]> git.notmuchmail.org Git - notmuch/log
notmuch
14 years agoFix Database().create_query() to actually initialize the Query with the correct object.
Sebastian Spaeth [Mon, 22 Mar 2010 10:19:51 +0000 (11:19 +0100)]
Fix Database().create_query() to actually initialize the Query with the correct object.

We were passing the internal pointer to the notmuch_database_t object to Query(), but it should be initialized with the Database() object itself. This rightfully raised an exception. Yay, for thorough sanity checking.

14 years agosetup.py and MANIFEST.in for the 0.1 release. Unfortunately post-version tag.
Sebastian Spaeth [Mon, 22 Mar 2010 09:48:03 +0000 (10:48 +0100)]
setup.py and MANIFEST.in for the 0.1 release. Unfortunately post-version tag.

14 years agoAdded tag v0.1 for changeset 27f38f75d27d
Sebastian Spaeth [Mon, 22 Mar 2010 09:40:08 +0000 (10:40 +0100)]
Added tag v0.1 for changeset 27f38f75d27d

14 years agoImplement Message.get_replies()
Sebastian Spaeth [Fri, 19 Mar 2010 13:24:40 +0000 (14:24 +0100)]
Implement Message.get_replies()

14 years agodocs: Add missing newline that broke formatting
Sebastian Spaeth [Fri, 19 Mar 2010 10:01:19 +0000 (11:01 +0100)]
docs: Add missing newline that broke formatting

14 years agoAdd Database.create_query() as a shorthand for db=Database();q=Query(db,"")
Sebastian Spaeth [Fri, 19 Mar 2010 09:55:06 +0000 (10:55 +0100)]
Add Database.create_query() as a shorthand for db=Database();q=Query(db,"")

This is a convenience extension to the C API. I hardly saves any typing, but
let's us automatically free the top-level Database() object when we delete
the Query().

14 years agoDatabase(): implement as of yet untested add_message() and remove_message()
Sebastian Spaeth [Fri, 19 Mar 2010 09:53:14 +0000 (10:53 +0100)]
Database(): implement as of yet untested add_message() and remove_message()

14 years agodocs: Various typo fixes in docstrings
Sebastian Spaeth [Fri, 19 Mar 2010 09:52:15 +0000 (10:52 +0100)]
docs: Various typo fixes in docstrings

14 years agoDatabase(): Factor out the test for an initialized self._db
Sebastian Spaeth [Fri, 19 Mar 2010 09:51:35 +0000 (10:51 +0100)]
Database(): Factor out the test for an initialized self._db

We need it in basically every function, so let's split this out.

14 years agoimplement Message.get_thread_id()
Sebastian Spaeth [Fri, 19 Mar 2010 08:46:44 +0000 (09:46 +0100)]
implement Message.get_thread_id()

14 years agonotmuch restore: only delete a message's old tags if they are not a subset of the...
Sebastian Spaeth [Fri, 19 Mar 2010 08:39:23 +0000 (09:39 +0100)]
notmuch restore: only delete a message's old tags if they are not a subset of the new ones.

14 years agohgignore: ignore test/test*
Sebastian Spaeth [Fri, 19 Mar 2010 08:38:43 +0000 (09:38 +0100)]
hgignore: ignore test/test*

14 years agonotmuch: implement restore. We pass now all 'dump' and 'restore' tests in the test...
Sebastian Spaeth [Fri, 19 Mar 2010 08:11:22 +0000 (09:11 +0100)]
notmuch: implement restore. We pass now all 'dump' and 'restore' tests in the test suite

14 years agonotmuch: Use Query.count_messages() rather len(Query.search_messages())
Sebastian Spaeth [Fri, 19 Mar 2010 07:47:14 +0000 (08:47 +0100)]
notmuch: Use Query.count_messages() rather len(Query.search_messages())

In my tests, this is about 10 times faster and returned the same results,
although both operations are really fast.

