Michal Sojka [Tue, 14 Feb 2012 17:09:47 +0000 (18:09 +0100)]
emacs: Fix display of highlighted line in notmuch-search
When notmuch-search-line-faces is used to set background color in search
results, the highlight of the current line is not always displayed
correctly. This patch fixes that by increasing the priority property of
the highlight overlay.
Justus Winter [Wed, 15 Feb 2012 21:25:13 +0000 (22:25 +0100)]
python: provide a Database.close function
Rename Database.__del__ to Database.close, move it just below the open
function and call close() in a newly created destructor just below the
constructor.
scandir() returns "strings allocated via malloc(3)" which are then
"collected in array namelist which is allocated via
malloc(3)". Currently we just free the array namelist. Instead, free
all the entries of namelist, and then free namelist.
entry only points to elements of namelist, so we don't free it
separately.
Dmitry Kurochkin [Mon, 13 Feb 2012 11:09:07 +0000 (15:09 +0400)]
emacs: allow to set RETAIN-STATE for `notmuch-show-refresh-view' interactively
The notmuch-show view refresh function (`notmuch-show-refresh-view',
bound to "=") accepts an optional RETAIN-STATE argument. The patch
allows to set this argument interactively by using "C-u =".
emacs: cleanup and simplify `notmuch-show-archive-thread' and related functions
Recent changes in notmuch-show tagging introduced some code
duplication. The patch cleanups and simplifies
`notmuch-show-archive-thread' function by using
`notmuch-show-tag-all', no longer used function are removed. After
the change, `notmuch-show-archive-thread' function becomes symmetric
with `notmuch-show-archive-message'.
A side effect of these changes is that `notmuch-show-archive-thread'
no longer calls "notmuch tag" for each message in the thread.
David Edmondson [Wed, 8 Feb 2012 08:02:14 +0000 (08:02 +0000)]
emacs: Allow the indentation of content to be toggled.
Very deeply indented content is sometimes difficult to
read (particular for something like patches). Allow the indentation of
the content to be toggled with '<'.
Indentation of the header lines is not affected, so it remains
possible to see the structure of the thread.
David Edmondson [Wed, 8 Feb 2012 08:02:12 +0000 (08:02 +0000)]
emacs: Allow `notmuch-show-mode' to display only matching messages.
The current behaviour (all messages shown, non-matching collapsed)
is retained as the default. Type '!' to switch to showing only
the matching messages - non-matching messages are not available.
'!' will switch back to showing everything.
David Edmondson [Wed, 8 Feb 2012 08:02:10 +0000 (08:02 +0000)]
emacs: Rework crypto switch toggle.
Re-work the existing crypto switch toggle to be based on a persistant
buffer-local variable.
To allow this, modify `notmuch-show-refresh-view' to erase and re-draw
in the current buffer rather than killing the current buffer and
creating a new one. (This will also allow more per-buffer behaviour in
future patches.)
Add a binding ('$') to toggle crypto processing of the current buffer
and remove the prefix argument approach that achieves a similar
result.
David Edmondson [Tue, 7 Feb 2012 17:26:11 +0000 (17:26 +0000)]
emacs: Ensure that gnupg output goes at the end of the buffer.
When showing the user some details of gnupg output, ensure that those
details are shown at the end of the gnupg status buffer
("*notmuch-crypto-gpg-out*"), otherwise it can end up mixed up with
earlier output.
Jani Nikula [Mon, 6 Feb 2012 19:57:22 +0000 (21:57 +0200)]
cli: convert "notmuch show" to use the new argument parser
Use the new notmuch argument parser to handle arguments in "notmuch
show". There are three minor functional changes:
1) Also set params.raw = TRUE when defaulting to raw format when part
is requested but format is not specified. This was a bug, and
--part=0 without --format=raw did not work previously.
2) Set params.decrypt = FALSE if crypto context creation fails.
3) Only use the parameters for the last --format if specified multiple
times. Previously this could have resulted in a non-working mixture
of parameters.
Austin Clements [Sat, 4 Feb 2012 21:24:26 +0000 (16:24 -0500)]
show: Simplify new text formatter code
This makes the text formatter take advantage of the new code
structure. The previously duplicated header logic is now unified,
several things that we used to compute repeatedly across different
callbacks are now computed once, and the code is simpler overall and
32% shorter.
