[ANN] notmuch-deliver

[Felipe Contreras]

On Mon, Nov 8, 2010 at 7:50 PM, Carl Worth  wrote:
> On Wed, 26 May 2010 17:01:34 +0300, Ali Polatel  wrote:
>> notmuch-deliver is a maildir delivery tool for notmuch mail indexer. It
>> reads from standard input, delivers the mail to the specified maildir
>> and adds it to the notmuch database. This is meant as a convenient
>> alternative to running notmuch new after mail delivery.
>
> Thanks for sharing this, Ali.
>
> What's the best way to advertise this to potential users?
>
> Should we include a separate utils directory in the notmuch repository
> with auxiliary programs like this?
>
> Or should we implement this functionality within the notmuch binary
> itself?
>
> I'm open to suggestions.
>
> If nothing else, the notmuchmail.org web page should grow a section to
> point to auxiliary programs like this that users might find helpful.
>
> Thanks again,

I think this should be part of notmuch itself, and there should be a
configuration to use this as Fcc, instead of relying on the mail
composer. This way both emacs and vim interfaces would share the same
configuration regarding the Fcc/Bcc preference.

-- 
Felipe Contreras

[PATCH] Properly quote Emacs' notmuch-command.

[Thomas Schwinge]

From: Thomas Schwinge 

It happens that my notmuch-command may look like this:

(setq notmuch-command "~/Mentor Graphics/command/notmuch")

Most of the times, notmuch-command is passed to call-process which will simply
do the right thing, but there are a few other cases where it is fed through the
shell and thus needs to be properly quoted / escaped.

I'm not entirely sure that shell-quote-wildcard-pattern is the correct function
to be used.  We can't use shell-quote-argument (which I first meant to use), as
that one will also escape the tilde (~).  Next idea was to first resolve the
tilde, and then use shell-quote-argument.  But I didn't find a suitable Emacs
function: we can't (unconditionally) use expand-file-name, as that one will
absolutize the filename: it will (wrongly) turn the default value `notmuch'
into `$CWD/notmuch'.

Signed-off-by: Thomas Schwinge 

---

Hallo!

As stated above, if someone has (more Emacs knowledge and) a better
solution, please shout.

An additional item to conider: this solves the issue with tilde
expansion, but what about other shell magic like filename expansion
(globbing; which is what shell-quote-wildcard-pattern explicitly does
*not* quote / escape)?


Once we have determined what we actually want, is it OK to send a
testsuite patch to make it use such a ``nonstandard'' path, for catching
regressions early in the future?


Gre,
 Thomas

 emacs/notmuch-lib.el  |4 ++--
 emacs/notmuch-show.el |4 ++--
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/emacs/notmuch-lib.el b/emacs/notmuch-lib.el
index dd180ee..fdeb32b 100644
--- a/emacs/notmuch-lib.el
+++ b/emacs/notmuch-lib.el
@@ -61,7 +61,7 @@ the user hasn't set this variable with the old or new value."
   (let ((long-string
 ;; Trim off the trailing newline.
 (substring (shell-command-to-string
-(concat notmuch-command " --version"))
+(concat (shell-quote-wildcard-pattern notmuch-command) " 
--version"))
0 -1)))
 (if (string-match "^notmuch\\( version\\)? \\(.*\\)$"
  long-string)
@@ -72,7 +72,7 @@ the user hasn't set this variable with the old or new value."
   "Return a value from the notmuch configuration."
   ;; Trim off the trailing newline
   (substring (shell-command-to-string
- (concat notmuch-command " config get " item))
+ (concat (shell-quote-wildcard-pattern notmuch-command) " config 
get " item))
  0 -1))

 (defun notmuch-database-path ()
diff --git a/emacs/notmuch-show.el b/emacs/notmuch-show.el
index 3a60d43..820f90f 100644
--- a/emacs/notmuch-show.el
+++ b/emacs/notmuch-show.el
@@ -957,12 +957,12 @@ than only the current message."
   (let (shell-command)
 (if entire-thread
(setq shell-command 
- (concat notmuch-command " show --format=mbox "
+ (concat (shell-quote-wildcard-pattern notmuch-command) " show 
--format=mbox "
  (shell-quote-argument
   (mapconcat 'identity 
(notmuch-show-get-message-ids-for-open-messages) " OR "))
  " | " command))
   (setq shell-command
-   (concat notmuch-command " show --format=raw "
+   (concat (shell-quote-wildcard-pattern notmuch-command) " show 
--format=raw "
(shell-quote-argument (notmuch-show-get-message-id)) " | " 
command)))
 (start-process-shell-command "notmuch-pipe-command" "*notmuch-pipe*" 
shell-command)))

-- 
tg: (b3caef1..) t/emcs-quote_notmuch-command (depends on: master)

[ANN] notmuch-deliver

[Thomas Schwinge]

Hallo!