14 years agoQuery(): Fix _nmlib -> nmlib type & clarify docs a bit
Sebastian Spaeth [Fri, 19 Mar 2010 07:44:34 +0000 (08:44 +0100)]
Query(): Fix _nmlib -> nmlib type & clarify docs a bit

14 years agoImplement Database.count_messages()
Sebastian Spaeth [Fri, 19 Mar 2010 07:32:29 +0000 (08:32 +0100)]
Implement Database.count_messages()

14 years agonotmuch: implement dump to file. Fake notmuch new by calling the real notmuch to...
Sebastian Spaeth [Thu, 18 Mar 2010 18:11:22 +0000 (19:11 +0100)]
notmuch: implement dump to file. Fake notmuch new by calling the real notmuch to enable running the testsuite

14 years agoadd original notmuch testsuite
Sebastian Spaeth [Thu, 18 Mar 2010 18:04:49 +0000 (19:04 +0100)]
add original notmuch testsuite

14 years agomention NOTMUCH_CONFIG in the notmuch docstring
Sebastian Spaeth [Thu, 18 Mar 2010 17:51:25 +0000 (18:51 +0100)]
mention NOTMUCH_CONFIG in the notmuch docstring

14 years agoDatabase(): honor NOTMUCH_CONFIG env variable for reading the standard database location.
Sebastian Spaeth [Thu, 18 Mar 2010 17:48:36 +0000 (18:48 +0100)]
Database(): honor NOTMUCH_CONFIG env variable for reading the standard database location.

Also replaces all Database.MODE.* defaults in the keyword arguments with their numeric replacement, they do not seem to be available yet at that time, so the python parser complained about them as being unknown.

14 years agoImplement Message.freeze(), thaw(), and remove_all_tags()
Sebastian Spaeth [Thu, 18 Mar 2010 08:34:10 +0000 (09:34 +0100)]
Implement Message.freeze(), thaw(), and remove_all_tags()

14 years agodocs: stub out remaining missing bindings to get a reminder
Sebastian Spaeth [Thu, 18 Mar 2010 08:02:59 +0000 (09:02 +0100)]
docs: stub out remaining missing bindings to get a reminder

14 years agoimplement Database.get_version() and Database.needs_upgrade()
Sebastian Spaeth [Thu, 18 Mar 2010 07:50:48 +0000 (08:50 +0100)]
implement Database.get_version() and Database.needs_upgrade()

14 years agonotmuch: implement tag command
Sebastian Spaeth [Wed, 17 Mar 2010 16:32:37 +0000 (17:32 +0100)]
notmuch: implement tag command

14 years agonotmuch: fix to use new Query.SORT enum
Sebastian Spaeth [Wed, 17 Mar 2010 16:11:58 +0000 (17:11 +0100)]
notmuch: fix to use new Query.SORT enum

14 years agoImplement Message.add|remove_tag() and documentation
Sebastian Spaeth [Wed, 17 Mar 2010 16:09:58 +0000 (17:09 +0100)]
Implement Message.add|remove_tag() and documentation

14 years agodocs: doc improvements
Sebastian Spaeth [Wed, 17 Mar 2010 14:14:44 +0000 (15:14 +0100)]
docs: doc improvements

14 years agomany doc improvements, fixed at least one bug due to parameter renaming
Sebastian Spaeth [Wed, 17 Mar 2010 11:32:22 +0000 (12:32 +0100)]
many doc improvements, fixed at least one bug due to parameter renaming

14 years agoImprove source documentation
Sebastian Spaeth [Wed, 17 Mar 2010 10:45:13 +0000 (11:45 +0100)]
Improve source documentation

14 years agoImplement an Enum class and make the STATUS object one
Sebastian Spaeth [Wed, 17 Mar 2010 10:44:36 +0000 (11:44 +0100)]
Implement an Enum class and make the STATUS object one

14 years agoREADME: add link to static docs
Sebastian Spaeth [Wed, 17 Mar 2010 08:28:39 +0000 (09:28 +0100)]
README: add link to static docs

14 years agoupdate documentation so it makes sense
Sebastian Spaeth [Tue, 16 Mar 2010 19:54:43 +0000 (20:54 +0100)]
update documentation so it makes sense

--HG--
extra : transplant_source : %3F%E9%23%2B%2A%83%06g%10R%CA%CD%27N5%3E%22g%1E%AD

14 years agoadd sphinx docs stubs (they don't make sense yet)
Sebastian Spaeth [Tue, 16 Mar 2010 19:36:04 +0000 (20:36 +0100)]
add sphinx docs stubs (they don't make sense yet)

--HG--
extra : transplant_source : %F3A%2C%D95%02%BF%1B%A1%D7%17%07%EDqh%23l8%93%20

14 years agohgignore: update to ignore static documentation
Sebastian Spaeth [Tue, 16 Mar 2010 19:35:34 +0000 (20:35 +0100)]
hgignore: update to ignore static documentation

--HG--
extra : transplant_source : %9C%CCh%60z%FE%7B1%AC%E6%88%DF7%00%22%7D%09i%EDL

