnuPG 2.1, it uses pinentry as its agent program.

The agent program does two things which cause the problem:

• Accept user input via graphic message box or cursor-based input box.
  GPG will not output anything to starndard output.
• Cache the master password for some time.
  GPG will directly use cached password without any prompt, only below:
  gpg: AES encrypted data
  gpg: encrypted with 1 passphrase

Both makes mew-passwd-filter has no chance to prompt the user, so Mew can't get any master password from user, mew-passwd-master is nil.
Then, when mew-passwd-save or mew-passwd-load is called, it will say "Master password is wrong".

Possible solution:

• For GPG agent:
  GnuPG provides one option named --pinentry-mode which can be set to loopback to force GnuPG use stdout and prompt password traditionally.
  One configure must be enabled in ~/.gnupg/gpg-agent.conf, with line below:
  allow-loopback-pinentry
  Also, when Mew start-process, it must provide option --pinentry-mode with
  value loopback.
  After doing this, Mew is able to get output from GPG.
• For password cache:
  GPG will output following two lines when using cached password:

gpg: AES encrypted data
gpg: encrypted with 1 passphrase

Inside `mew-passwd-filter', capture this situation and prompt for
master password if nil.

Talk is cheap, here's the code works for me:
(last two could be a little bit ugly... ;-)

• Inside mew-passwd-filter', within thecond' clause:

((string-match "encrypted with" string)
 (unless mew-passwd-master
   (mew-passwd-read-passwd "Master password: ")))

• Function for get the version of GPG.

(defun mew-passwd-check-gpg-version ()
  (let ((version (shell-command-to-string "gpg --version")))
    (setq version (substring version 0 (string-match (string ?\n) version)))
    (string-to-number (substring version (string-match "[1-9]" version)))))

• Inside `mew-passwd-save', when starting process.

    (setq pro (mew-start-process-lang
           mew-passwd-encryption-name (current-buffer) mew-prog-passwd
           (when (>= (mew-passwd-check-gpg-version) 2.1)
              "--pinentry-mode")
           (when (>= (mew-passwd-check-gpg-version) 2.1)
              "loopback")
       "-c" "--cipher-algo" mew-passwd-cipher
       "--yes" "--output" file tfile))

• Like 3, but in `mew-passwd-load'.

I hope it does make sense somehow.

Best wishes,
Danny

When mew tries to decrypt .mew-passwd.gpg it hangs forever when no graphical (i.e. gtk/qt) pinentry program is installed and gpg-agent is in use. I use pinentry-curses usually but when I additionally install pinentry-gtk2 I get a graphical prompt instead of it just hanging.

(setenv "GPG_AGENT_INFO" "") before starting mew seems to work around the issue however this is unsatisfactory as then gpg-agent's functionality is broken in any terminal started as a child process of emacs. When using this workaround I get the nice password prompt in emacs' minibuffer as I'd expect (and not the hang or a graphical prompt).

Sorry I can't provide more detailed information but the complete hang of emacs makes this bug very hard to debug, maybe you have at least some idea why this might be happening and what can be done about it.

Hi,
I started recently to have problems when signing messages in mew.
Not sure exactly what caused the change but I had to change my gpg and gpg-agent prefs to have mew happy. Furthermore mew is not able to parse the messages of my gpg.

gpg (GnuPG) 2.2.17
libgcrypt 1.8.5
Copyright (C) 2019 Free Software Foundation, Inc.

It seems that gpg answers do not follow what mew expect. After some investigation, I discovered how to setup debuging for gpg in mew and get the following information in the mew-debug buffer:

<PGP filter1>
[GNUPG:] KEY_CONSIDERED 027F76FFE5E74CFBCA473FDCC6C140BC32FC9604 2
[GNUPG:] BEGIN_SIGNING H10
[GNUPG:] USERID_HINT C6C140BC32FC9604 Dagnat Fabien <[email protected]>
[GNUPG:] NEED_PASSPHRASE C6C140BC32FC9604 C6C140BC32FC9604 1 0
[GNUPG:] INQUIRE_MAXLEN 100
[GNUPG:] GET_HIDDEN passphrase.enter

Clearly no case of mew-pgp.el function mew-pgp-process-filter1 matches this message. To make it work I had to patch mew-pgp-msg-enter-pass

(defconst mew-pgp-msg-enter-pass
  '("Enter pass phrase: "
    "Enter pass phrase: "
    "Enter pass phrase: "
    "Enter passphrase: "
    "passphrase.enter"))

I'm reluctant to submit a patch because I'm pretty sure it is not the right way of solving the problem, it is not sufficient, I have not found any description of the protocol external tools should follow when invoking gpg and I know too little in elisp...
Perhaps, if someone provides help, I would be able to propose a real patch.

Fabien

This is due to #121.
I'm now trying to fix this.

