When no --output option were given, change default to display senders
only.
When only --count option is given, display senders (in contrary to not
displaying anything).
Document that --count with --sort=**order** is not supported option
combination.
Clean up some whitespace in the documentation.
One test updated to have --output=count without sender nor recipient
output option.
Changed one tab in doc/man1/notmuch-show.rst to 8 spaces.
Removed one leading tab in doc/man5/notmuch-hooks.rst to make
the SYNOPSIS style consistent with other pages.
This filters out duplicate addresses from address command output.
It also also adds tests for the address command.
The code here is an extended version of a patch from Jani Nikula.
Now, when address related outputs are in a separate command, it makes
no sense to combine multiple --output options in search command line.
Using switch statement to handle different outputs is more readable
than a series of if statements.
This moves address-related functionality from search command to the
new address command. The implementation shares almost all code and
some command line options.
Options --offset and --limit were intentionally not included in the
address command, because they refer to messages numbers, which users
do not see in the output. This could confuse users because, for
example, they could see more addresses in the output that what was
specified with --limit. This functionality can be correctly
reimplemented for address subcommand later.
Also useless values of --exclude flag were not included in the address
command.
This was inspired by a patch from Jani Nikula.
NOTMUCH_OPT_INHERIT expects a notmuch_opt_desc_t * pointer in
output_var.
The "Unrecognized option" message was moved out of parse_option() to
not be emitted twice or when parsing a non-inherited option.
In the next commit, notmuch_search_command will be refactored to
several smaller functions. In order to simplify the next commit to
verbatim move of several lines to new functions with search_context_t*
argument, we convert all references to this structure to pointer
dereferences. To do so we rename the context variable and use the
original name ctx as the pointer to the renamed structure.
In order to share some command line options between search and address
subcommands we need to add corresponding variables to the context
structure. While we are at it, we also add notmuch_database_t to unify
parameters of all do_search_* functions and to simplify subsequent
commits.
Otherwise, there are no functional changes.
In the next commit the options structure will be extended by
non-option variables. Therefore we need a more generic name.
Just text replacement, no other changes.
The new manual pages converted from rst using sphinx or rst2man
has somewhat different syntax. man-to-mdwn.pl is now adjusted
to produce even better output from this syntax. The changes also
include using utf-8 locale (e.g. for tables and generated hypens)
and and quite a few bugs fixes.
This tool still produces better results than just using the
html pages generated using sphinx / rst2html. For example those
tools don't create inter-page hyperlinks -- and the preformatted
pages written by man-to-mdwn.pl just works well with manual page
content.
python3 doesn't allow dictionaries to be initialized with non-string
keywords. This presents problems on systems in which "python" means
"python3". We instead initalize the dictionary using the dict
comprehension and then update it with the values from the tree. This
will work with both python2 and python3.
There was theorical possibility that writing the config files could
have skipped (by interruption) after the instructions how to make
notmuch was printed out.
The new outputs allow printing senders, recipients or both of matching
messages. To print both, the user can use --output=sender and
--output=recipients simultaneously.
Currently, the same address can appear multiple times in the output.
The next commit will change this. For this reason, tests are
introduced there.
We use mailbox_t rather than InternetAddressMailbox because we will
need to extend it in a following commit.
This code is based on a patch from Jani Nikula.
This converts "notmuch search" to use the recently introduced
keyword-flag argument parser. At this point, it only makes the code
slightly less readable but following commits that add new --output
keywords will profit from this.
Many functions that implement the search command need to access command
line options. Instead of passing each option in a separate variable, put
them in a structure and pass only this structure.
This will become handy in the following commits.
Apparently the test binaries are built with minimal LDFLAGS, only
adding dependency specific LDFLAGS as needed. However because some of
the test binaries incorporate notmuch object files, it is necessary to
use the same link flags as notmuch. For example user provided
CFLAGS/CXXFLAGS/LDFLAGS with -fsanitize=undefined fails to build the
test binaries if the flags differ.
The post-new hook might no longer be needed or run very often if
notmuch insert is being used. Therefore a post-insert hook is needed
(arguably pre-insert not so much, so don't add one). Also add the
--no-hooks option to skip hooks.
Previously, it was necessary to link new messages to children to work
around some (though not all) problems with the old metadata-based
approach to stored thread IDs. With ghost messages, this is no longer
necessary, so don't bother with child linking when ghost messages are
in use.
This updates the thread linking code to use ghost messages instead of
user metadata to link messages into threads.
In contrast with the old approach, this is actually correct.
Previously, thread merging updated only the thread IDs of message
documents, not thread IDs stored in user metadata. As originally
diagnosed by Mark Walters [1] and as demonstrated by the broken
T260-thread-order test, this can cause notmuch to fail to link
messages even though they're in the same thread. In principle the old
approach could have been fixed by updating the user metadata thread
IDs as well, but these are not indexed and hence this would have
required a full scan of all stored thread IDs. Ghost messages solve
this problem naturally by reusing the exact same thread ID and message
ID representation and indexing as regular messages.
Furthermore, thanks to this greater symmetry, ghost messages are also
algorithmically simpler. We continue to support the old user metadata
format, so this patch can't delete any code, but when we do remove
support for the old format, several functions can simply be deleted.
[1] id:8738h7kv2q.fsf@qmul.ac.uk
This updates the message abstraction to support ghost messages: it
adds a message flag that distinguishes regular messages from ghost
messages, and an internal function for initializing a newly created
(blank) message as a ghost message.
