M4A has it's own multi-value spec that works similarly to vorbis, where
they just repeat the atom several times with multiple values. Since M4A
atoms are remapped to ID3v2 frames, this more or less requires us to
tolerate duplicate ID3v2 frames as well, which is frustratingly a spec
violation. It solves the problem though
Resolves#558.
Fix some immediate compile issues and some style nitpicks. This breaks
the system a bit, but I don't think I will have enough time to debug
fully for the forseeable future, so I want to get this out now.
Make it so that music items are meaningfully different when they were
created under different settings. This resolves an issue where music
information would not correctly update when separators or intelligent
sorting would change.
Resolves#546.
Use PlaySong modeling within settings too. This largely completes the
PlaySong refactor and fully adds the ability to play songs by
themselves as an option.
Resolves#424.
Fix an issue where the recent BackportBottomSheetBehavior update would
break queue sheet scrolling, as the queue sheet would be set to
invisible and prevent the findScrollingChild helper from being able
to locate it.
Fix this by removing that visibility check in the first place.
Fix an issue where the recent BackportBottomSheetBehavior update would
break queue sheet scrolling, as the queue sheet would be set to
invisible and prevent the findScrollingChild helper from being able
to locate it.
Fix this by removing that visibility check in the first place.
Further document the music indexing process.
It's so aggressively parallelized as to require some more extensive
comments to actually make it clear what's going on.
Recognize artists sort, albumartists sort, and album artists tags.
These are written by mediafile, so they are probably also written by
other software.
Fix an issue where an upwards nested scroll past the top of the queue
list would not actually collapse the sheet.
This was apparently caused by something introduced in recyclerview
1.3.0-alpha01. No idea why. Roll back to the working version.
Fix a regression where partial ReplayGain adjustments missing an album
or track component would not be indexed.
This was an oversight made when moving adjustments to the cache, as the
nullability model of ReplayGain adjustments changed.
Do not transfer the cache db between devices, drop it instead.
This is at best useless (timestamps and ids will be invalid) and at
worst actively detrimental (timestamps and ids are similar), so
better to reindex than try to read from the cache.
Resolves#467.
Fix issues stemming from how ExoPlayer apparently doesn't send a
playWhenReady event after being stopped.
This ended up breaking AudioEffect integration and notification
posting. I really don't know why player.stop() doesn't do this.
Fix an issue where the playback sheet will suddenly become "stuck" when
playback ends, preventing back navigation or playing new tracks.
This is, once again, caused by bottom sheet insanity. It does not
expose the target state of a sheet at all when it's settling, making
Auxio believe that it has to repeatedly "fix" the state of a hiding
playback sheet and leading to those aforementioned issues.
Fix this by band-aiding it with yet a n o t h e r bottom sheet behavior
extension that exposes the target state. This will be eventually be
used in the whole bottom sheet flow, as it allows me to abort
in-progress sheet transitions, but I'm not going to rock the boat like
this in a patch release. Probably 3.2.0.
Resolves#464.
Fix redundant separator parsing obliterating loading performance.
If there are no separators configured, the parsing function would not
short-circuit and instead do a useless O(n) iteraton (including
escaping!), massively reducing performance.
Song build performance still isn't the best (blame intelligent
sorting), but it's definitely better now.
Fix a few problems with the current comparison algorithm:
1. It wasn't actually comparing the raw music information, only the
UIDs, which was redundant.
2. The comparison in the main music loading process occurred on the
main thread, which causes massive freeze-up issues.
Resolves#457.
Drop support for Lollipop and Marshmallow.
I legit cannot build for them anymore. There's some insane dependency
tangle between Kapt and Room that makes it impossible to use anything
BUT Java 17 (maybe Java 11). If I use Java 17 or Java 11, suddenly the
whole app crashes since Lollipop and Marshmallow only support some weird
Java 8/11 chimera abomination. This issue has persisted probably since
the move to Java 11, making it basically unusable anyway, so at this point
just cut the charade and officially drop support.
Fix a visual error where adding to queue would visually place items
next to the starting index rather than at the end.
This was a result of forgetting to update the Add instructions returned
by Queue.addToNext to return the end of the queue rather than the
current index, which completely broke the queue visual.
Resolves#435.
Unify corner radius and icon size configurations under a new sizing
property. This is largely driven by a need to remove the typical
half-width icon sizing, as it results in blurry playing indicators
in some cases. This also co-incides with a change of parent image
icon sizes to 32dp over 28dp.
Resolves#415.
Group albums implicitly linked to an artist via the "artist" tag into
their own section called "Appears on".
This makes Auxio's artist model a bit more apparent to users.
