* qt: normalize large progress totals to kilobytes to avoid QProgressBar overflow (#4341); docs: update CONTRIBUTORS
* update
* revert: keep ProgressManager generic; no autoscaling in progress.py
* revert: keep ProgressManager generic; do scaling only in full sync
* Simplify sync progress update by scaling to KB, fix CONTRIBUTORS
* add anki_i18n and locale_config crates to launcher
* add launcher.ftl
* add tr to state
* replace most hardcoded strings with translations
* add support for `launcher` rustcfg to trim translations
* use marker structs to denote type of translations
* move underscores into generated code
* Update cargo-license, which may fix the license order issue (dae)
* Fix/Ensure fuzz doesn't go backward during rescheduling
Fixes https://github.com/ankitects/anki/issues/2694
* Fix
* Get previous_interval from LastRevlogInfo
* Fix
* Format
* Format
* Exclude lapses
* Force reconfigure in CI
The cached build.ninja may reference files that don't exist in the PR.
On a local build this tends to auto-fix itself as the build scripts detect
a quick failure and re-run the configure, but CI tends to be too slow.
https://github.com/ankitects/anki/pull/4364#issuecomment-3338026129
* Rename min/max to make it clear they restrict interval, not fuzz
* Wording tweaks/comments for clarity
---------
Co-authored-by: Damien Elmes <gpg@ankiweb.net>
* Export last_interval to Python
* Add last_interval field to StatsRevlogEntry
* Implement last_interval_secs function
* Update last_interval field type in stats.proto
* Update last_interval to use last_interval_secs
* Fix/first and latest review dates should only only consider entries with a rating.
* Update rslib/src/stats/card.rs
Co-authored-by: user1823 <92206575+user1823@users.noreply.github.com>
---------
Co-authored-by: user1823 <92206575+user1823@users.noreply.github.com>
* Refactor reviews_for_fsrs function for improved performance
Replaced the previous implementation with a more efficient approach using a single loop and pre-allocated vectors. This change reduces the complexity of creating FSRSItems and enhances overall performance, especially for larger datasets.
* collapse `if` statement
* When not training, only create the final FSRS item
* stats: link retention help box to "Desired retention" section
* docs: use HTTPS for fsrs and desiredRetention links
* Add Hanna Nilsén to CONTRIBUTORS
* Apply suggestion from @user1823
Co-authored-by: user1823 <92206575+user1823@users.noreply.github.com>
---------
Co-authored-by: Hanna Nilsén <hannanilsen@macbookair.students.netlogon.liu.se>
Co-authored-by: user1823 <92206575+user1823@users.noreply.github.com>
* add GetCustomColours rpc method
* save colours as rgb instead of argb
* show saved custom colours as possible options in colour picker
this is primarily for mobile clients, as qt currently ignores this
* save custom colours on colour picker change (for desktop)
* add SaveCustomColours rpc method
* restore custom colour palette on editor init
* save custom colour palette on colour picker open and input
there doesn't seem to be an event fired when the picker is
cancelled/closed, so it's still possible for work to be lost
* save colours on `change` instead of `input`
`input` is supposed to be fired on every adjustment to the picker
whereas `change` is only fired when the picker is accepted, but qt
seems to treat both as the latter, so this is currently a no-op
* Store colors in the collection
One minor tweak to the logic while I was there: an invalid color no
longer invalidates all the rest.
---------
Co-authored-by: Damien Elmes <gpg@ankiweb.net>
* Improved margins in Card Browser's "Preview" pane
* Alternate approach that looks good on Mac too
---------
Co-authored-by: Damien Elmes <gpg@ankiweb.net>
Co-authored-by: Damien Elmes <dae@users.noreply.github.com>
* Make timebox message translatable with flexible variable order
Currently, the timebox dialog message is built from two separate strings,
each containing one variable:
"{ $count } cards studied in" + "{ $count } minutes."
As a result, translators cannot freely reorder the variables in their translations.
This change introduces a single string with both variables, allowing translators
to adjust the order for more natural expressions in their languages.
* Preserve old string for now
* Ensure message doesn't display over two lines
---------
Co-authored-by: Damien Elmes <gpg@ankiweb.net>
With all the recent supply chain attacks, this seems prudent. There are
three in our current package list. esbuild's is just a performance
optimization (https://github.com/evanw/esbuild/issues/4085), and
dprint's gets done when we invoke .bin/dprint anyway. svelte-preprocess
simply prints something to the screen.