14 years agoimplement Message.get_date() and Message.get_header()
Sebastian Spaeth [Tue, 16 Mar 2010 16:10:57 +0000 (17:10 +0100)]
implement Message.get_date() and Message.get_header()

--HG--
extra : transplant_source : %90%E9%A5Z%D1%E6%26%1F%D3%8B%CCsbb%E9%17%FA%5CD%9F

14 years agoimplement stub for notmuch show. Remove finished ToDo items
Sebastian Spaeth [Tue, 16 Mar 2010 15:41:33 +0000 (16:41 +0100)]
implement stub for notmuch show. Remove finished ToDo items

--HG--
extra : transplant_source : 2%0B%97h%40Af%07%DD%F2%5ESX%20%05%3D%C0%C5v%F8

14 years agoimplement Query.set_sort() and sort notmuch dump by message id
Sebastian Spaeth [Tue, 16 Mar 2010 15:20:47 +0000 (16:20 +0100)]
implement Query.set_sort() and sort notmuch dump by message id

--HG--
extra : transplant_source : 1%BC%F3%ED%3C%C7c%0EFh%06%B6L%7C%8E%8F%EF%A2%8E%A8

14 years agoimplement notmuch dump and Messages.len()
Sebastian Spaeth [Tue, 16 Mar 2010 15:09:42 +0000 (16:09 +0100)]
implement notmuch dump and Messages.len()

--HG--
extra : transplant_source : %03_%C7%99L%16%7E%A8%B1p%11%11%A0%E3h%A3%AB%B6%FB%86

14 years agoimplement message counting
Sebastian Spaeth [Tue, 16 Mar 2010 14:53:54 +0000 (15:53 +0100)]
implement message counting

--HG--
extra : transplant_source : %8E%83%C2%83%FA%F8X%B6%16%1D%D3X%C4o%A2%A8%28%11G%AF

14 years agoimplement quoatation mangling in the notmuch binary
Sebastian Spaeth [Tue, 16 Mar 2010 14:40:13 +0000 (15:40 +0100)]
implement quoatation mangling in the notmuch binary

--HG--
extra : transplant_source : %E2I%C6%0A%05%3B%F3%27%07%96%DC%D6%91%C3%FA%8E%1B%5B%2B%3D

14 years agoFix iterator classes to not skip the first element when iterating
Sebastian Spaeth [Tue, 16 Mar 2010 14:00:42 +0000 (15:00 +0100)]
Fix iterator classes to not skip the first element when iterating

--HG--
extra : transplant_source : %19wvB%19A%0A%CD%E7%28-%F0%12j%7FG%0DD%16%F4

14 years agouse logging.debug for debug output. Implement notmuch search-tags
Sebastian Spaeth [Tue, 16 Mar 2010 13:57:07 +0000 (14:57 +0100)]
use logging.debug for debug output. Implement notmuch search-tags

--HG--
extra : transplant_source : %BAn%2B%93B%1EkU8%A6-I%5D%E1%E4%2B%D6E%0C%F5

14 years agoFixing the crashes that I was seeing
Sebastian Spaeth [Tue, 16 Mar 2010 13:45:01 +0000 (14:45 +0100)]
Fixing the crashes that I was seeing

--HG--
extra : transplant_source : %3E%B1%BB%DF%3C%22%A6%CC%90Z%E8v%95%962%DD%B2%A1%11%CC

14 years agoimplement Query() and search_messages()
Sebastian Spaeth [Tue, 16 Mar 2010 13:28:05 +0000 (14:28 +0100)]
implement Query() and search_messages()

However, this is still brittle and often segfaults for unknown reasons

--HG--
extra : transplant_source : %D1%FB%2B%FC%EA%E9%EB%3E%D2%13%DFV.o%D6%A6%CE%81%00%8C

14 years agoremove useless tags struct
Sebastian Spaeth [Tue, 16 Mar 2010 11:40:20 +0000 (12:40 +0100)]
remove useless tags struct

--HG--
extra : transplant_source : %EF%CD%D8K%E5%E1F%95%1E%F9%F3%88%E7%B0%600%21Wwp

14 years agoQuery() stub
Sebastian Spaeth [Tue, 16 Mar 2010 11:39:47 +0000 (12:39 +0100)]
Query() stub

--HG--
extra : transplant_source : %1E%14%7B%E7J%DD%96O%C7%E6%B5%FB%D8V%7B%0F%7C%25ot

14 years agosome more stubs
Sebastian Spaeth [Tue, 16 Mar 2010 11:28:05 +0000 (12:28 +0100)]
some more stubs

