Most people who write lots of lisp tend to only sparsely use empty
"separator" lines within forms. In lisp they feel unnecessary and
since most files stick to this convention we get a bit confused
when there are extra empty lines. It feels like the s-expressions
are falling into pieces.
All of this is especially true between a function's doc-string and
body because the doc-string is colored differently, which visually
already separates it quite sufficiently from the code that follows.
In notmuch-show buffer insert invisible U+200E LEFT-TO-RIGHT MARK
character at the beginning of message header paragraph if the From
header contains a right-to-left character. This ensures that the
header paragraph is always rendered in left-to-right mode.
See Emacs Lisp reference manual section "(elisp) Bidirectional
Display" for more info.
Fixed emacs docstrings to be consistent. No functional change.
- removed some (accidental) indentation
- removed some trailing newlines
- added trailing periods where missing (some exclusions)
Before this change, `load-prefer-newer' was ignored.
Set NOERROR and MUST-SUFFIX arguments of `load' to t, and NOSUFFIX
argument to nil, to preserve the behaviour of the deleted `let' form.
Starting with Emacs 27 the old `cl' implementation is finally
considered obsolete. Previously its use was strongly discouraged
at run-time but one was still allowed to use it at compile-time.
For the most part the transition is very simple and boils down to
adding the "cl-" prefix to some symbols. A few replacements do not
follow that simple pattern; e.g. `first' is replaced with `car',
even though the alias `cl-first' exists, because the latter is not
idiomatic emacs-lisp.
In a few cases we start using `pcase-let' or `pcase-lambda' instead
of renaming e.g. `first' to `car'. That way we can remind the reader
of the meaning of the various parts of the data that is being
deconstructed.
An obsolete `lexical-let' and a `lexical-let*' are replaced with their
regular variants `let' and `let*' even though we do not at the same
time enable `lexical-binding' for that file. That is the right thing
to do because it does not actually make a difference in those cases
whether lexical bindings are used or not, and because this should be
enabled in a separate commit.
We need to explicitly depend on the `cl-lib' package because Emacs
24.1 and 24.2 lack that library. When using these releases we end
up using the backport from GNU Elpa.
We need to explicitly require the `pcase' library because
`pcase-dolist' was not autoloaded until Emacs 25.1.
This is the non-deprecated way to use completing-read. Additionally
the old use was broken when using ivy for completing-read. For user's
using completing-read-default they won't see the default URL now, but
if they hit enter it will be visited. Alternatively they can select
it with M-n.
From the completing-read documentation for initial-input:
This feature is deprecated--it is best to pass nil for INITIAL-INPUT
and supply the default value DEF instead. The user can yank the
default value into the minibuffer easily using M-n.
Additionally collection is now all urls, rather than all but the
first. I'm not sure why "(cdr urls)" was previously done.
This is like notmuch-search-filter-by-tag, but creates a new search
rather than filtering the current search. We add this to
notmuch-common-keymap since this can be used by many contexts. We bind
to the key "t", which is the same key used by
notmuch-search-filter-by-tag in notmuch-search-mode-map. This is done
intentionally since the keybinding for notmuch-search-mode-map can be
seen as a specialization of creating a new search.
This change was motivated for use in "notmuch-hello". It is a more
convenient way to search a tag than expanding the list of all tags. I
also noticed many saved searches people use are simply tags.
Add x and X binds to notmuch-tree for functionally that we have in
notmuch-show.
The notmuch-tree-quit binding is somewhat redundant, since it is
handled by notmuch-bury-or-kill-this-buffer which is bound to q.
Signed-off-by: William Casarin <jb55@jb55.com>
This functions removes some duplicate logic between
notmuch-tree-{next,prev}-matching-message
We do this because we will be adding some additional logic similar to
the notmuch-show-next-open-message function, and it will help if this
logic is all in one place.
Signed-off-by: William Casarin <jb55@jb55.com>
This function captures some common logic when jumping to matching
messages in notmuch-tree mode.
We also add a new return value (t or nil), that indicates if there was
a next matching message in the thread to show.
Signed-off-by: William Casarin <jb55@jb55.com>
Without this change, we see the following warning when compiling the
elisp:
```
EMACS emacs/notmuch-crypto.elc
In end of data:
emacs/notmuch-crypto.el:266:1:Warning: the function
‘notmuch-show-get-message-id’ is not known to be defined.
