4.0.0 REGRESSION MEGATHREAD

[OxygenCobalt]

SUBMIT REGRESSIONS AND PROBLEMS WITH 4.0.0 HERE!

Some already known issues:

• Playback state is not persisting during the upgrade (I have yet to debug this due to time constraints) -> dev4
• Weird bouncing effect when navigating to album song (Likely regression) -> dev4
• Skipping effect when editing playlists -> dev4
• Fix broken fab shadow -> dev4
• Out of spec play/pause button -> dev4
• Certain themes are functionally identical -> dev4
• Broken home fast scroller when sorting by date added -> dev4
• Playlist editing isn't working -> dev3
• Cannot restart music loading -> dev3
• Music linking crashes on some libraries with a NullPointerException -> dev3
• No ID3v1 support -> dev3
• Crash when loading FLAC files -> dev2
• Playback controls fall behind queue bar (Unfixable system issue)
• Predictive back can stop working when navigating through library (Likely library issue, cannot be fixed right now)

Some untested stuff:

• Automatic reloading
• Cover behavior edge cases
• Long-term storage use by cover cache
• M3U and Vorbis edge cases

[grapheneloverdev]

#942

[dot166]

#946 ( fixed in 58e0956 )

[grapheneloverdev]

One small issue: the fast scrolling slider isn't pressable when near the shuffle button.

[strongville]

Crashed on background with the following stack trace provided by the app itself (dev2 on Android 15):

android.app.ForegroundServiceStartNotAllowedException: Service.startForeground() not allowed due to mAllowStartForeground false: service org.oxycblt.auxio/.AuxioService
	at android.app.ForegroundServiceStartNotAllowedException$1.createFromParcel(ForegroundServiceStartNotAllowedException.java:54)
	at android.app.ForegroundServiceStartNotAllowedException$1.createFromParcel(ForegroundServiceStartNotAllowedException.java:50)
	at android.os.Parcel.readParcelableInternal(Parcel.java:5105)
	at android.os.Parcel.readParcelable(Parcel.java:5087)
	at android.os.Parcel.createExceptionOrNull(Parcel.java:3267)
	at android.os.Parcel.createException(Parcel.java:3256)
	at android.os.Parcel.readException(Parcel.java:3239)
	at android.os.Parcel.readException(Parcel.java:3181)
	at android.app.IActivityManager$Stub$Proxy.setServiceForeground(IActivityManager.java:7193)
	at android.app.Service.startForeground(Service.java:776)
	at org.oxycblt.auxio.AuxioService$$ExternalSyntheticLambda0.invoke(SourceFile:388)
	at org.oxycblt.auxio.AuxioService.updateForeground(SourceFile:202)
	at org.oxycblt.auxio.music.service.IndexingHolder.onIndexingStateChanged(SourceFile:5)
	at org.oxycblt.auxio.music.MusicRepositoryImpl.access$emitIndexingProgress(SourceFile:97)
	at org.oxycblt.auxio.music.MusicRepositoryImpl$emitIndexingProgress$1.invokeSuspend(Unknown Source:12)
	at kotlin.coroutines.jvm.internal.BaseContinuationImpl.resumeWith(SourceFile:9)
	at kotlinx.coroutines.DispatchedTask.run(SourceFile:109)
	at androidx.core.app.ActivityRecreator$1.run(SourceFile:36)
	at kotlinx.coroutines.scheduling.TaskImpl.run(SourceFile:3)
	at kotlinx.coroutines.scheduling.CoroutineScheduler$Worker.run(SourceFile:93)
	Suppressed: android.app.ForegroundServiceStartNotAllowedException: Service.startForeground() not allowed due to mAllowStartForeground false: service org.oxycblt.auxio/.AuxioService
		... 20 more

Also, after crashing, the library loading is as slow as the first time adding the folder to Auxio (not sure if it is the time it will take when closed normally or because of the same crash).

[OxygenCobalt]

What do you mean in the background @strongville? Were you playing anything? Was Automatic Reloading on?

[strongville]

With background, I meant that I switched apps, but as it only happened once, I really don't remember if it really was like that (I could've left the app open on now playing screen).

The app was playing FLAC from an imported playlist.

Checking the settings, Automatic Reloading is off.

[OxygenCobalt]

Was it paused or not @strongville

[strongville]

Sorry, was not paused, the crash interrupted the playback.

[OxygenCobalt]

Wait, there are two errors here @strongville

Error 1 crashed the app and playback

Error 2 is a weird music loading error I have had a hard time tracking down since I believe it's an Android problem.

[strongville]

Yeah, was not sure how to report that, and at first stance looked related to me.

Thanks for checking out! I'll keep using dev versions and if I notice something else (or the same crash as error 1), I'll try to give more details how happened.

[OxygenCobalt]

Didn't think this was enough for a full issue, so I'll put it here
4.0.0-dev3 - funky shadow

Thanks, will fix with other UI stuff

[WreckingBANG]

Weird Date on all my Music. Everything seems to have the same really wrong Date.

[VoxelPrismatic]

@WreckingBANG please share a couple sample files

[OxygenCobalt]

No @VoxelPrismatic, this is probably actually a bug with the fast scroller

[VoxelPrismatic]

I could not reproduce with my library, @OxygenCobalt

[OxygenCobalt]

It's related to date added support @VoxelPrismatic, I've noticed this error before I just haven't acted on it.

[OxygenCobalt]

Most regressions I believe have been ironed out, a full 4.0.0 release is imminent.

Only remaining change might be to whether "Round mode" is enabled by default [Sharp covers will remain, however]. From what I've observed, more people turn Round mode on rather than off, so it seems to be the more sensible default on my end. It also is a much better companion to the new design. Once this decision is made I'll be prepared to finally launch.