--HG--
extra : transplant_source : %2A%C5%3C%C0u%B8%0DLx%3Eg%A2%06%F6%1E%B8%99C7%B3

14 years agononfunctional changes in notmuch stub
Sebastian Spaeth [Tue, 16 Mar 2010 11:24:52 +0000 (12:24 +0100)]
nonfunctional changes in notmuch stub

--HG--
extra : transplant_source : %F2%28%18m%91%F5%DA%08%17%AC/%9E%C2%AE%95%91%92%C4%A3%7F

14 years agoadd notmuch stub that will allow me to run the notmuch testsuite over the python...
Sebastian Spaeth [Tue, 16 Mar 2010 11:21:18 +0000 (12:21 +0100)]
add notmuch stub that will allow me to run the notmuch testsuite over the python bindings.

--HG--
extra : transplant_source : %E4%EF%29%A3%9BA%C57%7C%EC%B0%7BF%FB%00%F0%EB%97%92%E3

14 years agoFind the notmuch shared lib rather than hardode a location
Sebastian Spaeth [Tue, 16 Mar 2010 10:52:56 +0000 (11:52 +0100)]
Find the notmuch shared lib rather than hardode a location

--HG--
extra : transplant_source : %DEOdF%80%83%3A%3D%98%5C%AE%81%FE%BA%EF%1A%16%82u%E8

14 years agoImplement Message(), Database.find_message(), Database.create()
Sebastian Spaeth [Tue, 16 Mar 2010 10:27:31 +0000 (11:27 +0100)]
Implement Message(), Database.find_message(), Database.create()

Message() basically has get_message_id get_filename get_tags
Plus various bullet proofing and bug fixing.

--HG--
extra : transplant_source : O%3B1%EB%E0%D4pYrEY_%E3%0E%BA%C23%11a%B0

14 years agoimprove README example session to reflect reality
Sebastian Spaeth [Mon, 15 Mar 2010 16:12:52 +0000 (17:12 +0100)]
improve README example session to reflect reality

--HG--
extra : transplant_source : %15P%B8E%1A%CE%0D%DE%CEB%E0/%E04%2B%3F%255%A1%F9

14 years agoTags: clarify documentation: we can only iterate once over Tags, and free underlying...
Sebastian Spaeth [Mon, 15 Mar 2010 15:00:11 +0000 (16:00 +0100)]
Tags: clarify documentation: we can only iterate once over Tags, and free underlying tags once we iterated through.

--HG--
extra : transplant_source : %0BI%C9%899k%80%8C4j%07%85%038%86%04%2A%FE%A6%A3

14 years agoinitial revision. See README for what works
Sebastian Spaeth [Mon, 15 Mar 2010 14:47:15 +0000 (15:47 +0100)]
initial revision. See README for what works

--HG--
extra : transplant_source : %FBQ%B3%C5%0E%85%E8%0B%154%F5t%19%2A%C6p%A5%1Dd%F1

14 years agoCOPYING: add GPL v3+
Sebastian Spaeth [Tue, 16 Mar 2010 21:48:35 +0000 (22:48 +0100)]
COPYING: add GPL v3+

14 years agonotmuch-show: add tags to json output
Sebastian Spaeth [Thu, 11 Mar 2010 12:11:43 +0000 (13:11 +0100)]
notmuch-show: add tags to json output

The previous json patches forgot to add the notmuch tags to the json
output. This is inconsistent to the text output so here they are. We
just output a 'tags' field that sends notmuch tags as a json array.

Signed-off-by: Sebastian Spaeth <Sebastian@SSpaeth.de>
14 years agoemacs: Fix search refresh when on the last line of a search buffer.
Carl Worth [Wed, 10 Mar 2010 19:07:58 +0000 (11:07 -0800)]
emacs: Fix search refresh when on the last line of a search buffer.

We currently allow the cursor to be positioned on the "End of search
results" line after the last thread in a search buffer. When
refreshing on this line, there's no thread ID to be used as the
target.

Previously, a refresh from this case would result in a nil thread
target, but we were also using nil to indicate that the target thread
had been found. This caused the position to be lost during refresh,
(the cursor would move from the last line in the buffer to the first).

We fix this by using a magic string of "found" rather than nil to
clearly indicate whether the target thread has actually been found.

14 years agoemacs: Adjust search refresh to use a target line not a target position.
Carl Worth [Wed, 10 Mar 2010 19:05:33 +0000 (11:05 -0800)]
emacs: Adjust search refresh to use a target line not a target position.