```
Thanks to Örjan Ekeberg and David Edmondson for their followup about
this.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
Saved searches in notmuch-hello and notmuch-jump can specify whether
to use search mode or tree mode. This adds an option for them to
specify unthreaded mode.
We have shortcuts S and Z to let the user switch to Search view and
Tree view with the current search. Add U to let the user switch to
unthreaded view from the current search, and ensure that S and Z
switch from unthreaded to search and tree veiew respectively.
Tree mode allows the user to choose whether to use the split screen
displaying just the current message or a full screen displaying the
entire thread. As unthreaded mode is quite different in use the user
may want a different customisation for this mode.
It is likely that the user will want a different line format for
unthreaded mode from tree mode; in particular the thread structure
graphics are unnecessary in unthreaded mode.
Add a new customisable variable and set it to something sensible.
This commit introduces a new 'unthreaded' search mode where each
matching message is shown on a separate line. It shares almost all of
its code with tree view. Subsequent commits will allow it to diverge
slightly in appearance.
It causes this function to fail with:
let: Wrong type argument: null, t
Support for this was removed from Emacs in April
2019 (5c5e309527e6b582e2c04b83e7af45f3144863ac) because it never
worked correctly (apparently).
This also shouldn't be necessary as sentinels will not be called
unless emacs is idle or waiting for input. Therefore, the
`process-put' calls immediately following the `make-process' call
should always complete before the sentinel is first called.
This commit changes the behaviour of notmuch-mua-attachment-check
so that it stops searching for notmuch-mua-attachment-regexp when a
new mime-part is reached. This avoids false warnings when matching
words occur inside forwarded messages.
Rather than blocking emacs while gpg does its' thing, by default run
key retrieval asynchronously, possibly updating the display of the
message on successful completion.
This is an unbound function that is quite useful. It opens a selected
thread in notmuch-tree from the current search query.
Signed-off-by: William Casarin <jb55@jb55.com>
Debian's lintian has an informational alert
desktop-entry-lacks-keywords-entry, which recommends including
Keywords= in a .desktop file.
I dug around a bit in /usr/share/applications/*.desktop to make sure
that we covered the range of keywords other e-mail applications are
using. If anyone has other suggestions for keywords, they can add
them to this list.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
The extra flexibility of having both HAVE_EMACS (for yes, there is an
emacs we can use) and WITH_EMACS (the user wants emacs support) lead
to confusion and bugs. We now just force WITH_EMACS to 0 if no
suitable emacs is detected.
When we have not been able to evaluate the signature status of a given
MIME part, showing a content-free (and interaction-free) "[ Unknown
signature status ]" button doesn't really help the user at all, and
takes up valuable screen real-estate.
A visual reminder that a given message is *not* signed isn't helpful
unless it is always present, in which case we'd want to see "[ Unknown
signature status ]" buttons on all messages, even ones that don't have
a signing structure, but i don't think we want that.
Amended by db to drop the unused initialization of 'label'
`notmuch-search-interactive-region' was moved to notmuch-lib.el in
f3cba19f88 and renamed to
`notmuch-interactive-region' without making the old function
obsolete, thereby breaking user-commands which made use of it.
This commit marks the function as obsolete and makes it an alias for
the new function.
Apparently, message-default-charset is deprecated, which causes the
following warning messages during the build:
In notmuch-maildir-setup-message-for-saving:
emacs/notmuch-maildir-fcc.el:172:31:Warning: ‘message-default-charset’ is an
obsolete variable (as of 26.1); The default charset comes from the
language environment
In discussion with emacs upstream over on
https://debbugs.gnu.org/35370, it appears that we can just drop this
entirely and things should still work with emacs 25.
Since only the first line of the documentation is shown by the
help command, it is confusing when "x" and "a" seem to have the same
binding in show-mode. This commit makes the two function documentations
first lines different and (hopefully) clearer.
Instead of relying on the "In-Reply-To" header, use a buffer-local variable,
notmuch-message-queued-tag-changes, to add and remove tags to affected
messages when the message-send-hook is triggered.
Include the message-id of forwarded messages in the new message.
This ensures that the new (forwarding) message is linked to the
same thread as the message being forwarded.
