Writing m4a tags seems to render the file unplayable in some cases

[milesegan]

Reproducer

use std::{
    fs::{copy, remove_file},
    path::Path,
    process::Command,
};

use lofty::{Accessor, Probe, Tag, TagExt, TagType, TaggedFileExt};

pub fn write_tag(dest: &Path) {
    let mut dest_file = Probe::open(&dest)
        .expect("can't open")
        .read()
        .expect("can't read");
    let tag = Tag::new(TagType::Mp4Ilst);
    let mut dest_tag = dest_file.insert_tag(tag).expect("no tag");
    dest_tag.set_artist("test artist".to_string());
    dest_tag.save_to_path(dest).expect("can't write");
}

fn main() {
    let _ = remove_file("encoded.m4a");
    let _ = remove_file("tagged.m4a");
    Command::new("afconvert")
        .args(["-d", "aac", "-f", "m4af", "test.flac", "encoded.m4a"])
        .output()
        .expect("couldn't encode file");
    println!("encoded to encoded.m4a");
    copy("encoded.m4a", "tagged.m4a").expect("copy failed");
    println!("copied to tagged.m4a");
    write_tag(Path::new("tagged.m4a"));
    println!("wrote tag to tagged.m4a");
}

Summary

Writing a new tag to an m4a file seems to corrupt the file somehow and make it unplayable.

Running afplay test.m4a on the file generated above reports this error:

Error: AudioQueueStart failed ('bada')

And the file is also unplayable in the finder.

Note that this only seems to happen with m4a files created with afconvert. If I use ffmpeg to transcode the flac file to m4a and then write the tag with lofty the resulting file is playable.

Expected behavior

The file should still be a playable m4a file after writing the tag.

Assets

https://github.com/milesegan/lofty-m4a-bug

[Serial-ATA]

Thanks for the report and all the information. Could you also provide the file created with afconvert prior to running it through Lofty? I do not have access to afconvert (I think?).

[milesegan]

Sorry about that. I've updated the repo to include the file before tagging and I'm also attaching it here. I believe afconvert is part of the OS on Macs but that's the only way to get access to it.

Thanks for all your work on lofty, by the way!

encoded.m4a.zip

[Serial-ATA]

Awesome, thanks!

I'm pretty sure this is just caused by the offset atoms, which need to be updated anytime the position of atoms change. I forgot they existed, since I don't have any files using them in my library. I'll try and get that fixed in the next couple days.

[milesegan]

Thanks for your quick attention on this!

This year I'm going to try to help you out with some of the good first issue issues.

[Serial-ATA]

Could you try out #318 and see if it fixes the issue for you?

[milesegan]

Yep it seems to be working now. Thanks for the quick fix!