It doesn't make sense to move the cursor to some random point in the
middle of a line. We always want the refresh to leave the cursor at
the beginning of some line instead.

14 years agoMakefile: Fix Makefiles to depend on all child Makefile fragments.
Carl Worth [Wed, 10 Mar 2010 18:59:57 +0000 (10:59 -0800)]
Makefile: Fix Makefiles to depend on all child Makefile fragments.

We were previously maintaining two lists of the child Makefile
fragments---one for the includes and another for the dependencies. So,
of course, they drifted and the dependency list wasn't up to date.

We fix this by adding a single subdirs variable, and then using GNU
Makefile substitution to generate both the include and the dependency
lists.

Some side effect of this change caused the '=' assignment of the dir
variable to not work anymore. I'm not sure why that is, but using ':='
makes sense here and fixes the problem.

14 years agoMakefile: Use 'emacs --quick' for a less noisy build of "make install-emacs".
Carl Worth [Wed, 10 Mar 2010 18:58:46 +0000 (10:58 -0800)]
Makefile: Use 'emacs --quick' for a less noisy build of "make install-emacs".

I don't really notice if it goes any quicker, but it's sure nice to have
less spew now.

14 years agoMakefile: Add a message after "make install-emacs"
Carl Worth [Wed, 10 Mar 2010 18:50:20 +0000 (10:50 -0800)]
Makefile: Add a message after "make install-emacs"

More help to guide the new user here. Tell the user how to actually
invoke the emacs client now that it's installed.

14 years agoMakefile: Conditionalize the "make install" message.
Carl Worth [Wed, 10 Mar 2010 18:48:47 +0000 (10:48 -0800)]
Makefile: Conditionalize the "make install" message.

This is the same approach as with the 'all' target previously.

14 years agoMakefile: Simplify the conditional message of the all target.
Carl Worth [Wed, 10 Mar 2010 18:44:44 +0000 (10:44 -0800)]
Makefile: Simplify the conditional message of the all target.

We wamt a simple "make" to call the 'all' target and then print a
message when done, but we don't want "make install" which depends on
that same 'all' target to print the message.

We previously did this with a separate 'all-without-message' target,
which was inelegant because it caused all users of the target to
carefully depend on 'all-without-message' rather than 'all'.

Instead, we now use a single 'all' target but with a Makefile
conditional that examines the MAKECMDGOALS variable to determine
whether to print the message.

14 years agoMakefile: Add message to make install listing the other install targets.
Carl Worth [Wed, 10 Mar 2010 18:07:34 +0000 (10:07 -0800)]
Makefile: Add message to make install listing the other install targets.

Otherwise, it's hard for the user to know that things like install-emacs,
install-bash, and install-zsh even exist.

14 years agoMakefile: Add a meesage after "make" telling the user to run "make install"
Carl Worth [Wed, 10 Mar 2010 01:03:11 +0000 (17:03 -0800)]
Makefile: Add a meesage after "make" telling the user to run "make install"

As one command completes, it's kind of the tool to indicate which
command the user should execute next.

14 years agoMakefile: Rename all_deps to global_deps
Carl Worth [Wed, 10 Mar 2010 01:01:55 +0000 (17:01 -0800)]
Makefile: Rename all_deps to global_deps

The "all" inside this variable name was easy to confuse with the
separate "all" target. This variable specifies dependencies that apply
to every target, so use "global" instead.

14 years agoAdd is:<tag> as a synonym for tag:<tag> in search terms.
Carl Worth [Wed, 10 Mar 2010 00:03:58 +0000 (16:03 -0800)]
Add is:<tag> as a synonym for tag:<tag> in search terms.

I like the readability of this, it provides compatibility with people
trained in this syntax by sup, and it even saves one character.

14 years agoemacs: Move emacs UI (currently just one file) to subdirectory.
David Bremner [Wed, 10 Feb 2010 03:45:28 +0000 (23:45 -0400)]
emacs: Move emacs UI (currently just one file) to subdirectory.

Add emacs/Makefile.local and emacs/Makefile. Move emacs targets into
emacs/Makefile.local, but leave the byte compilation rule in the top
level Makefile.

14 years agoINSTALL: Include Fedora command for installing dependencies of notmuch.
Carl Worth [Tue, 9 Mar 2010 20:12:32 +0000 (12:12 -0800)]
INSTALL: Include Fedora command for installing dependencies of notmuch.

We already had this command in the error message from the configure script,
so we should include it here as well.