Fix notmuch-describe-key crashing for the following two cases
1. format-kbd-macro cannot deal with keys like [(32 . 126)], switch to
use key-description instead.
2. if a function in the current keymap is not bounded, it will crash
the whole process. We check if it is bounded and silently skip it to
avoid crashing.
Query the user if the message text indicates that an attachment is
expected but no MML referencing an attachment is found.
This is not enabled by default - see the documentation for
`notmuch-mua-attachment-check'.
Add a new binding when looking at messages, B, that prompts with a
list of URLs found in the current message, if any. Open the one that
is selected in a browser.
amended by db: s/--browse-urls/-browse-urls/
When invoking gpg as a backgrounded tool, it's important to let gpg
know that it is backgrounded, to avoid spurious prompts or other
breakage.
In particular, https://bugs.debian.org/913614 was a regression in
GnuPG which causes problems when importing keys without a terminal,
but gpg expects one.
Ensuring that notmuch-emacs always invokes gpg as a background process
should avoid some of these unnecessary failure.
Thanks to Justus Winter for finding this problem.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
These are intended to included in the sphinx manual for notmuch-emacs.
The stamp file makes it easier to depend on the docstrings from other
parts of the build
This small library is intended to support batch extraction of Emacs
Lisp docstrings from source files. Clients will need to include (or
replace) rstdoc.rsti.
When filtering by tags in notmuch-search-filter-by-tag, only return tags
related to the current query.
Before, it was returning all tags. There's no reason to refine the
current query with tags that don't exist in the current result set.
Signed-off-by: William Casarin <jb55@jb55.com>
`mm-inline-text-html-with-images' was removed from mm-decode.el in
2016 and replaced with `mm-html-inhibit-images'.
`gnus-select-frame-set-input-focus' was removed from gnus-util.el in
2016 and existed only for XEmacs compatibility.
On some platforms (e.g. macOS), it is necessary to add a real sentinel
process for the error buffer used by `notmuch-start-notmuch' rather
than a no-op sentinel.
Correct URLs that have crept into the notmuch codebase with http://
when https:// is possible.
As part of this conversion, this changeset also indicates the current
preferred upstream URLs for both gmime and sup. the new URLs are
https-enabled, the old ones are not.
This also fixes T310-emacs.sh, thanks to Bremner for catching it.
This brings the --decrypt argument to "notmuch reply" into line with
the other --decrypt arguments (in "show", "new", "insert", and
"reindex"). This patch is really just about bringing consistency to
the user interface.
We also use the recommended form in the emacs MUA when replying, and
update test T350 to match.
We also expand tab completion for it, update the emacs bindings, and
update T350, T357, and T450 to match.
Make use of the bool-to-keyword backward-compatibility feature.
Dynamically bind enriched-decode-display-prop when inserting
text/enriched part. This complements commit 9b05823838 for
emacs versions before 24.4 which do not have advice-add
functionality.
Since emacs 25.3 this particular bug is fixed.
Commit a7964c86d1 ("emacs: Sanitize authors and subjects in search
and show") added sanitization of header information for display. Do
the same for reply subjects.
This fixes the long-standing annoying artefact of certain versions of
mailman using tab as folding whitespace, leading to tabs in reply
subjects.
Quoting from the elisp reference:
For other types (e.g., lists, vectors, strings), two arguments
with the same contents or elements are not necessarily ‘eq’ to
each other.
Thanks to "Attic Hermit" for the fix.
Switch to a local version of enriched-decode-display-prop if we
encounter a text/enriched part. This is to mitigate
https://bugs.gnu.org/28350. Normally it would be prudent to remove the
override afterwards, but in this case just leave it in.
Notes from db:
This doesn't disable text/enriched, just one feature of it.
notmuch-tree did not protect against concurrent refreshes like
notmuch-search, meaning, hitting '=' (notmuch-refresh-this-buffer)
quickly will spawn multiple parallel notmuch processes, and clobber
the existing results in the current buffer.
* notmuch-tree.el: Add a guard to notmuch-tree-refresh-view similar to
the one in notmuch-search.
make-process is a new function introduced in Emacs 25, which provides
greater control over process creation. Crucially, it allows
separately redirecting stderr directly to a buffer, which allows us to
avoid needing to use the shell to redirect to a temporary file in
order to correctly distinguish stdout and stderr.
