To fully complete the ghost-on-removal-when-shared-thread-exists
proposal, we need to clear all ghost messages when the last active
message is removed from a thread.
Amended by db: Remove the last test of T530, as it no longer makes sense
if we are garbage collecting ghost messages.
implement ghost-on-removal, the solution to T590-thread-breakage.sh
that just adds a ghost message after removing each message.
It leaks information about whether we've ever seen a given message id,
but it's a fairly simple implementation.
Note that _resolve_message_id_to_thread_id already introduces new
message_ids to the database, so i think just searching for a given
message ID may introduce the same metadata leakage.
test thread breakage when messages are removed and re-added
This test (T590-thread-breakage.sh) has known-broken subtests.
If you have a two-message thread where message "B" is in-reply-to "A",
notmuch rightly sees this as a single thread.
But if you:
* remove "A" from the message store
* run "notmuch new"
* add "A" back into the message store
* re-run "notmuch new"
Then notmuch sees the messages as distinct threads.
This happens because if you insert "B" initially (before anything is
known about "A"), then a "ghost message" gets added to the database in
reference to "A" that is in the same thread, which "A" takes over when
it appears.
But if "A" is subsequently removed, no ghost message is retained, so
when "A" appears, it is treated as a new thread.
I see a few options to fix this:
ghost-on-removal
----------------
We could unilaterally add a ghost upon message removal. This has a
few disadvantages: the message index would leak information about what
messages the user has ever been exposed to, and we also create a
perpetually-growing dataset -- the ghosts can never be removed.
We could add a ghost upon message removal iff there are other
non-ghost messages with the same thread ID.
We'd also need to remove all ghost messages that share a thread when
the last non-ghost message in that thread is removed.
This still has a bit of information leakage, though: the message index
would reveal that i've seen a newer message in a thread, even if i had
deleted it from my message store
track-dependencies
------------------
rather than a simple "ghost-message" we could store all the (A,B)
message-reference pairs internally, showing which messages A reference
which other messages B.
Then removal of message X would require deleting all message-reference
pairs (X,B), and only deleting a ghost message if no (A,X) reference
pair exists.
This requires modifying the database by adding a new and fairly weird
table that would need to be indexed by both columns. I don't know
whether xapian has nice ways to do that.
scan-dependencies
-----------------
Without modifying the database, we could do something less efficient.
Upon removal of message X, we could scan the headers of all non-ghost
messages that share a thread with X. If any of those messages refers
to X, we would add a ghost message. If none of them do, then we would
just drop X entirely from the table.
---------------------
One risk of attempted fixes to this problem is that we could fail to
remove the search term indexes entirely. This test contains
additional subtests to guard against that.
This test also ensures that the right number of ghost messages exist
in each situation; this will help us ensure we don't accumulate ghosts
indefinitely or leak too much information about what messages we've
seen or not seen, while still making it easy to reassemble threads
when messages come in out-of-order.
Jani Nikula [Sun, 10 Apr 2016 19:43:22 +0000 (22:43 +0300)]
lib: fix handling of one character long directory names at top level
The code to skip multiple slashes in _notmuch_database_split_path()
skips back one character too much. This is compensated by a +1 in the
length parameter to the strndup() call. Mostly this works fine, but if
the path is to a file under a top level directory with one character
long name, the directory part is mistaken to be part of the file name
(slash == path in code). The returned directory name will be the empty
string and the basename will be the full path, breaking the indexing
logic in notmuch new.
Fix the multiple slash skipping to keep the slash variable pointing at
the last slash, and adjust strndup() accordingly.
David Bremner [Sat, 9 Apr 2016 01:49:50 +0000 (22:49 -0300)]
test: cope with glass backend file naming variations
In several places in the test suite we intentionally corrupt the Xapian
database in order to test error handling. This corruption is specific to
the on-disk organization of the database, and that changed with the
glass backend. We use the previously computed default backend to make
the tests adapt to changing names.
David Bremner [Sat, 9 Apr 2016 01:49:49 +0000 (22:49 -0300)]
configure: add test for default xapian backend
This is mainly for the test suite. We already expect the tests to be
run in the same environment as configure was run, at least to get the
name of the python interpreter. So we are not really imposing a new
restriction.