After commit cf354e4 is merged to master branch, make elisp fails with error messages as below.

Error: Symbol’s function definition is void: mew-match-string

I confirmed it with following two environments.

OS: FreeBSD 13.1-RELEASE amd64
Emacs: commit 0330cff65ae of upstream master branch

OS: 64bit Windows 10 22H2
Emacs: commit 95d5154feed of upstream master branch

Files mew-darwin.el, mew-unix.el, and mew-win32.el serve similar purposes, but it is difficult to understand how they differ, and the diffs between them are large and unweildy. Many of the differences could be eliminated by reordering some sections in the files and using a uniform style.

For example, mew-unix.el has sections in this order:

;;;
;;; Text/Html, Application/Xml
;;;

;;;
;;; Image
;;;

;;;
;;; MIME setting
;;;

By contrast, mew-darwin.el has sections in this order:

;;
;; MIME setting
;;

;;;
;;; Text/Html, Application/Xml, Image
;;;

Using the same order and number of sections would make the similarities and differences among the files more apparent, and it might enable refactoring some common code out of all three files.

It does not maintain backward-compatibility: https://lists.gnupg.org/pipermail/gnupg-announce/2017q3/000412.html

We should put the following to .emacs:

(setq mew-prog-gpg2 "gpg")

And put the following to .gnupg/gpg.conf:

no-auto-key-retrieve
auto-key-locate local

When openning some mails, it reports "wrong number of auguments error ..." and focus to the Subject window. I am using the newest Mew on freebsd 9.1.

Regards!

こんにちは。
MELPAでMewをダウンロードできるようですが、MELPAでMewだけをダウンロードしてもバイナリがないためMewを起動できません。

MewがMELPAに入った経緯を調べたのですが、purcellさんがささっと入れたものがそのまま残っているようです。
DanielG/ghc-mod#125

infoを見ましたが、MELPAに関する言及はないようです。MELPAユーザーはどのようにバイナリをインストールすれば良いでしょうか。。？

in my country, gmail is blocked by thg gfw in china. i use shadowsocks to over the wall, but the is issue is: how to setup proxy in emacs mew?

I am running mew 6.7 on Emacs 24.5.1 under Ubuntu Linux 15.10.

My .mew-alias file occasionally gets lost: I find that email address expansion using [TAB] does not work as expected, and then I see that the .mew-alias file has gotten very small and contains only few
entries. If I notice it early enough, .mew-alias.old still contains the full information, but often I don't. Then I need to restore .mew-alias from a backup.

Unfortunately, I cannot pin down the precise conditions when this happens. I guess it is related to system or X.org crashes, but I just tried killing an active mew session doing "killall -9 emacs", and in this case .mew-alias did not get lost. So, it must be something more complicated, perhaps a race condition.

From my automatic backups, I see that the problem started exactly at the time of the upgrade to mew 6.7 (not 6.6) and persisted since then.

Further, my automatic backups indicate that, when .mew-alias gets lost, the file .mew-alias gets "deleted" (which probably means it does not simply get overwritten or it does not get re-created fast enough for my backup program not to notice).

Hope this helps to narrow down the problem. Let me know if you need me to do some more tests.

Hello,

I recently upgraded to emacs 28 (now 28.2, but the problem also appeared on emacs 28.1), and emacs raises a lot of "Invalid function" errors for macros such as mew-called-interactively-p or mew-summary-with-mewl (just to give 2 examples, I think there are more).

The errors go away if I remove byte-compiled .elc files from the mew installation directory.

I wonder if the problem is related to recently introduced native compilation? I've tried disabling native compilation by setting native-comp-deferred-compilation to nil, but that didn't seem to have an effect.

[mew 7ab0241]
[GNU Emacs 30.0.50 (build 1, x86_64-pc-linux-gnu,
GTK+ Version 3.24.34, cairo version 1.16.0) of 2023-02-01]

Please consider the attached e-mail, generated by Thunderbird 102.6.0. If I reply to that by pressing 'R', the subject fields gets doubled and looks like this:

Subject: Re: Tie not working as expected,Re: Tie not working as expected

It's very easy to miss that, not manually correcting it – I already got complaints.

It seems to me that Mew has difficulties with OpenPGP/MIME signed messages. Maybe the simplest solution is to accept a single 'Subject' field only (namely the first)?

problematic-email.txt

When the following keys are imported,

• 1024D/97AA33D6 Dima Barsky [email protected]
• 1024D/1A944AD7 Martin Albert [email protected]

I write a mail with To: [email protected], and encrypt it,
then it is encrypted with Dima's key instead of Martin's key.

Is it a known issue? Should the decrypters be around with "<"
and ">" in mew-pgp-encrypt, mew-old-pgp-encrypt and so on?
Comments?