Resolves#411.
Fix two issues with the ways items are laid out:
1. Remove the automatic span size lookup. Now that ConcatAdpater is
used, this basically becomes impossible to really leverage.
2. Use a divider item instead of a divider item decoration. The
latter is too buggy in many contexts, like the search view.
Resolves#426Resolves#444
Add a real playlist naming dialog and UX flow.
This is a bit rough at the moment since theres a good amount of nuance
here. Should improve as the playlist implementation continues to grow
more fleshed out.
Move everything over to the media3 library instead of ExoPlayer.
Media3 is worse in every way. It labels half of ExoPlayer as "unsafe"
because it thinks that it's garbage uwu "helpful" abstractions are
perfectly servicable when in reality they are a pile of garbage filled
with insane performance issues, race conditions, and a seeming lack
of awareness to the sheer absurdity of android's media APIs. It is
absolutely horrible, but ExoPlayer will stop being maintained soon
and I will have to move over for further maintenance.
Refactor the weird picker god module into specific sub-impls in
playback and a new navigation package.
I cannot keep this unified. The needs are too different among each
picker. Better to keep it separate, especially in preparation for
the playlist dialogs.
If the user clicks on the playback bar in any context, including the
queue view, open the playback panel.
This adds another means to closing the queue that does not involve
swiping.
Resolves#402.
Make the numeric sort name handling added prior dependent on a new
"Intelligent Sorting" setting that also controls the article checks.
This kind of behavior might not be desirable in all cases, and it makes
the setting more consistent anyway.
Make the numeric sort name handling added prior dependent on a new
"Intelligent Sorting" setting that also controls the article checks.
This kind of behavior might not be desirable in all cases, and it makes
the setting more consistent anyway.
Due to a sudden pivot to starting from repository backwards with
playlists, there were some half-baked changesets lying around that
I forgot to revert. Do that.
Broadcast the audiofx session when playing/pausing rather than starting
and stopping.
This is apparently the more correct way to do this.
Resolves#391.
Add fine-grained updates to the queue view.
Should do the following:
1. Make queue updates faster (no diff calculation)
2. Resolve some bugs regarding duplicate queue items.
3. Finally complete the new list framework.
Resolves#350Resolves#335Resolves#380
Not adding this broke MP4-AAC files for some insane reason. Google devs
try not to maintain two near-identical products with only slightly
infuriating differences challenge (IMPOSSIBLE)
Resolves#388.
Add a setting to remove hard-coded sort names based on articles.
This feature is nice, but does not work with some non-english music.
Those individuals should have the ability to disable it.
The implementation honestly is not the greatest, primarily because it
does a 100% reload when it could just regenerate the library. Auxio's
current music model isn't really designed for that, so it will do this
until a need to that kind of "soft reload" really arises.
Resolves#359.
Fix visual clipping on the shuffle FAB's shadow.
Turns out padding, while slower, is actually the better inset handling
method, as it allows me to avoid visual clipping in some cases.
Make sorting direction (ascending/descending) explicit in the UI and in
the code.
Instead of a boolean flag, two distinct "ascending" and "descending"
options are used instead. This should be much clearer.
Accept positions that are zeroed, but only if there are non-zero total
values as well.
Sometimes zeroed positions are deliberate, but other times they are
placeholders meant to indicate a lack of a position. To resolve this,
Auxio now considers zeroed track/disc numbers in the presence of
non-zero track/disc numbers to be valid.
Disable the new instructions-based system for now.
The way I was doing reflection was likely far too unsafe and prone to
bugs. I'll want to do it some other way.
Completely rework the detail list implementations so that resorting the
song list causes a replace operation instead of a diff operation.
This finally makes the list experience consistent across the app.
Make list instructions generic in preparation for the detail list
update.
Detail views need their own instructions datatype, so this is meant to
allow that to be implemented without issue.
Split up the settings ui into four categories.
This should reduce the visual load on the user as Auxio continues to
accrue possible configuration options.
Resolves#323.
Reorganize the music folders dialog to be more visually straightforward
than prior, primarily by grouping the folder elements into the same
visual region.
Resolves#318.
Make all adapters relying on diffing unified into a DiffAdapter
superclass that can then accurately respond to the new
UpdateInstructions data.
UpdateInstructions is still not fully used everywhere, but will be
soon.
When the back button is pressed, clear the current selection before
navigating back.
This is something I was planning to do but then completely forgot about
when implementing multi-select.
Resolves#316.
Switch back to using settings-specific listeners rather than the
SharedPreference listener.
Again, this is due to the need to decouple android code from settings.
It also allows us to fully obscure the details of what settings we are
actually working with.