On Mon, 08 Nov 2010 09:50:46 -0800, Carl Worth  wrote:
> On Wed, 26 May 2010 17:01:34 +0300, Ali Polatel  wrote:
> > notmuch-deliver is a maildir delivery tool for notmuch mail indexer. It
> > reads from standard input, delivers the mail to the specified maildir
> > and adds it to the notmuch database. This is meant as a convenient
> > alternative to running notmuch new after mail delivery.
> 
> Thanks for sharing this, Ali.
> 
> What's the best way to advertise this to potential users?

I recently put a description and link onto the notmuch web pages.

> Should we include a separate utils directory in the notmuch repository
> with auxiliary programs like this?

I wouldn't do so.  But that is not a very strong opinion of mine.

In general, I like it if I see a repository containing one specific tool,
and that one cleanly interfaces through specified interfaces with another
tool.  These two things are no longer as cleanly visible once
notmuch-deliver was part of the notmuch repository.


> Or should we implement this functionality within the notmuch binary
> itself?

That's another option, of course.  (And a separate discussion.)


> I'm open to suggestions.
> 
> If nothing else, the notmuchmail.org web page should grow a section to
> point to auxiliary programs like this that users might find helpful.

I'm working on that (and other parts of the web pages) as I go on with
exploring the ``notmuch world''.

I'll also take the liberty to put stuff from the mailing list or IRC
discussions into web pages, for we have to document this notmuch beast
;-), and it's better to have a generic place to refer people to, instead
of discussing the same things more than once.


If someone disagrees with any of this, I'm open to discuss these items.


Gr??e,
 Thomas
-- next part --
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: not available
URL: 
<http://notmuchmail.org/pipermail/notmuch/attachments/20110111/fc445790/attachment.pgp>

Re: [ANN] notmuch-deliver

[Felipe Contreras]

On Mon, Nov 8, 2010 at 7:50 PM, Carl Worth cwo...@cworth.org wrote:
 On Wed, 26 May 2010 17:01:34 +0300, Ali Polatel a...@exherbo.org wrote:
 notmuch-deliver is a maildir delivery tool for notmuch mail indexer. It
 reads from standard input, delivers the mail to the specified maildir
 and adds it to the notmuch database. This is meant as a convenient
 alternative to running notmuch new after mail delivery.

 Thanks for sharing this, Ali.

 What's the best way to advertise this to potential users?

 Should we include a separate utils directory in the notmuch repository
 with auxiliary programs like this?

 Or should we implement this functionality within the notmuch binary
 itself?

 I'm open to suggestions.

 If nothing else, the notmuchmail.org web page should grow a section to
 point to auxiliary programs like this that users might find helpful.

 Thanks again,

I think this should be part of notmuch itself, and there should be a
configuration to use this as Fcc, instead of relying on the mail
composer. This way both emacs and vim interfaces would share the same
configuration regarding the Fcc/Bcc preference.

-- 
Felipe Contreras
___
notmuch mailing list
notmuch@notmuchmail.org
http://notmuchmail.org/mailman/listinfo/notmuch

[PATCH] Have to configure and build inside the source directory.

[Thomas Schwinge]

From: Thomas Schwinge tho...@schwinge.name

Signed-off-by: Thomas Schwinge tho...@schwinge.name

---
 configure |   11 +++
 1 files changed, 11 insertions(+), 0 deletions(-)

diff --git a/configure b/configure
index c58dd0f..b5e446c 100755
--- a/configure
+++ b/configure
@@ -188,6 +188,17 @@ developers. Then, please email those details to the 
Notmuch list
 (notmuch@notmuchmail.org) so that we can hopefully make future
 versions of notmuch easier for you to use.
 
+EOF
+
+if ! { :  configure; } 2 /dev/null; then
+cat EOF
+*** Error: You have to configure and build in the source directory.
+
+EOF
+exit 1
+fi
+
+cat EOF
 We'll now investigate your system to verify that all required
 dependencies are available:
 
-- 
tg: (b3caef1..) srcdir (depends on: master)
___
notmuch mailing list
notmuch@notmuchmail.org
http://notmuchmail.org/mailman/listinfo/notmuch

better timeout patch.

[David Bremner]

Here is a timeout patch that actually seems to work. Also attached are
the diffs to debian/changelog to update to 0.5+nmu2.

From f8e8ea0058ca2e57fececc3f01cc05ece0c3c3c8 Mon Sep 17 00:00:00 2001
From: David Bremner brem...@unb.ca
Date: Sun, 19 Dec 2010 21:48:59 -0400
Subject: [PATCH 1/2] tests: Add optional use of timeout utility, if present. Each top level
test (basic, corpus, etc...) is run with a fixed timeout of 10m.