* notmuch-lib.el: Use make-process when it is available; fall back to
the previous method when not.
Load subprocess error output to a string in the callers, and propagate
the error messages as a string parameter instead of a path to file
names.
Required to be able to avoid using temporary files for subprocess
error output.
* notmuch-lib.el: Update notmuch-check-async-exit-status,
notmuch-check-exit-status: accept an err parameter instead of
err-file; shift the responsibility of loading error messages from
files up the call stack.
The queries "" and "*" are special cased in the notmuch library to
match all messages, but only if they're the entire query. They can't
be combined with other queries, such as "* AND foo", in which case
they "leak" down to the Xapian query parser.
Notmuch show and tree buffers inadvertently combine the thread query
with said special queries, causing incorrect collapsing of
messages. Handle the special queries specially. We already do a
similar thing in notmuch-search-filter.
A leading / in paths in a .gitignore file matches the beginning of the
path, meaning that for patterns without slashes, git will match files
only in the current directory as opposed to in any subdirectory.
Prefix relevant paths with / in .gitignore files, to prevent
accidentally ignoring files in subdirectories and possibly slightly
improve the performance of "git status".
We have a steady trickle of people using notmuch-emacs from melpa with
distro packages of notmuch, and then being confused when it doesn't
work. Try to warn people what a foot-gun this is; this commentary
should be copied to the melpa web site.
There are some cases like remote usage where this might cause
problems, but those users can easily customize the variable. The
inconvenience seems to be outweighed by the security benefit for most
users.
Since the error field is unused by the emacs front end, no changes are
needed other than bumping the format version number.
As it is, this is a bit overengineered, but it will reduce duplication
when we support gmime 3.0
Emacs' major modes can facilitate navigation in their buffers by
supporting Imenu. In such major modes, launching Imenu (M-x imenu)
makes Emacs display a list of items (e.g., function definitions in a
code buffer). Selecting an item from this list moves point to this
item.
This patch adds Imenu support to both notmuch-show and notmuch-search
buffers:
* in notmuch-show, Imenu will present a list of all messages in the
currently visible thread;
* in notmuch-search, Imenu will present a list of all messages in the
search buffer.
Previously notmuch tree did not specify the format-version when
calling notmuch. This meant that when the structured output was
slightly changed (in commit 14c60cf168)
stash filename broke. This fixes this breakage by specifying the
format-version.
Adjusted notmuch-crypto gpg call-process function to respect
the GPG program set by the EasyPG epg-gpg-program variable.
This is to correct a problem observed on NixOS where only gpg2 is
installed by default. The Notmuch Emacs frontend fails when trying to
access someone's key to verify their signature when it cannot find the
GPG binary.
This backports the fix from emacs master (commit
77bbca8c82f6e553c42abbfafca28f55fc995d00) to notmuch-emacs to wrap
long headers.
This fixes the test introduced in the previous changeset.
With the mailto: handling in notmuch-emacs-mua, we can update the
desktop file to advertize we can be set as the default application to
handle email. While at it, add GenericName and Comment to be more
informative.
With --hello, notmuch-emacs-mua will run (notmuch) if mailto: url is
not given.
Gnus seems to display application/zip and application/tar by
default. This doesn't seem desirable so we override it.
We only override if the user has not customized
mm-inline-override-types themselves.
Fix the deprecation warning:
In notmuch-maildir-fcc-make-uniq-maildir-id:
emacs/notmuch-maildir-fcc.el:279:53:Warning: ‘system-name’ is an obsolete
variable (as of 25.1); use (system-name) instead
I've used (system-name) since at least 2011, so it must have been
around quite a while.
We now have several customizable options for address completion. There
is a customize group notmuch-address but it only contains one of these
options. Add all the others, and make it part of the notmuch customize
group.
This allows the user to save the address hash so that it is much
faster for the first completion after a restart. This defaults to off
as there are privacy implications to saving this information.
The code tries hard to avoid overwriting the wrong file. It also notes
if changes have been made to any of the relevant user settings, so
that the user does not get surprising results (i.e., outdated options
being used). Finally it stores some version information so that is
easy for us to update the format of the save file.
This makes the code access notmuch-address-full-harvest-finished via a
helper function, notmuch-address--harvest-ready. Later we will use
this to check whether we can load the harvest instead of regenerating
it.