Unifying the header logic causes this to format some dates slightly
differently, so the two affected test cases are updated.
Austin Clements [Sat, 4 Feb 2012 21:24:25 +0000 (16:24 -0500)]
show: Convert text format to the new self-recursive style
This is all code movement and a smidgen of glue. This moves the
existing text formatter code into one self-recursive function, but
doesn't change any of the logic. The next patch will actually take
advantage of what the new structure has to offer.
Note that this patch retains format_headers_message_part_text because
it is also used by the raw format.
Jani Nikula [Thu, 2 Feb 2012 14:58:41 +0000 (16:58 +0200)]
emacs: add default value to notmuch-search-line-faces
Add default value to notmuch-search-line-faces to show "unread"
messages in bold, and "flagged" messages in blue, to have some visual
indication of important messages in search results. This should be
helpful for new users.
"unread" tag is quite obvious, and handled specially both in the lib
and emacs ui. "flagged" is synced to maildir F flag in the lib. If one
syncs the maildir to IMAP, this also translates to corresponding IMAP
flag. (This is "starred" in GMail and Android.)
Jani Nikula [Tue, 31 Jan 2012 17:29:06 +0000 (19:29 +0200)]
emacs: make show view a/A/x/X key bindings more consistent
Modify the show view key bindings as follows to make them more
consistent:
'a' = Archive current message, then move to next message, or show next
thread from search if at the last message in thread.
'A' = Archive each message in thread, then show next thread from
search.
'x' = Archive current message, then move to next message, or exit back
to search results if at the last message in thread.
'X' = Archive each message in thread, then exit back to search
results.
The changes make the key bindings more consistent in two ways:
1) 'a'/'A' both advance to the next thread like 'a' used to.
2) 'x' operates on messages and 'X' on threads like 'a'/'A'.
emacs: s/tags/tag-changes/ for arguments of tagging functions
This makes the argument names more consistent and clear. The
following functions changed: `notmuch-tag',
`notmuch-search-tag-thread', `notmuch-search-tag-region' and
`notmuch-search-tag-all'.
Since `notmuch-tag' is a non-interactive function and hence is meant
to be invoked programmatically, it should accept zero tags. Also, the
tagging operations (bound to "*", "+", "-") would accept empty input
without an error.
emacs: relax tag syntax check in `notmuch-tag' function
The tag syntax check in `notmuch-tag' function was too strict and did
not allow nmbug tags with "::". Since the check is done for all
tagging operations in Emacs UI, this basically means that no nmbug
tags can be changed. The patch relaxes the tag syntax check to allow
any tag names that do not include whitespace characters.
emacs: separate history for operations which accept single and multiple tags
Some tag-related operations accept a single tag without prefix
(`notmuch-select-tag-with-completion'), others accept multiple tags
prefixed with '+' or '-' (`notmuch-read-tag-changes'). Before the
change, both functions used a single default minibuffer history. This
is inconvenient because you have to skip options with incompatible
format when going through the history. The patch adds separate
history lists for the two functions. Note that functions that accept
the same input format (e.g. "+", "-", "*") share the history list as
before.
The patch adds `notmuch-show-tag-all' function bound to "*" in
notmuch-show view. The function is similar to the
`notmuch-search-tag-all' function for the notmuch-search view: it
changes tags for all messages in the current thread.
test: fix emacs tests after tagging operations changes
After the recent tagging operations changes, functions bound to "+"
and "-" in notmuch-search and notmuch-show views always read input
from the minibuffer. Use kbd macros instead of calling them directly.
emacs: make "+" and "-" tagging operations in notmuch-show more flexible
Before the change, "+" and "-" tagging operations in notmuch-show view
accepted only a single tag. The patch makes them use the recently
added `notmuch-read-tag-changes' function, which allows to enter
multiple tags with "+" and "-" prefixes. So after the change, "+" and
"-" bindings in notmuch-show view allow to both add and remove
multiple tags. The only difference between "+" and "-" is the
minibuffer initial input ("+" and "-" respectively).
emacs: make "+" and "-" tagging operations in notmuch-search more flexible
Before the change, "+" and "-" tagging operations in notmuch-search
view accepted only a single tag. The patch makes them use the
recently added `notmuch-read-tag-changes' function (renamed
`notmuch-select-tags-with-completion'), which allows to enter multiple
tags with "+" and "-" prefixes. So after the change, "+" and "-"
bindings in notmuch-search view allow to both add and remove multiple
tags. The only difference between "+" and "-" is the minibuffer
initial input ("+" and "-" respectively).
emacs: move tag format validation to `notmuch-tag' function
Before the change, tag format validation was done in
`notmuch-search-operate-all' function only. The patch moves it down
to `notmuch-tag', so that all users of that function get input
validation.