[Fuoxden]

Most regressions I believe have been ironed out, a full 4.0.0 release is imminent.

Only remaining change might be to whether "Round mode" is enabled by default [Sharp covers will remain, however]. From what I've observed, more people turn Round mode on rather than off, so it seems to be the more sensible default on my end. It also is a much better companion to the new design. Once this decision is made I'll be prepared to finally launch.

Very Exciting news! I can't wait for the full 4.0.0 release!

[e-zk]

#960 :(
Occurring after updating to dev-4

[Richard38907]

since 4.0.0 bottom android navigation bar has an ugly black background instead of the transparent one present on 3.x.x.

This black bar is only present on the menus.

[VoxelPrismatic]

since 4.0.0 bottom android navigation bar has an ugly black background instead of the transparent one present on 3.x.x.

This black bar is only present on the menus.

I can't reproduce this, on any set/auto theme or system light/dark theme.

Sony Xperia 1 V; Android 15 (OEM); Auxio 4.0.0-dev4

[Richard38907]

Strange, I have this black bar on my two OnePlus ( OxygenOS and AOSP roms)🤔
This behaviour happens with all themes activated (light, dark and pure black)

[VoxelPrismatic]

What's your build?

[Richard38907]

The same as you

[VoxelPrismatic]

The interesting thing is you also have tab icons. I do not.

[Richard38907]

It's because of the screen density (DPI). Try to set a bigger display size and you will see the difference.

[VoxelPrismatic]

1644x3840 ought to be enough, but nope

[OxygenCobalt]

I think it's because I switched to a different edge-to-edge system in an attempt to fix a bug @Richard38907, I'll see if I can override it with the old transparent bars.

[strongville]

ReplayGain info for OPUS files is not detected.
Using Vorbis tags on those files, as created by default on foobar2000.

foobar2000 showing RG info for an example track on my library:

Track info for same file on Auxio dev04:

OPUS File used:
https://files.catbox.moe/cg2db7.opus

[OxygenCobalt]

Probably an accidental tag error, will investigate @strongville

[OxygenCobalt]

Okay @strongville, the problem is that your file is not using R128 tags but the internal base gain field. OPUS has it's own internal "base gain" value that's supposed to take precedence over ReplayGain. This is automatically handled by the system media decoder, and so there's not really a good way for me to set it up such that it can be displayed in the app without accidentally being double-applied. It'll get iced for now until I rejig my app data structures again to handle gain values that should be displayed but not used.

[strongville]

Ok, I think I'm following until the part about two different gain values being present on OPUS files that are possible to be applied (for that file -8.81dB in opus gain, -3.31dB on ReplayGain info).

Should I try to remove the OPUS gain? (I think I'm using same kind of metadata as my FLAC files and for those works just fine, and on Auxio OPUS files are using the preamp for non-RG-info-found files)

[OxygenCobalt]

Ok, I think I'm following until the part about two different gain values being present on OPUS files that are possible to be applied (for that file -8.81dB in opus gain, -3.31dB on ReplayGain info).

Yeah, it's really confusing design choice on Xiph's end IMO. In general, where you might put an ReplayGain Album Gain in a FLAC file, you put a Base/Header Gain in an OPUS file.

Should I try to remove the OPUS gain? (I think I'm using same kind of metadata as my FLAC files and for those works just fine, and on Auxio OPUS files are using the preamp for non-RG-info-found files)

You can, but taggers tend to default to writing it, so it's more a problem on my end that I'm not displaying it (Hence #968).

I think Auxio does need to rejig it's OPUS Gain handling considering #969.

@strongville

[hknutsen]

@strongville I wrote a script that uses opus-tools to convert my FLAC files to Opus files that I play through Auxio on my phone.

Here's a comment from that script that describes how gain is stored in Opus files:

# The Opus encoder provided by opus-tools will propagate tags from the input
# FLAC file to the output Opus file, except "REPLAYGAIN_*" tags.
#
# Opus follows the EBU R128 specification for loudness normalization.
# According to the Opus specification, gain must be stored in the "Output
# Gain" field in the ID header. Media players should apply this gain by
# default. Additional track and album gain can be stored in the
# "R128_TRACK_GAIN" and "R128_ALBUM_GAIN" tags in the comment header.
# Ref: https://datatracker.ietf.org/doc/html/rfc7845#section-5.2.1
#
# If the input FLAC file has a "REPLAYGAIN_ALBUM_GAIN" tag, its value will be
# converted to the R128 reference level and stored in the "Output Gain" field
# of the output Opus file. If the input FLAC file has a
# "REPLAYGAIN_TRACK_GAIN" tag, its value relative to the album gain will be
# converted to the R128 reference level and stored in the "R128_TRACK_GAIN"
# tag of the output Opus file.
# Ref: https://github.com/xiph/opus-tools/blob/v0.2/src/flac.c#L179-L193
#
# Some media players might require ReplayGain to be turned off in order apply
# the default output gain (i.e. the album gain) without applying the
# additional track gain.

Auxio currently seems to implement support for Opus files as intended by Xiph (developers of Opus):

• Album gain is read from the file header (applied by system media decoder).
• Additional track gain relative to the album gain is read from the R128_TRACK_GAIN tag (applied by Auxio).

With that in mind, here's how to "properly" use gain for Opus files in Auxio:

• If you want to use album gain, set "ReplayGain strategy" to "Off". The system media decoder will apply album gain.
  • If you want to apply pre-amp when using album gain, use "Adjustment without tags".
• If you want to use track gain, set "ReplayGain strategy" to "Prefer track". The system media decoder will apply album gain, then Auxio will apply additional track gain.
  • If you want to apply pre-amp when using track gain, use "Adjustment with tags".