When I bind a keybind to self-insert-command's key on global-map,
I encountered an error.

my environment:
mew-version: Mew version 6.5
OS: Ubuntu 12.04
Emacs-version: 24.3.1

setting is like this:

(define-key global-map
  "a" (lambda () (interactive)
        (call-interactively 'self-insert-command)))

I saw following error message when I typed "a" after typed "g" at mew-summary-mode.

mew-input-folder-self-insert: Wrong type argument: symbolp, #[nil "ÀÁ!�" [call-interactively self-insert-command] 2 nil nil]

It seems the problem have at mew-minibuf.el.
I found following code.(line number 363)

  ((not (string-match "self-insert-command" (symbol-name gfunc)))
   (mew-input-folder-display "not allowed"))

image-type-available-p is defined multiply in auxiliary.el.
This causes byte-compile warning.

Could you please push the tags, v6.5rc2 and later?

Recent enigmail versions encrypt the subject (if the corresponding option is set) and uses a (configurable) default string (Encrypted Message) for the normal subject header.
Two questions about this:

1. Will Mew support this nice feature for sending new mails?
2. If not, Mew should at least handle incoming messages more cleverly. The subject string for a reply should be constructed differently. Currently, the reply to such a enigmail message is: Re: here is your readable subject,Encrypted Message. The ,Encrypted Message part should be avoided.

System: MacOS High Sierra using the latest Mew from the git repo and Emacs 26.2.

Problem:

Intermittently, but consistently once it first happens (until a restart of Emacs), when creating a mail message, new or in reply etc, I am getting a read only error when attempting to edit the body of the resulting draft message. I can edit the header of the email, however, just not the body of the draft.
There seem to be two ways this problem can manifest:

1. (In a minimal Emacs config with only Mew loaded) After a long series of replies with perhaps the first 10 to 20 working fine.

Details for this case:

At least once, I found the problem first manifests when you try to create a draft, say in reply to an email or via 'w', and that goes fine. You C-c C-c to send it, and then the minibuffer returns 'nil' (despite prior sends succeeding) and the draft does not go away, i.e., you stay in its buffer. Subsequent check reveals that it wasn't sent. All subsequent attempts to edit the body portion of any message (header is still editable) for this draft or any other draft fails with a message in the minibuffer about the text being read-only.

With mew-debug set to t, here is a Messages buffer log excerpt from the aforementioned failing sequence of events. In this particular case, I can edit the draft body successfully, but I get this error when I attempt to send the message via C-c C-c and 'y':

   Connecting to the IMAP server...done
    Communicating with the IMAP server...
    Retrieving 1 message in background...
    1 message retrieved
    (New file)             <--------- Here is create a reply to the newly received message
    Saving file /Users/username/Mail/mew/draft/1...
    Wrote /Users/username/Mail/mew/draft/1
    Draft is prepared
    Really send this message? (y or n) y
    Saving file /Users/username/Mail/mew/draft/1...
    Wrote /Users/username/Mail/mew/draft/1
    Making a message...
    let*: Text is read-only   <--- This, apparently, is where the issue occurs. Not sure in Mew's code where this occurs. 

2. (In my usual maximal Emacs config with evil-mode, org-mode, etc. loaded) Right away in the first reply to an email most times.

Relevant output of Messages log:

Draft is prepared
funcall-interactively: Text is read-only [2 times]

I've bisected my config with respect to (2) trying to identify which package may be interacting to cause this issue and haven't found an obvious offender if there is one at all. Indeed, manifestion (1), above, has happened to me with just Mew loaded and nothing else.

How to reproduce the problem:

Not sure honestly. But, since I'd think we'd want Mew to work with lots of common packages like evil-mode and org-mode, load at least evil-mode and org-mode and then test Mew. If no one else was getting this issue, say, even after a long test reply thread (as I tested between a couple of my own email accounts), then I would have more info to try to debug.

Other possibly not relevant information:

Permissions when this happens look like (notice draft, inbox, postq, queue - all folders that Mew creates - username is a placeholder for my real username):

    /Users/username/Mail/mew:
    total used in directory 40K available 56550100
    drwxr-xr-x  8 username staff  256 Nov 17 16:22  #imap
    drwxr-xr-x 10 username staff  320 Nov 17 16:32 .
    drwxr-xr-x 27 username staff  864 Nov 18 10:43 ..
    drwxr-xr-x  2 username staff   64 Nov 17 12:47  attach
    drwxr-xr-x 79 username staff 2.5K Nov 18 10:31  backup
    drwx------ 12 username staff  384 Nov 18 10:43  draft
    drwx------  4 username staff  128 Nov 18 10:30  inbox
    drwx------  2 username staff   64 Oct 18 18:28  postq
    drwx------  8 username staff  256 Nov 18 10:31  queue
    drwxr-xr-x 79 username staff 2.5K Nov 18 10:26  trash

and within the draft folder, after several failed draft edits and/or sends things look like:

  /Users/username/Mail/mew/draft:
  total used in directory 128K available 47213328
  drwx------ 33 username staff 1.1K Nov 22 12:29 .
  drwxr-xr-x 10 username staff  320 Nov 17 16:32 ..
  -rw-r--r--  1 username staff 1.7K Nov 18 13:38  #1
  -rw-r--r--  1 username staff  679 Nov 18 13:54  #2
  -rw-r--r--  1 username staff  683 Nov 18 13:55  #3
  -rw-r--r--  1 username staff  428 Nov 21 21:19  #4
  -rw-r--r--  1 username staff  428 Nov 21 21:19  #5
  -rw-r--r--  1 username staff  378 Nov 21 21:20  #6
  -rw-r--r--  1 username staff  107 Nov 22 11:30  #7
  -rw-r--r--  1 username staff  342 Nov 22 12:29  #8
  -rw-------  1 username staff   15 Nov 22 12:29  .mew-mtime
  -rw-------  1 username staff   10 Nov 18 13:31  .mew-sinfo
  -rw-------  1 username staff   51 Nov 18 13:31  .mew-sinfo.old
  -rw-------  1 username staff    0 Nov 18 13:31  .mew-summary
  -rw-------  1 username staff 1.7K Nov 18 13:38  1
  -rw-------  1 username staff  170 Nov 18 13:38  1.mdi
  -rw-r--r--  1 username staff  284 Nov 18 08:44  1~
  -rw-------  1 username staff  679 Nov 18 13:54  2
  -rw-------  1 username staff  172 Nov 18 13:54  2.mdi
  -rw-r--r--  1 username staff  385 Nov 18 10:42  2~
  -rw-------  1 username staff  683 Nov 18 13:55  3
  -rw-------  1 username staff  172 Nov 18 13:55  3.mdi
  -rw-------  1 username staff  428 Nov 21 21:19  4
  -rw-------  1 username staff   69 Nov 21 21:19  4.mdi
  -rw-------  1 username staff  428 Nov 21 21:19  5
  -rw-------  1 username staff   69 Nov 21 21:19  5.mdi
  -rw-------  1 username staff  378 Nov 21 21:20  6
  -rw-------  1 username staff   69 Nov 21 21:20  6.mdi
  -rw-------  1 username staff  107 Nov 22 11:30  7
  -rw-------  1 username staff   69 Nov 22 11:30  7.mdi
  -rw-------  1 username staff  342 Nov 22 12:29  8
  -rw-------  1 username staff   69 Nov 22 12:29  8.mdi
  -rw-------  1 username staff  107 Nov 22 11:32  8~

Update & workaround for now:

For now, I launch another Emacs that loads a minimalist config with just Mew (no evil-mode, no org-mode, etc.):

#!/bin/bash
~/.nix-profile/Applications/Emacs.app/Contents/MacOS/Emacs -q -l slim-init.el

As per (1), this still does not resolve the issue permanently, but does allow me to use Mew mostly without issues. The problem is intermittent with either my regular config or the minimal config (takes longer too manifest in the minimal config as opposed to immediately in my usual config) and therefore difficult to debug. If the problem manifests, I can usually restart Emacs and then Mew will behave properly for some unknown amount of time.

I also have Gnus and notmuch working fine, but I really like Mew!

Unfortunately, the function mew-biff-setup in mew-net.el assumes global-mode-string is always a lisp. But there is an easy fix: check to see if it is string, and if it is, set global-mode-string to be a new list that contains the old string value of global-mode-string .

Here's a patch.
mew.patch

@tats I'm going to release Mew v6.9. To view files nicely on github, many files were moved into sub-directories. I believe that all documents are updated. It would be appreciated if you did final check.

TODO:

• Add XOAUTH2 support, merge from https://github.com/yoshinari-nomura/Mew/tree/mew-support-xoauth2
• Use json-read-from-string instead of json-parse-string for Emacs 26
• Disable XOAUTH2 by default, especially to use IMAP PLAIN
• Support mew-config-alist to use XOAUTH2 for each servers
• Update documentation for XOAUTH2 in info/mew.texi

cf. https://groups.google.com/g/mew-ja/c/Hn9L27ll-eY

I would like to change the master branch to main.

Hi,

I have a problem with creating encrypted mails and specifically S/MIME. Over time, many S/MIME certificates have been collected by Mew, many of which have expired. If I now want to encrypt a mail with S/MIME, it often happens that this is not possible because there are several certificates in my key ring that match the To: address. The error message then also reads: "Ambigious specification for <[email protected]>". However, the currently valid certificate is available, it should only be possible to select it. Is it possible to include such a selection? The same applies to signing. There, too, I cannot select a certificate if there are several expired personal certificates.

Bernhard