There is really no need to have a separate install target for the
desktop file. Just install the desktop file with emacs, with a
configure option to opt out.
If the --hello parameter is given, display the notmuch hello buffer
instead of the message composition buffer if no message composition
parameters are given.
Signed-off-by: Jani Nikula <jani@nikula.org>
Some of the recent changes to the emacs code have used functions
introduced in emacs 24. The functions used are read-char-choice and
setq-local. This changeset adds a file notmuch-compat.el which
contains compatibility functions so that it should work on emacs
23.
Note, since these functions are taken almost unchanged from the emacs
source they are copyright the Free Software Foundation, and the header
in the file reflects that.
Provide functionality to resume editing a message previously saved with
notmuch-draft-save, including decoding the X-Notmuch-Emacs-Secure
header.
Resume gets the raw file from notmuch and using the emacs function
mime-to-mml reconstructs the message (including attachments).
'e' is bound to resume a draft from show or tree mode.
This provides initial support for postponing in the emacs frontend;
resuming will follow in a later commit. On saving/postponing it uses
notmuch insert to put the message in the notmuch database
Current bindings are C-x C-s to save a draft, C-c C-p to postpone a
draft (save and exit compose buffer).
Previous drafts get tagged deleted on subsequent saves, or on the
message being sent.
Each draft gets its own message-id, and we use the namespace
draft-.... for draft message ids (so, at least for most people, drafts
are easily distinguisable).
We want to use "e" for editting postponed messages in show, and in
tree view, so remove the binding for the function which does
(In message pane) Activate BUTTON or button at point
This hook can be used to update the message based on the results of
address completion. For example using message-templ or gnus-alias to set
the From address based on the To address just completed.
The post-completion command is added to the notmuch-company backend to
ensure that the hook is also called company completion is started
without going through notmuch-address-expand-name. See the docstring of
`company-backends' for more information.
This makes replying to a message in tree view, use the decrypted state
from the message pane if it is open. Previously it just used the
global decryption state from notmuch-crypto-process-mime.
In particular if notmuch-crypto-process-mime is nil, and the user
views the messages (which doesn't decrypt), toggles decryption in the
message pane, and then replies, the reply will be decrypted.
This makes $ in the tree pane toggle decryption in the message
pane. Without this the user can only decrypt the message pane by
switching to it, or by setting decryption on globally by setting
notmuch-crypto-process-mime to t.
The notmuch-tag-flagged, notmuch-search-flagged-face and
notmuch-crypto-part-header faces defaulted to "blue", which is nearly
unreadable when a dark background is in use. This is addressed by using
"LightBlue1" for dark backgrounds.
As a side effect, these faces are now no-op definitions for grayscale or
mono displays.
The normal tag commands in search mode tag the all threads meeting the
region when called interactively. This makes them do the same when
called non-interactively. This is a change in the api.
This makes tag changes appear in the message pane as well as in the
tree window.
Note that the message pane is reloaded each time a message is viewed
so the tags shown in the message pane can still be different from
those in the tree window. Usually this will just be that the tag
change is still shown as a change (strikethough underline etc) in the
tree window, and are shown after the change in the message
pane. However, if something else updates the database then the tags
shown can be genuinely different.
The command notmuch-interesting-buffer has got out of date -- it
doesn't mention notmuch-tree, and it still refers to message-mode not
notmuch-message-mode. Update both of these.
This fixes the bug that notmuch-cycle-notmuch-buffers does not include
notmuch-tree or notmuch-message-mode buffers in its cycling.
notmuch-show--build-buffer now queries a list of queries built by the
former. This simplifies the logic. It also provides an easy place to
experiment with alternate sets of queries for given notmuch-show-*
variables (e.g. users can use advice-add to do so in a surgical way).
This commit expands docstrings for notmuch-fcc-dirs and
notmuch-maildir-fcc-with-notmuch-insert to describe how quoted strings
are processed and make the ability to configure sent folders containing
whitespace more discoverable.
In commit 2a7b11b064 the default value
for notmuch-search-line-faces was changed so that it didn't match the
specification in the corresponding defcustom. This meant that it was
difficult for the user to customize this variable as they got a type
mismatch error.
Note anyone who had already customised this variable would not see
this bug as their customisation would match the defcustom.