David Bremner [Sat, 9 Apr 2016 01:49:48 +0000 (22:49 -0300)]
test/atomicity: guard chert-only optimization
This should potentially be updated to have an equivalent optimization
for the glass backend, but it in my unscientific tests, the glass backend
without the optimization is faster then the chert backend with.
Please put my address in CC when replying. Thanks!
From 4b9ab261a0ea8a31065e310c5150f522be86d37b Mon Sep 17 00:00:00 2001
From: stefan <aeuii@posteo.de>
Date: Fri, 8 Apr 2016 22:47:06 +0200
Subject: [PATCH] emacs: make use of `message-make-from'
David Edmondson [Sun, 3 Apr 2016 15:38:22 +0000 (16:38 +0100)]
emacs: Always insert crypto buttons.
When no decryption or signature examination is
happening (i.e. `notmuch-crypto-process-mime' is `nil') insert buttons
that indicate this, rather than remaining silent.
David Edmondson [Sun, 3 Apr 2016 16:04:38 +0000 (17:04 +0100)]
emacs: Allow part preferences to depend on message content.
Currently the preference for which sub-part of a multipart/alternative
part is shown is global. Allow to the user to override the settings on a
per-message basis by providing the ability to call a function that has
access to the message to return the discouraged type list.
David Bremner [Wed, 30 Mar 2016 02:03:02 +0000 (23:03 -0300)]
cli/new: add better error messages for FILE_ERROR
The code in add_file seems to assume that NOTMUCH_STATUS_FILE_ERROR is
never returned from add_message. This turns out to be false (although it
seems to only happen in certain fairly rare race conditions).
David Bremner [Sun, 27 Mar 2016 20:25:11 +0000 (17:25 -0300)]
nmbug: ignore # comments
Lines starting with # have always (for a long time, anyway) been ignored
by notmuch-restore, but have not been generated by notmuch-dump
previously. In order to make nmbug robust against such output, ignore
comment lines.
David Edmondson [Tue, 8 Mar 2016 17:12:59 +0000 (17:12 +0000)]
emacs: Improve the acquisition of text parts.
`notmuch-get-bodypart-text' assumed that it is always possible to
acquire text/* parts via the sexp output format. This is not true if the
part in question has a content type of application/octet-stream but is
being interpreted as text/* based on the extension of the part filename.
Rework `notmuch-get-bodypart-text' to use the raw output format to
address this and make the implementation common with that of
`notmuch-get-bodypart-binary'.
in d27d90875dfb1 (2016-02-20) notmuch-mua-reply-insert-header-p-function
was set to notmuch-show-reply-insert-header-p-never as its default was
changed to something else. Now that default is set back to *-never so
this change done in d27d90875dfb1 is not needed anymore.
W. Trevor King [Thu, 24 Mar 2016 17:08:18 +0000 (10:08 -0700)]
NEWS: Consolidate nmbug-status and notmuch-report sections
These are the same tool; the nmbug-status text just landed before the
name change. We can also drop the message-url details from NEWS,
since they're already in the man page.
W. Trevor King [Sat, 2 Jan 2016 06:08:04 +0000 (22:08 -0800)]
notmuch-report: Add notmuch-report(1) and notmuch-report.json(5) man pages
To describe the script and config file format, so folks don't have to
dig through NEWS or the script's source to get that information.
The Makefile and conf.py are excerpted from the main doc/ directory
with minor simplifications and adjustments. The devel/nmbug/ scripts
are largely independent of notmuch, and separating the docs here
allows packagers to easily build the docs and install the scripts in a
separate package, without complicating notmuch's core build/install
process.
W. Trevor King [Sat, 2 Jan 2016 06:08:03 +0000 (22:08 -0800)]
notmuch-report.json: Rename from status-config.json
status-config.json wasn't obviously associated with the old
nmubg-status, now notmuch-report. The new name is
${CONFIGURED_SCRIPT}.json, so the association should be clear.
W. Trevor King [Sat, 2 Jan 2016 06:08:02 +0000 (22:08 -0800)]
notmuch-report: Rename from nmbug-status
This script generates reports based on notmuch queries, and doesn't
really have anything to do with nmbug, except for sharing the NMBGIT
environment variable.
Michal Sojka [Tue, 29 Dec 2015 21:55:40 +0000 (22:55 +0100)]
doc/reply: Clarify how reply-to header is handled
Current documentation and comments in the code do not correspond to
the actual code and tests in the test suite ("Un-munging Reply-To" in
T230-reply-to-sender.sh). Fix it.
