Hi,

I've been using ytermusic a lot and it's been great. I still have a few quality of life improvements I'm thinking about discussing, but this one is definitely the biggest issue, since it actually prevents me from listening to certain music.

The Youtube Music search being what it is, sometimes the right album or song does not show up in the results. In the official website/mobile app, you have to - for example - navigate manually to the artist to get to the album and start playing it.

Making the search better seems like a lot to ask, so here's my idea: a mode where you paste something like https://music.youtube.com/playlist?list=<playlist_id> (or just the playlist ID), then ytermusic shows you the songs in that playlist like in the search view, and you can then decide to play it, just like from the search view.

How does that sound?

Linked with #12
More details in #11

I'm not 100% sure how to reproduce this, but when trying out the basic functionality, I got at least one song with weird compression artifacts for a couple of seconds (sounds like it's underwater in one ear).

A workaround would be to allow resetting the cache for a given song/album (when album support is available), which I think is always a good option to have. :)

I've updated to the latest version (which changed where my caches are, which I'm not complaining about, this was a good change), and now I can't really figure out what the intended way of getting music to play is.

After a random amount of clicking and spacebar action it stars playing, but there is no clear indication of why.

Hi, awesome project! I was really looking for a proper yt music api with offline caching program, so, thanks!

Unfortunately, when I run the program (built from source) nothing is populated in the TUI.
I looked at the log file and found only a single line (which is this song):

InvalidJsonCantFind("navigationEndpoint", "{\"text\":\"In Heat (feat. Hentai Xander)\"}")

I set up the headers according to the instructions, ran ytermusic, pressed f, entered a search term, pressed enter. Nothing happened, no error message printed. Am I supposed to press something else?

I just saw this neat crate pop-up in the Rust subreddit and it made me think that it could be integrated into ytermusic without too much trouble.

The idea would be to use it to get data to feed into a tui widget to replace the current seekbar with a waveform of the currently playing song.

Basically like this plugin in Foobar2000: https://foobar.hyv.fi/?view=foo_wave_minibar_mod

If music is playing, the ctrl+c signal does not propagate to the task playing the music, which forces the user to ctrl+c again. Same goes for ctrl+d, except the second chord still needs to be ctrl+c.

I'm assuming this is related to https://tokio.rs/tokio/topics/shutdown?

• Split the main function into smaller functions with associated doc
• Remove all functions that aren't related to argument parsing and main function from main.rs
• Improve runtime launch checks and error messages

While it's useful for Windows, it isn't on Linux as PipeWire already allows for per-application volume control (see below image). The volume slider tui just makes managing audio harder as it's independent from the native audio controller.

I assume this is an issue with sound under WSL.

After following the steps I get:

michael@Michaels-PC:~/ytermusic$ sudo ./ytermusic
ALSA lib confmisc.c:855:(parse_card) cannot find card '0'
ALSA lib conf.c:5178:(_snd_config_evaluate) function snd_func_card_inum returned error: No such file or directory
ALSA lib confmisc.c:422:(snd_func_concat) error evaluating strings
ALSA lib conf.c:5178:(_snd_config_evaluate) function snd_func_concat returned error: No such file or directory
ALSA lib confmisc.c:1334:(snd_func_refer) error evaluating name
ALSA lib conf.c:5178:(_snd_config_evaluate) function snd_func_refer returned error: No such file or directory
ALSA lib conf.c:5701:(snd_config_expand) Evaluate error: No such file or directory
ALSA lib pcm.c:2664:(snd_pcm_open_noupdate) Unknown PCM default
ALSA lib confmisc.c:165:(snd_config_get_card) Cannot get card index for 0
ALSA lib confmisc.c:165:(snd_config_get_card) Cannot get card index for 0
ALSA lib confmisc.c:165:(snd_config_get_card) Cannot get card index for 0
ALSA lib confmisc.c:165:(snd_config_get_card) Cannot get card index for 0
Cannot connect to server socket err = No such file or directory
Cannot connect to server request channel
jack server is not running or cannot be started
JackShmReadWritePtr::~JackShmReadWritePtr - Init not done for -1, skipping unlock
JackShmReadWritePtr::~JackShmReadWritePtr - Init not done for -1, skipping unlock
Cannot connect to server socket err = No such file or directory
Cannot connect to server request channel
jack server is not running or cannot be started
JackShmReadWritePtr::~JackShmReadWritePtr - Init not done for -1, skipping unlock
JackShmReadWritePtr::~JackShmReadWritePtr - Init not done for -1, skipping unlock
ALSA lib pcm_oss.c:397:(_snd_pcm_oss_open) Cannot open device /dev/dsp
ALSA lib pcm_oss.c:397:(_snd_pcm_oss_open) Cannot open device /dev/dsp
ALSA lib pulse.c:242:(pulse_connect) PulseAudio: Unable to connect: Connection refused