The goal here is to treat a hung test as a failure. The emacs test
for sending mail is known to be problematic on the debian
autobuilders. This is both a bandaid fix for that, and a sensible
long term feature.
---
 test/notmuch-test |   10 +-
 1 files changed, 9 insertions(+), 1 deletions(-)

diff --git a/test/notmuch-test b/test/notmuch-test
index b51045a..b75bc40 100755
--- a/test/notmuch-test
+++ b/test/notmuch-test
@@ -21,9 +21,17 @@ TESTS=basic new search json thread-naming raw reply dump-restore uuencode threa
 # Clean up any results from a previous run
 rm -r test-results /dev/null 2/dev/null
 
+# test for timeout utility
+if command -v timeout /dev/null; then
+TEST_TIMEOUT_CMD=timeout 10m 
+echo INFO: using 10 minute timeout for tests
+else
+TEST_TIMEOUT_CMD=
+fi
+
 # Run the tests
 for test in $TESTS; do
-	./$test $@
+$TEST_TIMEOUT_CMD ./$test $@
 done
 
 # Report results
-- 
1.7.2.3

From 6998bffbaf2b4d683b8ce0b1706b3636d60a6559 Mon Sep 17 00:00:00 2001
From: David Bremner brem...@unb.ca
Date: Tue, 14 Dec 2010 22:24:38 -0400
Subject: [PATCH 2/2] Add NMU stanzas to changelog

---
 debian/changelog |   14 ++
 1 files changed, 14 insertions(+), 0 deletions(-)

diff --git a/debian/changelog b/debian/changelog
index cc758b9..790f34b 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,3 +1,17 @@
+notmuch (0.5+nmu2) experimental; urgency=low
+
+  * Non-maintainer upload.
+  * Second try at timeout for test. Put timeouts at top level.
+
+ -- David Bremner brem...@debian.org  Sun, 19 Dec 2010 21:40:08 -0400
+
+notmuch (0.5+nmu1) experimental; urgency=low
+
+  * Non-maintainer upload.
+  * Add a timeout to emacs tests to hopefully work around build failures.
+
+ -- David Bremner brem...@debian.org  Tue, 14 Dec 2010 22:23:51 -0400
+
 notmuch (0.5) unstable; urgency=low
 
   * new: maildir-flag synchronization
-- 
1.7.2.3

___
notmuch mailing list
notmuch@notmuchmail.org
http://notmuchmail.org/mailman/listinfo/notmuch

Tag changes not persisted in message view

[Günter Ladwig]

Hi,

I installed notmuch yesterday and imported an account with about 9000 
mails. The import worked fine and the Emacs client shows all messages.


However, I ran into an problem with changing tags in the message view: I 
noticed that mails were not marked as read (i.e., their unread tag was 
not removed) after reading them in the message view. After some 
experimentation I can say that adding or removing tags in the message 
view does not work at all (neither manually nor implicitly through 
reading or archiving). While the UI is updated (i.e., the tags shown in 
the headers), the changes are not persisted. Neither the search results 
view nor the command line client show the changes made in the message 
view. After reopening a message, the tags will again appear as before.


Changing tags from the search results view or the command line works 
fine, even though it is slow (I think this is caused by that Xapian bug 
I saw some references to). Adding and removing tags in the message view 
is however instantaneous (although ineffective), which seems to indicate 
that Xapian was not called (or incorrectly).


I'm using the latest version from the git repository (although I tried 
0.5 as well, without any difference). I'm on Mac OS with Cocoa Emacs 
23.2.1 (tried the console version as well). I also tried a pristine 
Emacs configuration to remove interference with other Emacs packages, 
but no such luck. Here's output from configure:


Checking for Xapian development files... Yes (1.0.15).
Checking for GMime development files... Yes (gmime-2.4).
Checking for Glib development files (= 2.14)... Yes.
Checking for talloc development files... Yes.
Checking for valgrind development files... No (but that's fine).
Checking if emacs is available... Yes.
Checking which platform we are on... Mac OS X.
Checking for getline... No (will use our own instead).
Checking for strcasestr... Yes.
Checking for rpath support... No (nothing to worry about).
Checking for -Wl,--as-needed... No (nothing to worry about).
Checking for available C++ compiler warning flags...
-Wall -Wextra -Wwrite-strings -Wswitch-enum
Checking for available C compiler warning flags...
-Wall -Wextra -Wwrite-strings -Wswitch-enum -Wmissing-declarations

Any ideas?

Cheers,
Günter
___
notmuch mailing list
notmuch@notmuchmail.org
http://notmuchmail.org/mailman/listinfo/notmuch