The variable notmuch-pick-message-buffer should be buffer local but
instead notmuch-pick-message-buffer-name (a non-existent variable) was
made buffer local.
The function notmuch-pick-refresh-result (used to update tag changes)
was not quite correct: sometimes it got the choice between the subject
and " ..." wrong. This was always true but the new code often calls
this (when opening a message in the message pane to remove the unread
tag) while the async pick process is still running and this caused
mistakes which made the tests fail.
Thus we store the previous subject with the message.
This function was a horrible hack (sleeping while waiting for the
correct message). The new target code can just open the message in the
message window when it arrives.
Previously pick removed the unread tag from its tag display: since the
tag change is now customisable use the customised variable.
This only affected the tags displayed, not the tags in the database as
notmuch-show (in the view pane) did the actual tag changes.
Previously pick set a prefix argument prior to calling show in the
message pane to tell show to only show matching messages. This sets
notmuch-show-only-matching-messages instead which is much cleaner and
will work even if the user has configured show to default to showing
only matching messages.
This adds a target message for pick which it will jump to when (if) it
appears. It adds the target to notmuch-pick-from-show-current-query so
that pick goes straight to the message that was current in the show
view and it adds target to notmuch-pick-refresh-view so that the
current message is preserved.
Pick keeps point roughly at the top of the buffer while inserting
messages at the end as they come in (from the async
parser). Previously the save-excursion to do this was done once for
each thread inserted: now it is done for each individual message.
The advantage is that the message insertion code can decide where to
leave point. In the next patch point will be left on the target message.
Note notmuch-pick-insert-msg is unchanged as that is used by the tag
display update code.
Details:
- $pipe_decode is turned off, to prevent message-id from being
filtered out by "ignore" settings in the muttrc.
- Original values for $pipe_decode and $wait_key are saved and restored.
- The macros, being much longer now, are line wrapped for improved
readability.
On Thu, Feb 14, 2013 at 12:36:58AM +0100, Profpatsch wrote:
> On 13-02-13 02:35pm, Kevin J. McCarthy wrote:
> > A more likely idea is to check whether you have $pipe_decode set.
>
> BRILLIANT!
> So much for copying a basic rc from someone else.
> Of course, that was it and I’m officially an idiot.
Neat, thanks Kevin for debugging the issue down to $pipe_decode (which
I've never used, mutt never stops to amaze me :-)).
> And apparently Mail::Internet errors out if there is no Message-ID.
> (Which mentioned in the docs at CPAN…)
>
> Mystery solved.
Right, but still a more graceful failure model would be nice.
Please find attached a patch that in such cases should 1) give a
supposedly nice error message explaining what's going on and 2) empty
the results dir to avoid showing you unrelated results. It works for me.
But extra checking never hurts, in particular for the tag action, which
I don't personally use.
I guess it would also be nice to actually disable $pipe_decode in the
relevant Mutt macros, but I'm not sure about to do that without
interfering with user desired configuration. Kevin: do you know if there
is a common Mutt trick to store the value of a variable before changing
it, and restoring it a posteriori? More isolation for this kind of
things in Mutt would definitely be welcome...
Cheers.
--
Stefano Zacchiroli . . . . . . . zack@upsilon.cc . . . . o . . . o . o
Maître de conférences . . . . . http://upsilon.cc/zack . . . o . . . o o
Debian Project Leader . . . . . . @zack on identi.ca . . o o o . . . o .
« the first rule of tautology club is the first rule of tautology club »
From b67ab95855ce7d279d8c0b3ddcbc20e679afc70b Mon Sep 17 00:00:00 2001
From: Stefano Zacchiroli <zack@upsilon.cc>
Date: Thu, 14 Feb 2013 09:31:37 +0100
Subject: [PATCH] notmuch-mutt: more graceful handling of missing Message-Id
errors
in particular:
- the "thread" action would print an error and empty results dir
- the "tag action would print an error
As discussed in id:871udhcmks.fsf@zancas.localnet, notmuch-vim doesn't
really meet the standards of the CLI, emacs interface, or python
bindings in terms of being well maintained.
There seems to be consensus to use presence in contrib as
documentation of limited support by the notmuch developers; in fact
nmbug is pretty integrated into our current development process, so
devel seems more appropriate.
nmbug pull only merges upstream master, but nmbug push tries to push
all local branches. The asymmetry results in conflicts whenever there
have been changes in the config branch in the origin:
$ nmbug push
To nmbug@nmbug.tethera.net:nmbug-tags
! [rejected] config -> config (non-fast-forward)
error: failed to push some refs to 'nmbug@nmbug.tethera.net:nmbug-tags'
hint: Updates were rejected because a pushed branch tip is behind its remote
hint: counterpart. If you did not intend to push that branch, you may want to
hint: specify branches to push or set the 'push.default' configuration
hint: variable to 'current' or 'upstream' to push only the current branch.
'git push origin' exited with nonzero value
To fix this, only push the master branch on nmbug push. Any config
changes need to be done manually via git anyway.