14 years agoINSTALL: Add a pointer to ./configure --help
Carl Worth [Tue, 9 Mar 2010 20:11:47 +0000 (12:11 -0800)]
INSTALL: Add a pointer to ./configure --help

We have some good documentation in ./configure --help, so we should
direct users to it.

14 years agolib: Silence a compiler warning.
Carl Worth [Tue, 9 Mar 2010 20:05:43 +0000 (12:05 -0800)]
lib: Silence a compiler warning.

The original code was harmless, but apparently some compilers aren't
able to think deep enough to catch that.

14 years agoemacs: Fix refresh of search results to leave cursor on current thread.
Carl Worth [Tue, 9 Mar 2010 19:36:08 +0000 (11:36 -0800)]
emacs: Fix refresh of search results to leave cursor on current thread.

This has been broken since the addition of support for streaming
search results to the emacs interface. With the old apparoach it was
easy to simply wait for all search results to land in the buffer and
then search for the desired line. With streaming results, we have to
save the target off to the side and allow the process filter and
sentinel functions to correctly respond when the target thread
appears.

14 years agoFix a few documentation typos in notmuch.h
Fernando Carrijo [Sat, 5 Dec 2009 19:32:34 +0000 (17:32 -0200)]
Fix a few documentation typos in notmuch.h

Fix a few documentation typos in notmuch.h

Signed-off-by: Fernando Carrijo <fcarrijo@yahoo.com.br>
14 years agoUpdate documentation of notmuch_query_create
Fernando Carrijo [Sat, 5 Dec 2009 14:22:05 +0000 (12:22 -0200)]
Update documentation of notmuch_query_create

Commit cd467caf renamed notmuch_query_search to notmuch_query_search_messages.
Commit 1ba3d46f created notmuch_query_search_threads. We better keep the docs
of notmuch_query_create consistent with those changes.

Signed-off-by: Fernando Carrijo <fcarrijo@yahoo.com.br>
Edited-by: Carl Worth to explicitly list the full name of each
function being referenced.

14 years agonotmuch show: Don't show empty headers.
Carl Worth [Tue, 9 Mar 2010 18:12:58 +0000 (10:12 -0800)]
notmuch show: Don't show empty headers.

This is a fairly old regression. There has always been code to avoid
printing empty headers (such as Cc or Bcc with no values), but it has
been broken since notmuch_message_get_header was changed to return
an empty string rather than a NULL pointer for these fields.

14 years agoTODO: Add proposal for a saved-search interface.
Carl Worth [Tue, 9 Mar 2010 18:09:40 +0000 (10:09 -0800)]
TODO: Add proposal for a saved-search interface.

Adding this to our TODO list so that it doesn't get forgotten.

14 years agoTODO: Remove many items that have been completed recently.
Carl Worth [Tue, 9 Mar 2010 18:06:37 +0000 (10:06 -0800)]
TODO: Remove many items that have been completed recently.

There's been a lot of good work done, and we've been doing a generally
poor job of noticing when some of the tasks we've completed were
already on our TODO list.

So here's a careful scan, removing all items I could find that have
already been done.

14 years agoemacs: Fix documentation of notmuch-search-remove-tag.
Carl Worth [Tue, 9 Mar 2010 17:54:53 +0000 (09:54 -0800)]
emacs: Fix documentation of notmuch-search-remove-tag.

The behavior was changed in commit 4aff2ca55bfe285ced36e9fe02c907d8b4120672
to affect all messages in the thread (and not only those matching
the current search) but the documentation was not updated (until
now).

14 years agolib: Document what move_to_next does at the end of the list.
Carl Worth [Tue, 9 Mar 2010 17:24:45 +0000 (09:24 -0800)]
lib: Document what move_to_next does at the end of the list.

Explicitly mention that there's an invalid position after the last
item in the list.

14 years agolib: Rename iterator functions to prepare for reverse iteration.
Carl Worth [Tue, 9 Mar 2010 17:22:29 +0000 (09:22 -0800)]
lib: Rename iterator functions to prepare for reverse iteration.

We rename 'has_more' to 'valid' so that it can function whether
iterating in a forward or reverse direction. We also rename
'advance' to 'move_to_next' to setup parallel naming with
the proposed functions 'move_to_first', 'move_to_last', and
'move_to_previous'.

14 years agoemacs: Fix backspace to not scroll more than the previous message
Carl Worth [Tue, 9 Mar 2010 16:11:45 +0000 (08:11 -0800)]
emacs: Fix backspace to not scroll more than the previous message

The bug was occuring due to counting invisible lines, but then
scrolling past them since they are invisible.