David Bremner [Fri, 29 Jan 2016 11:53:11 +0000 (07:53 -0400)]
test: replace single giant term with large phrase
We only need a long string, not a single long term to trigger batch
mode. The giant term triggers a bug/incompatibility in Xapian 1.3.4
that throws an exception because it is longer than the Xapian term size
limit.
Tomi Ollila [Tue, 1 Mar 2016 19:30:07 +0000 (21:30 +0200)]
lib: NOTMUCH_DEPRECATED macro also for older compilers
Some compilers (older than gcc 4.5 and clang 2.9) do support
__attribute__ ((deprecated)) but not
__attribute__ ((deprecated("message"))).
Check if clang version is at least 3.0, or gcc version
is at least 4.5 to define NOTMUCH_DEPRECATED as the
latter variant above. Otherwise define NOTMUCH_DEPRECATED
as the former variant above.
For a bit simpler implementation clang 2.9 is not included
to use the newer variant. It is just one release, and the
older one works fine. Clang 3.0 was released around 2011-11
and gcc 5.1 2015-04-22 (therefore newer macro for gcc 4.5+)
Jani Nikula [Sun, 13 Mar 2016 12:27:17 +0000 (14:27 +0200)]
emacs: hello: promote '?' as the universal help key
Move the brief help text at the bottom of the hello screen to the
notmuch-hello-mode help, and promote '?' as the universal help key
across Notmuch. This unclutters the hello screen, and allows for a
more verbose description in the mode help. Hopefully, this change is
useful for both experienced and new users alike.
While at it, improve the links to Notmuch and hello screen
customization.
David Bremner [Sat, 13 Feb 2016 17:11:18 +0000 (13:11 -0400)]
STYLE: document some rules about variable declarations
No-one seemed opposed to C99 style loop variable declarations. The
requirement to declare variables at the top of blocks is maybe a little
more contested, but I believe it reflects the status quo.
David Edmondson [Sat, 20 Feb 2016 18:24:38 +0000 (18:24 +0000)]
emacs/mua: Generate improved cited text for replies
Use the message display code to generate message text to cite in
replies.
For now we set insert-headers-p function to
notmuch-show-reply-insert-header-p-never so that, as before, we don't
insert part buttons.
With that choice of insert-headers-p function there is only one
failing test: this test has a text part (an email message) listed as
application/octet-stream. Notmuch show displays this part, but the
reply code omitted it as it had type application/octet-stream. The new
code correctly includes it. Thus update the expected output to match.
David Edmondson [Sat, 20 Feb 2016 18:24:36 +0000 (18:24 +0000)]
emacs/show: Make the insertion of part headers overridable.
This allows callers of notmuch-show-insert-bodypart to use a `let'
binding to override the default function for specifying when part
headers should be inserted.
We also add an option to never show part buttons which will be used by
the test suites for the reply tests.
David Edmondson [Thu, 11 Feb 2016 08:33:27 +0000 (08:33 +0000)]
emacs: Report a lack of matches when calling `notmuch-show'.
If the basic query passed to `notmuch-show' generates no results, ring
the bell and inform the user that no messages matched the query rather
than displaying an empty buffer and showing an obscure error.
Similarly when refreshing a `notmuch-show' buffer and no messages match.
Mark Walters [Sat, 13 Jun 2015 08:08:10 +0000 (09:08 +0100)]
emacs: Bind filter in search to 'l'
Change the key binding for filter (or "limit") in search-mode. This
gives consistency with the new filter in show-mode, and frees 'f' for
forward-thread in the future.
Tomi Ollila [Fri, 22 Jan 2016 22:01:53 +0000 (00:01 +0200)]
configure: pass HAVE_TIMEGM to build
Checking the existence of timegm() function and setting
configure internal variable ${have_timegm} was done, but
actually defining HAVE_TIMEGM in build was not done --
meaning that compat timegm() was always part of final
notmuch binaries.
notmuch.Database.remove_message should raise exception on failure
The notmuch python bindings document that database.remove_message
should raise an exception when the message removal fails, but they
don't actually do it.
David Bremner [Sun, 16 Aug 2015 17:41:16 +0000 (19:41 +0200)]
debian: add gpgsm as build dependency
It's not needed for the actual build, but it is needed to run the
SMIME tests; <!nocheck> means it can be omitted if the tests are not
going to be run.
