When studying, the learning count now indicates the number of
learning cards due within the learn ahead limit, instead of the total
number of learning steps required to complete that day.
Also fix the ineffective limit clauses in the learning counts.
To avoid all sorts of problems, we need to ensure cards scheduled with
the V2 scheduler are not studied in older clients. Unfortunately we
can't just bump the file's schema version, as the existing clients will
freely import files created with newer versions. This patch changes
that, so things should be a bit easier in the future.
In the mean time, we need a way to prevent older clients from importing
files created with the V2 scheduler. To do this, we switch to using a
'collection.anki21' file in the archive, and include a dummy collection
.anki2 file.
The code has been tested with both deck and collection packages, but
exporting deck packages w/ scheduling info will remain disabled until
the V2 scheduler has had more testing.
We can't preserve the original queues when in preview mode, as
otherwise the due counts report the remaining steps of cards in
the learning queue, instead of just 1.
Rather than the rather complicated approach of making the learning and
deck list code aware of the current mode we're in, preview mode moves
all cards to the review queue when the filtered deck is built - just as
cards are moved to the new queue in Anki 2.0.x. The reason for the
review queue is that users were frequently confused when cards appeared
as new - hopefully this is slightly less confusing.
- fetch reviews from all child decks at once, sorted by due order
- shuffle the gathered cards as we did previously
- review limits on child decks are ignored - only the current deck and
its parents control what the limit is
- to make the deck list consistent with actual counts, we can't sum the
child counts, as the sum in the parent limit>child limit case may not
reflect the actual number of cards that would be presented
the previous approach meant we weren't able to preserve the card state
exactly when cards were in learning, since we didn't record the step
position prior to cards being moved into the filtered deck.
it also meant the answer buttons needed to change depending on state - 4
for cards in learning/review, but 2 when the card is on the final step
or is a review.
instead, in preview mode cards always have 2 buttons: again will repeat
again after a delay, and good immediately removes the card and restores
it to its previous state.
to accomplish this, we use a separate queue #, as the learn count
always needs to have a 1:1 correspondence to the number of cards
- move fuzzing into _constrainedIvl() so it's applied prior to limits
like maxIvl
- don't fuzz early reviews, so cards get the same interval if a filtered
deck is rebuilt again
- pasting in prewrap mode had been broken for a while and nobody noticed
- needs changes to export code and cloze handling code to deal with the
newlines
- newline rewriting in filterNode() is likely fragile
- may revisit post 2.1
- enable input keys for video seeking
- output debug info to stdout so it's not caught by anki's error handler
- before sending a command, check the process is still alive, and
restart if necessary - otherwise a user who closes a video window will
end up with errors
- allows translations of filename
- allows users to keep multiple collection exports in the same folder
- provides a clearer distinction between deck and collection packages
- the collection/backup .apkg special cases will continue to work in
future 2.1.x releases
- drop support for $$ in favour of separate opening and closing tags
- add shortcuts to add mathjax
- don't highlight any clozes in blue if field contains mathjax chars
Experiment with ignoring dirMod and scanning the folder for changes on
each sync. Use scandir() for better performance. Please let me know if
this causes a noticeable slowdown for you.
Users can now mark individual cards with one of four different coloured
flags, instead of relying on a tag that applied to the whole note.
- replaced marking functionality in reviewer and browser with new
flag options
- added flag:x search
- marked and leech tags now show in normal tag list in filter screen,
instead of being treated specially
- the other clients will need updating to set and shown the flags, but
flags set in the beta should be preserved by the other clients
- js code that was previously bundled in .py files is now in the
web folder
- add helpers to create links to bundled files, and update
stdHtml() to accept a list of javascript files to include
instead of text
- render MathJax in card layout and preview screens - these should be
updated in the future to update the document dynamically like the
reviewer does
- start media server earlier so it can be used to serve content for
the toolbar, etc
- work around a bug in WebEngine on Windows that could cause the
media server to hang
- add option to wrap html in implicit pre-wrap environment during
editing and review - defaults to off
- update paste filter to convert divs/Ps to newlines and non-breaking
spaces to normal ones
- catch enter key and write \n instead of creating a new div
also:
- remove extra caretToEnd() call that is no longer required
- add dd/dt/dl to allowed tags
It looks like pyqt5 is playing dirty and checking the number of args of
functions before calling them. When using hooks.wrap, pyqt5 thinks it
can pass any amount of arguments (because *args) and you get exceptions
like this inside the wrap function, when calling the 'old' function:
>TypeError: onFindDupes() takes 1 positional argument but 2 were given
This commit fixes it by preserving the signature of the wrapped method,
by adding an optional dependency on the "decorator" module.
Making it an optional dependency is probably not the wisest idea but
since this is a small edge case it might be smoother to start like this.
I also added functools.wraps() as a fallback, which won't help much but
is slightly more correct.
See this article for details: https://hynek.me/articles/decorators/