We can't use c-u = as some of the notmuch-show refresh function
already uses that. This is a global keybinding for a relatively
infrequent function but M-= seems unlikely to cause conflicts.
notmuch-refresh-all-buffers calls each buffer's major mode specific
refresh function using the generic notmuch-refresh-this-buffer function.
Since the earlier changesets have stopped the refresh functions from
forcing the buffers to be redisplayed this can refresh buffers that
are not currently displayed without disturbing the user. This is very
useful for silent async background updating the emacs display when new
mail is fetched.
Signed-off-by: Ioan-Adrian Ratiu <adi@adirat.com>
This updates all windows displaying a notmuch-show buffer when the
buffer refresh function is called.
Each window displaying a notmuch-show buffer has its own currently
displayed message based on the (point) location. We store the state
of all displayed windows when refreshing a notmuch-show buffer and
re-apply the current shown message (point) for all windows.
Implementation note: Each window has it's own (point) value, besides
the buffer's (point) value. Sometimes these values are identical like
in the case where a single window displays a buffer. When multiple
windows display a buffer, (point) returns each window's specific value.
What we are storing in this changeset is the window values not the
buffer point values. The buffer's point is returned only if no window
is displaying the buffer, a case we do not care about here.
Signed-off-by: Ioan-Adrian Ratiu <adi@adirat.com>
There's no reason to completely kill a buffer while refreshing its
search results because the buffer name is constant between refreshes
(based on the search query), only its contents may change and notmuch
search kills all local variables, so it's safe to reuse.
Reusing the same buffer also makes it possible to do things like
refreshing a buffer which is not focused or even not shown in any
window - this will be used in the next commits to add auto-refresh
capabilities to all existing notmuch buffers + a function to call
after syncing mail to refresh everything.
Signed-off-by: Ioan-Adrian Ratiu <adi@adirat.com>
If no-display is non-nil when calling notmuch-search then do not force
the search buffer to be displayed.
Signed-off-by: Ioan-Adrian Ratiu <adi@adirat.com>
Make the notmuch-hello refresh function (notmuch-hello-update) not
force the buffer to be displayed. All the callers call it when the
buffer is already displayed so it will only affect non-interactive
callers. Since it is just a trivial wrapper of notmuch-hello anyone
who wants to force the buffer to be displayed should just call
notmuch-hello.
If an encrypted multipart message is received which contains html and
notmuch-multipart/alternative-discouraged is set to discourage "text/plain",
any encrypted parts are not decrypted during generation of the reply
text. This fixes that problem by making sure notmuch-mua-reply does
that.
notmuch jump allows the user to specify a key sequence rather than
just a single key for its bindings. However, it doesn't show what has
already been typed so it can be difficult to see what has
happened. This makes each key press appear, and the jump menu reduce
to the possible follow up keys.
We also bind backspace (emacs symbol DEL) to go back up a level in the
subjumpmaps, and to exit from the top level.
In commit 297d091e1c the key for
switching to the reverse keymap was changed from r to k but the
docstring for the defcustom was not updated. This changeset updates
the documentation to match.
This seems to fix a problem with emacs 25 creating partial buttons by
calling n-s-b-l with a region that does not include the whole button.
I'm not 100% sure it's legit to act outside the region passed by
jit-lock, but goto-address-fontify-region (where I borrowed the code
from) already does this, so this patch to not make things worse.
Emacs message-send seems to ignore a secure mml tag anywhere except at
the start of the body, and it must be followed by a newline. Since
this is almost certainly not desired we check for it, and require user
confirmation before sending.
As the setup before message-send or message-send-and-exit is getting
more complicated it is convenient to unify the two correspoinding
notmuch functions.
This commit adds a common message-send function for message-send and
message-send-and-exit. At the moment the overlap is small, but the
message-send function will get more complex.
The current code for address completion takes the list of possible
completions (whether generated internally or externally), makes the
first match the initial value for the completion, and puts all the
others (but not the first match) into the possible completions.
This has the nice effect that the <down> key takes you immediately to
the next completion (whereas if the first match were included in the
possible completions it would take you to the first match
again).
However, it has two side effects. First, once you have completed to
the full match you find it says and try completing again you get told
"no match" not "sole completion". Secondly, if you delete some of the
text and try completing you don't get the first match as an option.
This tries to get round most of these problems by including the full
list of possible completions, but with the first match moved to the
very end of the list.
