- new per-character tokenizer that is easier to read and maintain
- OR searches now supported
- grouping with ( and ) now supported
- sorting is off by default now
- searching in html is disabled for now, but may come back later
- use ? arguments instead of named arguments to ease porting to other
platforms
- field:foo is no longer wrapped in implicit wildcards
- avoid joining notes table if it not required - about 30% faster
- fixed sql injection bug in findNids()
- commands no longer have to take care of negation themselves
- commands can return nothing to immediately flag the query as invalid
The way we were handling 1 day+ learning intervals was not great - they'd show
up at the start of a new day before normal reviews, meaning the hardest cards
came first. In previous Anki versions we deliberately sorted the queue in the
opposite order to prevent that. When relearning the cards the next day, if you
failed a card and expected to see it in 10 minutes that wouldn't happen
either, as all the overdue cards took precedence.
To fix this, we put cards that are due tomorrow or later into a separate queue
(queue 3), and pull cards from that queue only after the reviews are done. In
the future it might also be nice to move overdue learning cards into that
queue automatically at the start of a session.
When a user has learning steps that extend past the daily cutoff, we end up
counting them all instead of only the ones that would be done today. In order
to avoid this without expensive calculations or db schema changes, we
calculate the number of steps until the daily cutoff and pack it into the left
column, as totalLeft + leftToday*1000.
- When we reset() the deck, we need to update the daily counts for all decks,
not just the selected ones, or the deck list is inaccurate
- We don't want to save the decks at that time, or simply opening Anki on
another machine will cause a conflict with all decks