David Bremner [Sun, 16 Aug 2015 17:41:13 +0000 (19:41 +0200)]
test: add broken S/MIME signature verification test for notmuch CLI
The test is pretty much cut and paste from the PGP/MIME version, with
obvious updates taken from notmuch output. This also requires setting
up gpgsm infrastucture.
Tomi Ollila [Wed, 6 Jan 2016 19:04:03 +0000 (21:04 +0200)]
devel/emacs: add devel/try-emacs-mua
devel/try-emacs-mua provides an easy way to try and experiment
with the notmuch emacs client distributed in emacs subdirectory of
the notmuch source tree.
try-emacs-mua starts a new emacs process and if initial checks pass
*scratch* buffer is filled with information of how to begin.
Normal emacs command line arguments can be used, like -q or -Q.
These arguments are appended verbatim to the starting emacs process.
If the emacs version in use is smaller than 24.4, special care is taken
to ensure that notmuch*.elc files older than corresponding .el files
are not loaded. Since emacs 24.4, setting `load-prefer-newer' variable
takes care of this.
W. Trevor King [Wed, 30 Dec 2015 21:04:13 +0000 (13:04 -0800)]
nmbug-status: Fix unbalanced <p> tags in default header/footer
These were broken by b70386a4 (Move the generated date from the top of
the page to the footer, 2014-05-31), which moved 'Generated ...' to
the footer with the opening tag, but didn't replace the blurb opening
tag or add a closing tag after 'Generated ...'.
W. Trevor King [Thu, 7 Jan 2016 13:51:29 +0000 (05:51 -0800)]
nmbug-status: Style headers with smaller fonts
We only use h1 through h3, and David prefers smaller headers [1], so
shift over to the font sizes usually used for h2 through h4 [2,3,4].
I haven't bothered with the W3C's default margins, since a bit of
extra whitespace doesn't seem like a big deal.
Michal Sojka [Wed, 6 Jan 2016 21:28:45 +0000 (22:28 +0100)]
emacs: Don't use nconc on quoted list
As pointed out by David Bremner, Elisp manual says "A common pitfall
is to use a quoted constant list as a non-last argument to ‘nconc’."
Since this was the case in recently added code, we fix it here.
Michal Sojka [Sat, 2 Jan 2016 16:47:45 +0000 (17:47 +0100)]
emacs: Handle switch-function argument of notmuch-mua-mail
notmuch-mua-mail ignored the switch-function argument and always used
the function returned by notmuch-mua-get-switch-function instead. In
order to support standard emacs interfaces (compose-mail in this
case), this commit changes notmuch-mua-mail to use the switch-function
argument if it is non-nil and notmuch-mua-get-switch-function
otherwise.
Michal Sojka [Sat, 2 Jan 2016 16:47:43 +0000 (17:47 +0100)]
emacs: Fix mail composition under Emacs 23
Commit 570c0aeb40bd0c3af8174624a55e968f62c44f09 reworked
notmuch-mua-mail function in a way that worked only under Emacs 24.
The reason was that message-setup-1 took one argument less in Emacs
23.
We fix this by only supplying the return-action argument when it is
actually set by the caller.
W. Trevor King [Sat, 2 Jan 2016 06:08:01 +0000 (22:08 -0800)]
nmbug-status: Avoid hard-coded filename in error message
We already have a 'filename' variable with the name, so stay DRY and
use that variable here.
Also fix a missing-whitespace error from bed8b674 (nmbug-status:
Clarify errors for illegible configs, 2014-05-10), wrapping on the
sentence to match similar error-generation earlier in this function.
Recent addition of notmuch-message-mode introduced several problems:
1. When message-setup-hook is used to set buffer local variables,
these settings are not effective, because all buffer local
variables are immediately erased by notmuch-message-mode
initialization.
2. message-mode-hook gets invoked twice - first when message-mail
invokes message-mode and second when notmuch-mua-mail invokes
notmuch-message-mode.
This commit fixes these problems by replacing a call to message-mail
with notmuch-specific code that is (hopefully) equivalent to
message-mail functionality before introduction of
notmuch-message-mode.
We first initialize notmuch-message-mode with
notmuch-mua-pop-to-buffer, which is a modified version of
message-pop-to-buffer and then call message-setup-1, which is the only
functionality of message-mail that is needed for notmuch.