We add a hook to the show buffer in the message window to close the
message window when that buffer quits. It checks that the
message-window is still displaying the show-message buffer and then
closes it.
Previously running search or pick from the pick buffer did not close
the message pane (if open). This meant that then new search ends up in
a very small window. Fix this so that the message pane is
shut. However, make it so that the pane is shut after the search
string is entered in case the user is basing the search on something
in the current message.
Currently pick just uses notmuch-show to display messages in the
message pane: this means that they get indented just as show
would. However, since pick is only displaying one message at a time
there is no need to indent so override the indentation.
Update pick's archive message to respect notmuch-archive-tags. Also
split archive message into an archiving part and a separate
"then-next" part, to move more inline with show. Update the keybinding
so default behaviour is unchanged.
Previously if you carried on past the last message in a pick view pick
would get confused and `forget' about the split pane and would try and
re-split when moving up again. This was due to faulty logic in
notmuch-pick-show-message: something that should have been in the (when message)
clause was not.
Thanks to jrollins for the bug report.
The test should be run using the wrapper run-tests.sh. This links
the tests into the normal notmuch TEST_DIRECTORY and runs them from
there. After the test is complete then the links are removed.
If the user pressed return on the end result status line it gave a
blank message. Modify the function notmuch-pick-get-message-id to
return nil rather than an empty message-id in this case to fix this.
This also fixes a bug in the (lack of) quoting of the id string.
Newer patch email containing In-Reply-To: to an email sent some time ago
(i.e. to a "thread") was not visible in that "thread" in patch view when
another patch "thread" was submitted in between. This change collects
all messages in every (notmuch-created) thread together before printing
all these threads out in a patch view.
Thanks to Ethan Glasser-Camp for initial review and suggestions with
code examples.
In latest configuration quite a few long views were added to the
Notmuch Patches page. To ease navigating to the views a 'Views'
section was added to the beginning of page containing hyperlink
to every view.
When the From: field in patch email does not contain 'realname'
field, the patch listing does not show anything as patch sender.
In this case use the part before '@' in mail address as the sender
identification in patch listing.
Add a --remove-dups flag which removes duplicate files from search and
thread results. Uses fdupes if installed. Otherwise it runs a size and
Digest::SHA scan on each file to detect duplicates.
Signed-off-by: Stefano Zacchiroli <zack@upsilon.cc>
The initial nmbug-status was pretty consistent in it's whitespacing
but a few lines had some leftover slips. Those are now "corrected".
Also, most of the code used ' as quoting char. As in Python one can
use ' and " interchangeably some code used " instead of '. However
the usage of those were inconsistent. Now all quotes that python
parses are ':s (only quoted content uses ":s).
No functional changes.
In reformatting the line 111 accidentally indented to one indentation
level too much (happens easily when interactively indenting python
code using emacs). The line now has 4 spacess less indentation, thus
restoring it to the block level it belongs.
This is (almost) the same script as has been used for
http://nmbug.tethera.net/status for a while now. The only change is
that the configuration is not hardcoded anymore. By default the config
is fetched from a special branch in the nmbug repo that contains only
config info. The idea is that push access to this branch can be
restricted a bit more than the tags, since it will change the
appearence of the web pages.
If any of the forked process exits with nonzero value, terminate
current operation -- nonzero exit value indicates failure and
then there is no point continuing.
When asked for interactive help, lookup notmuch-search-terms(7)
instead of notmuch(1). Syntax of notmuch queries used to be described
in the latter, but has recently been moved to the former.
Closes: #675073 (in the Debian BTS)
The main idea is consider the notmuch database as analogous to the
work-tree. A bare git repo is maintained in the users home directory,
with a tree of the form tags/$message-id/$tag
Like notmuch and git, we have a set of subcommnds, mainly modelled on
git.
Implementation wise, the heavy lifting is in the following functions.
commit xapian -> git
checkout git -> xapian
merge fetched git + git -> xapian
status find differences between xapian, git, and remote git.
The central implementation trick, from an idea I think due to
tomprince on IRC is manipulate the git index directly from the xapian
tag information. The merge routine is still done using a temporary
checkout as I wasn't able to get it working with the index only.
There are also some convenience wrappers around git commands, like "fetch"
that essential just set GIT_DIR in the environment.
In order to encode tags (viewed as octet sequences) into filenames,
we whitelist a smallish set of characters and %hex escape anything outside.
The prefix is omitted in git, which lets one save and restore to
different prefixes (although this is only lightly tested).
Thanks to Tomi Ollila for a huge amount of feedback and patches while
putting this together.
We won't use all of the included build infrastructure files, but adding them
nevertheless helps to track changes that are applied to them upstream.
Signed-off-by: Thomas Schwinge <thomas@schwinge.name>
The original "contrib" name is lousy. Everything in notmuch has been
contributed, and we are integrating as much of it as possible, (rather
than making users grub through contrib looking for useful pieces to
install).
Meanwhile, the only things we have in contrib are command-line
completion scripts, so "completion" makes more sense as a name, (and
helps make "./configure" slightly less ambiguous).
