With talloc, we were already freeing all memory by the time we exited
the loop, but that didn't help with excess use of memory inside the
loop, (which was mostly from tallocing some objects with the incorrect
parent).
Thanks to Andrew Tridgell for sitting next to me and teaching me to
use talloc_report_full to find these leaks.
As mentioned in the recent test commits, this also fixes the missing
'[' and ']' characters in the --format=json variant of "notmuch search
--output=tags" as well.
This change affects both text and json formats. Now, text format
behaves as before commit 6dcb7592, but json format is changed.
Earlier, the empty search returned '[]', now it returns ''. The emacs
interface seems not to be affected by this change.
Now that notmuch_query_search_threads can return NULL, (for example,
due to a Xapian exception), we need to handle that case (rather than
just segfault). It's simple enough to just return a non-zero exit
code.
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'.
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).
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.
This was a poor workaround around the fact that the existing
notmuch_threads_t object is implemented poorly. It's got a fine
iterartor-based interface, but the implementation does all of the
work up-front in _create rather than doing the work incrementally
while iterating.
So to start fixing this, first get rid of all the hacks we had working
around this. This drops the --first and --max-threads options from the
search command, (but hopefully nobody was using them
anyway---notmuch.el certainly wasn't).
This way, the user gets a steady (but bursty) stream of reults. We
double the chunk size each time since each successive chunk has to
redo work from all previous chunks.
Of course, the overall time is thereby slower, as the price we pay for
increased responsiveness. With a search returning about 17000 thread
results I measured a total time of 48.8 seconds before this change and
58.4 seconds afterwards.
We only rarely need to actually open the database for writing, but we
always create a Xapian::WritableDatabase. This has the effect of
preventing searches and like whilst updating the index.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Carl Worth <cworth@cworth.org>
That is, give a nice error message and exit if no search terms are
provided. Thanks to Priit Laes <plaes@plaes.org> for reporting the
error and providing an early version of the fix.
This is what most people want for a _search_ command. It's often
different for actually reading mail in an inbox, (where it makes more
sense to have results displayed in chronological order), but in such a
case, ther user is likely using an interface that can simply pass the
--sort=oldest-first option to "notmuch search".
Here we're also change the sort enum from NOTMUCH_SORT_DATE and
NOTMUCH_SORT_DATE_REVERSE to NOTMUCH_SORT_OLDEST_FIRST and
NOTMUCH_SORT_NEWEST_FIRST. Similarly we replace the --reverse option
to "notmuch search" with two options: --sort=oldest-first and
--sort=newest-first.
Finally, these changes are all tracked in the emacs interface, (which
has no change in its behavior).
This is one of those cases where total time is not the metric of
interest. We increase the total time of the search, (by doing some
redundant work for the initial threads). But more significantly, we
give the user *some* results nearly instantaneously, (so that the user
might see the result of interest without ever even waiting for the
complete results to come in).
Note that the difference between thread results in date order and
thread results in reverse-date order is not simply a matter of
reversing the final results. When sorting in date order, the threads
are sorted by the oldest message in the thread. When sorting in
reverse-date order, the threads are sorted by the newest message in
the thread.
This difference means that we might want an explicit option in the
interface to reverse the order, (even though the default will be to
display the inbox in date order and global searches in reverse-date
order).
Note that we don't print the number of *unread* messages, but instead
the number of messages that matched the search terms. This is in
keeping with our philosophy that the inbox is nothing more than a
search view. If we search for messages with an inbox tag, then that's
what we'll get a count of. (And if somebody does want to see unread
counts, then they can search for the "unread" tag.)
Getting the number of matched messages is really nice when doing
historical searches. For example in a search like:
notmuch search tag:sent
(where the "sent" tag has been applied to all messages originating
from the user's email address)---here it's really nice to be able to
see a thread where the user just mentioned one point [1/13] vs. really
getting involved in the discussion [10/29].
This time, things are actually tested. The current results aren't
exactly the same as previous results since the incremental search
doesn't necessarily see all the new messages that pertain to the
thread. This means that some author names are missing.
I plan to fix this by doing an additional database search for all
messages in each thread. Of course, this will also be different than
before since now the result will display *all* authors in the thread
(rather than only those that matched the search) but that's probably
what we really want to display anyway.
The library interface now allows the caller to do incremental searches,
(such as one page of results at a time). Next we'll just need to hook
this up to "notmuch search" and the emacs interface.
It's important to have the names present for determining whether a
thread is worth reading or not. We may want to think about
abbreviating the list somehow if it is excessively long (or redundant
as in bugzilla-daemon, bugzilla-daemon, bugzilla-daemon, etc.).
All of the following commands:
notmuch dump
notmuch reply
notmuch restore
notmuch search
notmuch show
notmuch tag
were calling notmuch_database_open with an argument of NULL. This was
a legitimate call until the recent addition of configuration, after
which it is expected that all commands will lookup the correct path in
the configuration file. So fix all these commands to do that.
Also, while touching all of these commands, we fix them to use the
talloc context that is passed in rather than creating a local talloc
context. We also switch from using goto for return values, to doing
direct returns as soon as an error is detected, (which can be leak
free thanks to talloc).
Now that the client sources are alone here in their own directory,
(with all the library sources down inside the lib directory), we can
break the client up into multiple files without mixing the files up.
The hope is that these smaller files will be easier to manage and
maintain.