Michal Sojka [Tue, 29 Dec 2015 08:52:17 +0000 (09:52 +0100)]
test: Unset ALTERNATE_EDITOR before running emacsclient
ALTERNATE_EDITOR causes emacsclient to run an alternate editor if the
emacs server is not ready. This can collide with intended
functionality in test-lib.sh.
If the ALTERNATE_EDITOR is set but empty, emacsclient runs emacs
daemon and tries to connect to it. When this happens the emacs run by
test-lib.sh fails to start the server and the subsequent attempts to
use the server fail because the daemon started by emacsclient does not
know about notmuch-test-progn. This leads to test suite failure due to
time out on any emacs test.
Michal Sojka [Tue, 29 Dec 2015 09:30:26 +0000 (10:30 +0100)]
test: Always use paths without symlinks
When notmuch sources are at a symlinked path, some tests fail because
one part of the test uses physical path and another uses logical
path (with symlinks). For example the following test output is
produced when the test is started from /home/src/symlink-to-notmuch,
which is a symlink to /home/src/notmuch.
FAIL notmuch-fcc-dirs set to a string
--- T310-emacs.26.OUTPUT 2015-12-29 08:54:29.055878637 +0000
+++ T310-emacs.26.EXPECTED 2015-12-29 08:54:29.055878637 +0000
@@ -1,5 +1,5 @@
From: Notmuch Test Suite <test_suite@notmuchmail.org>
To:
Subject:
-Fcc: /home/src/notmuch/test/tmp.T310-emacs/mail/sent-string
+Fcc: /home/src/symlink-to-notmuch/test/tmp.T310-emacs/mail/sent-string
--text follows this line--
nil
This commit makes all paths in test scripts physical. With it, all
tests pass even when run from a symlinked directory.
Mark Walters [Wed, 28 Oct 2015 18:28:39 +0000 (18:28 +0000)]
emacs: poll: return useful errors when poll fails.
Previously poll called from emacs would fail silently. This makes it
return a useful error message.
In the non-deprecated case of notmuch new and appropriate hooks, it
uses notmuch-call-notmuch-process which gives an error and
additionally puts the stdout/stderr etc in the *Notmuch errors*
buffer.
In the deprecated case of a custom poll script it only returns an
error message.
Commit based on a bug report, and a potential fix, by Ketil Malde.
Mark Walters [Thu, 12 Nov 2015 11:02:19 +0000 (11:02 +0000)]
emacs: hello: fix accidental modification of widget-keymap
In emacs24 we use make-composed-keymap. It seems that if only a single
map is specified then emacs just resuses it rather than creating a
copy of it. Thus use make-sparse-keymap to force a copy.
Steven Allen [Sun, 25 Oct 2015 21:30:39 +0000 (17:30 -0400)]
forbid atomic transactions on writable, upgradable databases
We can't (but currently do) allow upgrades within transactions because
upgrades need their own transactions. We don't want to re-use the
current transaction because bailing out of an upgrade would mean loosing
all previous changes (because our "atomic" transactions don't commit
before hand). This gives us two options:
1. Fail at the beginning of upgrade (tell the user to end the
transaction, upgrade, and start over).
2. Don't allow the user to start the transaction.
I went with the latter because:
1. There is no reason to call `begin_atomic` unless you intend to to
write to the database and anyone intending to write to the database
should upgrade it first.
2. This means that nothing inside an atomic transaction can ever fail
with NOTMUCH_STATUS_UPGRADE_REQUIRED.
Tomi Ollila [Wed, 2 Sep 2015 14:34:39 +0000 (17:34 +0300)]
emacs: notmuch-show-view-raw-message clears buffer, makes it read-only
notmuch-show-view-raw-message() re-uses buffer created with same
name (same Message-Id:) but it did not erase it before filling.
If this ever happened, there were duplicated (potentially overlapping)
content in the buffer. Now this is fixed.
Apparently since emacs 24.5 the (view-buffer) makes the buffer read-only;
so this problem would not have happened there, just that
notmuch-show-view-raw-message() failed. This is fixed by setting
inhibit-read-only t before erasing and filling the buffer. The emacs 24.5
feature having raw message buffer read-only is also now explicitly set to
the buffer so the same experience is available with emaces < 24.5.