14 years agoemacs: Unbreak notmuch-show-rewind (the function to which Backspace is bound)
Carl Worth [Tue, 9 Mar 2010 01:24:34 +0000 (17:24 -0800)]
emacs:  Unbreak notmuch-show-rewind (the function to which Backspace is bound)

Commit 095a02211e696434e5b41a85ab516b3a639f9a9b broke the backspace
key by trying to execute nil as a function, (which obviously won't
work), when it was intended as a return value. Fix this now, (and pine
for a test suite that exercises the emacs user-interface of notmuch).

14 years agoemacs: Make 'n' and 'p' navigate only open messages.
Carl Worth [Fri, 26 Feb 2010 19:30:12 +0000 (11:30 -0800)]
emacs: Make 'n' and 'p' navigate only open messages.

And add new 'N' and 'P' keybindings for navigating through messages
that are open or closed.

14 years agoemacs: Don't open unread messages by default.
Carl Worth [Fri, 26 Feb 2010 18:40:23 +0000 (10:40 -0800)]
emacs: Don't open unread messages by default.

When searching for an individual message, (by message id, say), it's
really annoying to have the entire thread open just because the thread
was archived without ever having been read.

This means that the "unread" tag is a lot less special, and it really
just exists as a mild cue for the user.

14 years agocontrol: Update package description to follow upstream README.
Carl Worth [Thu, 25 Feb 2010 19:11:59 +0000 (11:11 -0800)]
control: Update package description to follow upstream README.

Justin B Rye pointed out (in Debian bug #566282) that a user with mail
in mbox format can spend a lot of time investigating notmuch before
realizing that mbox is not supported. Head that off with a more
detailed mention in the package description.

14 years agoREADME: Mention that notmuch only supports maildir or mh format.
Carl Worth [Thu, 25 Feb 2010 19:03:56 +0000 (11:03 -0800)]
README: Mention that notmuch only supports maildir or mh format.

Justin B Rye pointed out (in Debian bug #566282) that a user with mail
in mbox format can spend a lot of time investigating notmuch before
realizing that mbox is not supported. Head that off with a more
detailed mention in the README blurb.

14 years agoemacs: Avoid removing the unread tag due to internal navigation
Carl Worth [Thu, 25 Feb 2010 01:02:31 +0000 (17:02 -0800)]
emacs: Avoid removing the unread tag due to internal navigation

Sometimes the internals of the implementation navigate among messages,
(as opposed to the user explicitly requesting the next message to be
shown). In these cases we don't want to remove the unread tag from the
message navigated to.

This fixes a bug where invocation of notmuch-show-next-unread-message
would clear the unread tag from all messages in a thread.

14 years agoSimplify "unread" tag handling in emacs UI.
Jameson Rollins [Tue, 19 Jan 2010 22:54:16 +0000 (17:54 -0500)]
Simplify "unread" tag handling in emacs UI.

This patch is intended to greatly simplify the handling of the
"unread" tag in the emacs UI.  This patch adds a new function
'notmuch-show-mark-read', that removes the "unread" tag in
notmuch-show-mode.  This function is then executed as a
notmuch-show-hook, and by notmuch-show-next-message.  All of the
functions that explicitly marked messages as unread are removed or
renamed.

The idea here is that the user should never have to worry about
manipulating the "unread" tag.  The tag should persist until the user
actually views a message, at which point it should be automatically
removed.  This patch simplifies the notmuch.el quite a bit, removing a
lot of somewhat redundant functions, and reducing clutter in the name
and key-mapping space.

14 years agoTODO: We should fix the --format=json option to not imply --entire-thread.
Carl Worth [Tue, 23 Feb 2010 20:51:23 +0000 (12:51 -0800)]
TODO: We should fix the --format=json option to not imply --entire-thread.

What we print and how we print it are orthogonal options, so --format
shouldn't change what messages are printed.

14 years agonotmuch search: Use "thread" rather than "id" when formatting with JSON
Carl Worth [Tue, 23 Feb 2010 20:41:59 +0000 (12:41 -0800)]
notmuch search: Use "thread" rather than "id" when formatting with JSON

The text output uses thread:<foo>, (which is a syntax directly supported
by "notmuch show"), so make the json output be "thread", "<foo>" rather
than "id", "<foo>". This should help avoid confusion of thread IDs with
message IDs, (which do use the "id" prefix in searches).