David Bremner [Sat, 4 Feb 2012 18:40:24 +0000 (13:40 -0500)]
emacs: use mark instead of point-max in MML quoting.
As Aaron explains in id:"m2vco72tf3.fsf@wal122.wireless-pennnet.upenn.edu"
Using point-max would include the signature in the quoting as well.
It would probably be fairly odd to want to put an MML tag in one’s
signature, but that doesn’t mean that we should break that usage.
We had to use point-max in the 0.11.1 bug-fix release, because the
mark functionality was added post 0.11.
emacs: fix `notmuch-wash-region-to-button' to work at beginning of buffer
`Notmuch-wash-region-to-button' is the function that creates hidden
regions with buttons for signatures, citations and original messages.
Before the change, it did not work correctly if the to-be-hidden
region started at the beginning of a message: the visibility toggle
button was hidden as well. The patch fixes this. There are two parts
in the fix:
* Use `insert-before-markers' instead of `insert' for creating the
button, so that it does not get added to the hidden overlay.
* Stop using PREFIX argument for adding a newline before the button.
The newline should not be added before a button at the beginning of
buffer.
David Edmondson [Mon, 30 Jan 2012 16:52:20 +0000 (16:52 +0000)]
emacs: Move the blank line from the bottom of the headers to the top of the body.
The blank line doesn't really change position, but is now considered
to be part of the body rather than part of the headers. This means
that it is visible when the body is visible rather than when the
headers are visible.
Austin Clements [Sun, 29 Jan 2012 05:50:10 +0000 (00:50 -0500)]
lib: Use talloc to simplify cleanup in notmuch_database_open
Previously, we manually "free"d various pointers in
notmuch_database_open. Use a local talloc context instead to simplify
cleanup and eliminate various NULL pointer initializations and
conditionals.
Austin Clements [Sun, 29 Jan 2012 05:50:08 +0000 (00:50 -0500)]
lib: Don't delete uninitialized pointers
In the error-handling paths of notmuch_database_open, we call
notmuch_database_close, which "delete"s several objects referenced by
the notmuch_database_t object. However, some of these pointers may be
uninitialized, resulting in undefined behavior. Hence, allocate the
notmuch_database_t with talloc_zero to make sure these pointers are
NULL so that "delete"ing them is harmless.
Aaron Ecay [Fri, 3 Feb 2012 10:24:08 +0000 (11:24 +0100)]
emacs: quote MML tags in replies
Emacs message-mode uses certain text strings to indicate how to attach
files to outgoing mail. If these are present in the text of an email,
and a user is tricked into replying to the message, the user’s files
could be exposed.
Edited-by: Pieter Praet <pieter@praet.org>: Rebased to release branch.
emacs: modify the default show-mode key bindings for archiving
This changes the default key bindings for the 'a' key in notmuch-show
mode. Instead of archiving the entire thread, it now just archives
the current message, and then advance to the next open message
(archive-message-then-next). 'A' is now bound to the previous
archive-thread-then-next function.
emacs: add option to show-next-{, open-}message functions to pop out to parent buffer if at end
This will allow for keybindings that achieve a smoother message
processing flow by reducing the number of key presses needed for most
common operations.
This adds two new message archiving functions that parallel the thread
archiving functions: notmuch-show-archive-message{,-then-next}. The
former also takes a prefix argument to unarchive the message (ie. put
back in inbox).
emacs: break out thread navigation from notmuch-show-archive-thread
This function is now just for archiving the current thread. A new
function is created to archive-then-next. The 'a' key binding is
updated accordingly.
This will allow people to bind to the simple thread archiving function
without the extra navigation. The archive-thread function now also
takes a prefix to unarchive the current thread (ie. put the whole
thread back in the inbox).
emacs: break up notmuch-show-archive-thread-internal into two more generally useful functions
Break up notmuch-show-archive-thread-internal into two new functions:
notmuch-show-tag-thread-internal: applies a tag to all messages in
thread. If option remove flag is t, tags will be removed instead of
added.
notmuch-show-next-thread: moves to the next thread in the search
result. If given a prefix, will show the next result, otherwise will
just move to it in the search view.