ALSA lib pulse.c:242:(pulse_connect) PulseAudio: Unable to connect: Connection refused

ALSA lib confmisc.c:165:(snd_config_get_card) Cannot get card index for 0
ALSA lib confmisc.c:165:(snd_config_get_card) Cannot get card index for 0
ALSA lib confmisc.c:165:(snd_config_get_card) Cannot get card index for 0
ALSA lib confmisc.c:165:(snd_config_get_card) Cannot get card index for 0
panicked at 'called `Option::unwrap()` on a `None` value', src/systems/player.rs:78:29

Currently, searching triggers an API call for (almost?) each keystroke, which pollutes the (google account) search history on top of being wasteful of resources. The local results maybe could still be matched on each keystroke given that there is no search history nor real performance issues.

A debounce of 150-200ms is usually a good bet?

(I'm writing this issue as I come across it, I will take time to answer the other issue I've opened. :) )

So I just discovered this project and it seems really cool. Unfortunately, I can't seem to get it working.

When starting up the program, I cannot search for playlists or anything, and the log file has an InvalidHTMLFile error at the bottom, followed by a large amount of text.

Regarding the cookie, the readme says "Find the request to the music.youtube.com...", but there are so many requests that I am a bit confused about which one I should use. I have tried POST, GET, and HEAD requests, and I am only looking at OK responses. I have copied the cookie to the headers file and added a user agent. I am certain the issue is from the cookie and not the user agent.

I am using the 'alpha-0.0.10' release as the most recent doesn't have an exe so I couldn't get it running.

I have looked at issues to see if anyone else is having/has had a similar issue, and none of the suggested fixes are working.

Any help would be appreciated!

I've just come back from a work trip so this might be ~a week old, but the online features of ytermusic do not seem to work anymore. The only playlists showing up are the "last playlist" and "local musics" (sic), and the search is only looking for local songs.

I've updated my Cookie header, but it doesn't work either. Was there a change in the API?

This will most likely not end up being relevant but I'll include it anyway:
OS: Arch Linux x86_64
Install method: cloned from repo, built using rustup 1.26.0-3

Actual issue and question:
I'm able to launch ytermusic without any errors being thrown but my personal playlists aren't showing up at all.
Instead I get generic suggestions for people in my region, this makes me feel like I'm not actually logged in but I have no idea if its possible to check.

Am I supposed to copy the cookie from a specific request on music.youtube.com? Currently I have copied the string from the first music.youtube.com request that pops up after refreshing the page. Aka the cookie: field under Request Headers.

My account is using security key 2fa, my assumption is that this doesn't matter since it should pass if the cookie is valid. My browser is Brave.
Obviously I won't post my cookie string, but I assume that user agent is fine. This is the general spec of my headers.txt
Cookie: VERYLONGCOOKIESTRING
User-Agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/113.0.0.0 Safari/537.36
Each field is on a separate line.

When I make a build for MacOS with cargo +nightly build --release I get the following error:

error[E0277]: `souvlaki::Error` doesn't implement `std::fmt::Display`
  --> src/systems/player.rs:26:9
   |
23 |       handle_error_option(
   |       ------------------- required by a bound introduced by this call
...
26 | /         MediaControls::new(PlatformConfig {
27 | |             dbus_name: "ytermusic",
28 | |             display_name: "YTerMusic",
29 | |             hwnd: None,
30 | |         }),
   | |__________^ `souvlaki::Error` cannot be formatted with the default formatter
   |
   = help: the trait `std::fmt::Display` is not implemented for `souvlaki::Error`
   = note: in format strings you may be able to use `{:?}` (or {:#?} for pretty-print) instead
note: required by a bound in `handle_error_option`
  --> src/errors.rs:14:8
   |
8  | pub fn handle_error_option<T, E>(
   |        ------------------- required by a bound in this
...
14 |     T: std::fmt::Display,

Whenever the playlist fills up (which I'm not too sure how, or with what limits?), clicking on a given track will move around all tracks with no clear way of simply looking at the current playlist to pick a song.

Thanks for implementing scrolling. :)

Scrolling in music player view still triggers SoundAction::Previous(1).apply_sound_action(self); or SoundAction::Next(1).apply_sound_action(self);, making it impossible to scroll through the playlist without changing the current song, something I personally do a lot.

A separate counter should be kept for the "currently selected" and "currently playing", offering a similar interface to the playlist view.

If I have the file in /home/user/.ytermusic, then I add that folder to the PATH, and run it, I get the " The 'headers.txt' file is not present" message, even though the headers.txt file is in the .ytermusic folder and configured properly.

It's been a few days at least.