14 years agojson: Add copy of MIT license text from cJSON
Carl Worth [Tue, 23 Feb 2010 20:49:33 +0000 (12:49 -0800)]
json: Add copy of MIT license text from cJSON

When we incorporate external code, we should include its license,
(particularly when one of the terms of the license is to include it in
copies).

14 years agoAdd an "--format=(json|text)" command-line option to both notmuch-search and notmuch...
Scott Robinson [Thu, 31 Dec 2009 15:17:40 +0000 (11:17 -0400)]
Add an "--format=(json|text)" command-line option to both notmuch-search and notmuch-show.

In the case of notmuch-show, "--format=json" also implies
"--entire-thread" as the thread structure is implicit in the emitted
document tree.

As a coincidence to the implementation, multipart message ID numbers are
now incremented with each part printed. This changes the previous
semantics, which were unclear and not necessary related to the actual
ordering of the message parts.

14 years agoTODO: Rename the proposed --for option to --output
Carl Worth [Tue, 23 Feb 2010 19:27:43 +0000 (11:27 -0800)]
TODO: Rename the proposed --for option to --output

We've been using --output in IRC and on the mailing list for a while,
(someone had the good sense to point out that --for would defeat
command-line completion since it's a prefix of the proposed --format).

14 years agonotmuch.1: Use bold and indentation for the NOTMUCH_CONFIG variable.
Carl Worth [Tue, 23 Feb 2010 19:21:41 +0000 (11:21 -0800)]
notmuch.1: Use bold and indentation for the NOTMUCH_CONFIG variable.

This seems to be the standard method of formatting an environment
variable section within a man page.

14 years agoAdd ENVIRONMENT VARIABLES section to the man page
James Westby [Sat, 19 Dec 2009 13:12:15 +0000 (13:12 +0000)]
Add ENVIRONMENT VARIABLES section to the man page

Briefly describe the NOTMUCH_CONFIG variable there.

14 years agonotmuch.el: Emphasize the 'i' of 'ID' in the documentation for 'c i'.
Carl Worth [Sat, 20 Feb 2010 20:27:37 +0000 (12:27 -0800)]
notmuch.el: Emphasize the 'i' of 'ID' in the documentation for 'c i'.

We're using 'i' in the keybinding, so it helps to have a capital
'I' in the help string to empahsize the point.

14 years agoChange the stash keybinding from 'z' to 'c'. And use 'i' for message ID.
Carl Worth [Sat, 20 Feb 2010 20:23:13 +0000 (12:23 -0800)]
Change the stash keybinding from 'z' to 'c'. And use 'i' for message ID.

In spite of being implemented with the word "stash" in the function
names, the documentation (and hence help strings) for each function
already use the word "Copy" to describe the action. So 'c' is a much
more natural key-binding, (particularly since 'z' didn't map to any
real word anyway).

We also use 'i' for the message ID copying of the submap. This is
intended to align mnemonically with the "id:" prefix already used
for message IDs.

14 years agonotmuch.el: add a submap (on "z" for "ztash") to stash things.
David Bremner [Fri, 12 Feb 2010 03:01:08 +0000 (23:01 -0400)]
notmuch.el: add a submap (on "z" for "ztash") to stash things.

Provide key bindings for stuffing various RFC822 header fields and other metadata
into the emacs kill-ring as text. The bindings are as follows:

z F notmuch-show-stash-filename
z T notmuch-show-stash-tags
z c notmuch-show-stash-cc
z d notmuch-show-stash-date
z f notmuch-show-stash-from
z m notmuch-show-stash-message-id
z s notmuch-show-stash-subject
z t notmuch-show-stash-to

14 years agonotmuch.el: convert sparse keymap to a list in notmuch-substitute-one-command-key...
David Bremner [Fri, 12 Feb 2010 03:01:07 +0000 (23:01 -0400)]
notmuch.el: convert sparse keymap to a list in notmuch-substitute-one-command-key-with-prefix

The previous version would crash when a key was bound to a sparse
keymap, since apparently these are not straightforward lists.  The
usage of map-keymap is a bit obscure: it only has side-effects, no
return value.

14 years agoAdd functions notmuch-show-get-(bcc, cc, date, from, subject, to).
David Bremner [Fri, 12 Feb 2010 02:44:10 +0000 (22:44 -0400)]
Add functions notmuch-show-get-(bcc, cc, date, from, subject, to).

Return the corresponding header field for the current message as a
string.  These are thin wrappers around notmuch-show-get-header, which
means they each cause a full parse of the RFC822 header. The main idea
is to fix an api.