Two new interactive functions, notmuch-show-{add,remove}-tag-thread,
are also added. Together, these provide a better suit of thread
tagging and navigation tools.
The higher level thread archiving functions are modified to use these
new function.
Tomi Ollila [Mon, 30 Jan 2012 10:31:25 +0000 (12:31 +0200)]
moved _config_(get|set)_list () functions earlier in the file
Moved static functions _config_get_list () and _config_set_list ()
closer to the beginning of file so that their definition is known
(without adding forward declarations) in upcoming changes.
Dmitry Kurochkin [Wed, 25 Jan 2012 01:10:53 +0000 (05:10 +0400)]
emacs: use a single history for all searches
There are two ways to do search in Emacs UI: search widget in
notmuch-hello buffer and `notmuch-search' function bound to "s".
Before the change, these search mechanisms used different history
lists. The patch makes notmuch-hello search use the same history list
as `notmuch-search' function.
Dmitry Kurochkin [Wed, 25 Jan 2012 01:10:52 +0000 (05:10 +0400)]
emacs: bind "s" to `notmuch-search' in notmuch-hello buffer
Before the change, "s" in notmuch-hello buffer would jump to the
search box. The patch changes the binding to `notmuch-search' which
is consistent with all other notmuch buffers.
Tomi Ollila [Tue, 24 Jan 2012 20:55:59 +0000 (22:55 +0200)]
uncrustify.cfg: label indent, some known types, not, # and ##
Adjusted some uncrustify variables to get closer to prevailing style:
* Label indent (for goto) relative to current indentation.
* Registered GMimeObject and mime_node_t being as types.
* Space after ! (not) operator.
* No space after 'stringify' (#) preprosessor token.
* No spacing change around ## (option not versatile enough).
There are at least 3 cases where attention needs to be paid:
* If there is newline between function name and open paren in function
call, the paren (and args) are indented too far right.
* #define HOUR (60 *MINUTE) -- i.e. no space after star (*).
* void (*foo)(args) -- i.e no space between (name) and (args).
David Edmondson [Tue, 24 Jan 2012 16:14:05 +0000 (16:14 +0000)]
test: Add `test_emacs_expect_t'.
Add a new test function to allow simpler testing of emacs
functionality.
`test_emacs_expect_t' takes one argument - a lisp expression to
evaluate. The test passes if the expression returns `t', otherwise it
fails and the output is reported to the tester.
David Edmondson [Tue, 24 Jan 2012 16:14:04 +0000 (16:14 +0000)]
test: Don't return the result of checking for running emacs to the tester.
When checking for a running emacs, test_emacs evaluates the empty list
'()'. This returns 'nil' when emacs is running, which is then
prepended to the actual test result. Given that it is not part of the
actual test output the test harness can incorrectly report test
failure (or success).
Austin Clements [Mon, 23 Jan 2012 23:33:10 +0000 (18:33 -0500)]
show: Introduce mime_node formatter callback
This callback is the gateway to the new mime_node_t-based formatters.
This maintains backwards compatibility so the formatters can be
transitioned one at a time. Once all formatters are converted, the
formatter structure can be reduced to only message_set_{start,sep,end}
and part, most of show_message can be deleted, and all of
show-message.c can be deleted.
Austin Clements [Mon, 23 Jan 2012 23:33:09 +0000 (18:33 -0500)]
mime node: Record depth-first part numbers
This makes the part numbers readily accessible to formatters.
Hierarchical part numbering would be a more natural and efficient fit
for MIME and may be the way to go in the future, but depth-first
numbering maintains compatibility with what we currently do.
Pieter Praet [Mon, 23 Jan 2012 04:22:33 +0000 (05:22 +0100)]
test: only exclude "deleted" messages from search if explicitly configured
Currently, the 'search.exclude_tags' option is automatically set to
"deleted;spam;" if it's missing from the config file.
This violates the Principle of Least Surprise, so update the tests to
*only* expect the exclusion of messages which are tagged "deleted" if the
'search.exclude_tags' option is explicitly set *and* contains that tag.