spotify: update sendPong, handle MP3_160_ENC #120
Conversation
|
Awesome! How did you manage to reverse engineer it? |
|
Pretty badass 😎 |
…ify-web into theSmallNothing-master # Conflicts: # lib/spotify.js
The small nothing master
|
Cheers guys, @LinusU JPEXS now supporting Alchemy opcode decompilation helped, it would have been a lot more painful to work with the control flow on an unending stream of bytecodes. That cross-referenced with assembly of C++ stdlib allowed all the inlined functions to be removed, leaving what you see in the commit as the main algorithm. Weirdly, there did seem to be a second encryption routine in there that wasn't used, might be something to keep an eye on in the future. |
|
Haven't done extensive testing yet, but this appears to work fine! Very nice work |
add lib/crypto.js: handles all key generation and decryption functionality modify lib/spotify.js: handles sneaky `album_art` messages modify lib/track.js: `play` now includes decryption transformer in pipeline modify lib/util.js: added hex2bin and bin2hex helper functions modify package.json: commit changes require newer `superagent`
ghost
changed the title
modify package.json: remove references to `protobuf` modify lib/schemas.js: remove references to `protobuf`, make `protobufjs` the only parser remove proto/*.desc: only required for the `protobuf` module
So as it turns out, the old code seems to have taken the very first "album_art" message over to every single track that followed, if an Album or Playlist was fetched. This code basically runs it through a new Context within the code-execution VM every time one of those messages is being received.
Fix for trying to fetch a whole album / playlist at once. Fix library method not being able to fetch saved tracks.
|
Is it possible to use part of this code to decrypt 160kb/s MP3s from the Spotify Web-Player or did they changed something in the meantime? |
|
I've already tried without success unfortunately. If somebody finds a way to do it cc me please :) |
|
Strange... The swf key is definitly the same as before - this was the reason I found this website ;-) |
|
I know, but apparently the resulting file is garbage. They may have changed something, or it is me that I'm not good at implementing stuff |
Had a bit of spare time over the holidays to look into this. Apologies for the wall of code (courtesy of Google Closure), it was either that or ~350 lines of pretty random assignments. Anyone manages to cut it down let me know, I'd love to know how.