The logs say Error downloading -YXG0fHEFmY: YouTube returned an unexpected response: watch html did not contain a PlayabilityStatus [src/tasks/download.rs] upon failure.

I accidentally removed my .cache folder (I'm sad) and now I can't listen to any music. I haven't looked into the change yet.

I frequently listen to new albums on repeat to get a better feel for them, and sometimes individual songs that I really enjoy. Having a way of setting "repeat album" and "repeat song" would be very nice.

The issue I can see with "repeat album" would be that it could confuse the user intent of whether they're asking to replay the whole media player history (if other things were played before), or just the current album/playlist. Not sure if you simply want a third repeat mode. :)

What do you think?

Sometimes, a song will fail to download and will appear red. I would like to be able to retry this (because it works most of the time if I retry the entire album it's from) without disturbing the current playlist.

It is really convenient to install packages with cargo install package_name. It would be really appreciated if we could install ytermusic this way too.

During troubleshooting for #47, I re-created my headers.txt file and it seems to never be valid even though I did everything the same as before. I've tried multiple times, not sure if anything has changed.

This doesn't seem related to the CONTROL detection but rather a bug in CPAL or RustyBackend
To investigate

The current behavior when the user opens a playlist is to start 4 parallel jobs to download 4 (arbitrary?) songs from the playlist to local cache, then start playing as soon as one is available.

Since streaming is not yet an option, I see two things that would greatly improve the user experience:

• ensuring that the user has at least one song to play as fast as possible
• allowing the user to override the current preferred song to download

The second one should be its own issue because it opens other questions of UI (do we have a way of showing all songs in the playlist and giving control etc.), I only mention it here to be thorough.

When I look at my drive activity, I see that ytermusic is only writing at ~120kb/s when downloading the videos, which is much, much slower than either my gigabit ethernet or my PCIE-4 NVME SSD are capable of handling. This results in having to wait more than 1 minute to start playing a song (and I say a song because I can't really choose which one is being downloaded yet), while 3 others are at 80% completion.

Is there a hard reason why we can't download faster from Youtube? If so, investigating streaming might be the path forwards. On the other hand, if we're limited by the implementation somehow, is the async download pool slow because it's trying to be too clever and not prioritize enough, etc?

I've been using ytermusic for some time. Today I've added (using browser) some new songs to one of my playlists. When I restarted ytermusic (in order to refresh the content of my playlist), it now shows only two items

1. last played playlist which contains old data and
2. local music

There used to be along with those two items several others playlists and albums, but now they are all gone.

Maybe Google has changed something in their data format.

• themes (change colors, texts and more) Maybe some kind of HTML like declaration
• controls (key mapping)
• networking (limits, number of worker...)

Some time I want to remvoe songs to free some space.
How can I do that?

Currently, the program stores the headers.txt header file containing Youtube Music cookie and User Agent in the same directory where the program is executed. I propose moving this header file to a fixed and more appropriate location, such as XDG_CACHE_HOME or $HOME/.cache.

This change would enhance program's cleanliness and facilitate packaging for various Linux distributions. While it's not recommended to scatter files needed by programs in random locations (like /usr/bin or similar prefixes if using an application launcher) cache directories are designed to cater to such use cases.

Your thoughts on this improvement are welcome.

Since you already have the functionality to listen to the playlist of liked songs, could you add the function to save the songs in this playlist?
With a new command, such as CTRL+S (save).

Currently if you start a playlist, then go back to the playlist list view, then back inside the playlist, the first song is started again from the start. IMO it shouldn't do anything.

I guess this ties into the (maybe configurable) idea of not auto-playing when going into a playlist, which would allow for browsing, helping #26 be even more useful.

IMO it makes sense to display that information to give more context to the user.

ytermusic is a great software and I enjoy using it. However, both my laptop and my desktop run on Linux (Arch Linux and NixOS) and I notice some performance issue.

More specifically, on my laptop, it seems that running the app makes the rest of the DE (KDE Plasma) really slow when I change volume and/or tracks. For instance, the way the volume slider interacts is very unreactive (it takes likes 1 or 2 seconds to actually de/increase the volume by 5%, a thing that you can "spam" usually).

Also, on desktop, this time running the app seems to make more laggy the switch between windows.

I know this bug report is really lacking details, and I hope maybe adding more info, but for now, these are the only info I can give. Nonetheless, if I had to make an educated guess, I would say it is related to the way you use the DBUS API, and would suggest a very high number of calls to the DBUS API, such important it starts to make lag the DE (the Audio apps, or more).

I see 3 ways for going about this.

1. Don't draw the text background at all except for when the selector is hovering on an item.
2. Draw the the text background to the exact character length of the item.
3. Implement some form of transparency configuration option.