If you have a blob of encrypted data, that should normally tell you nothing about the plaintext, except its approximate length (assuming it wasn't padded). Normally, the length of the plaintext should tell you very little about its content; it doesn't do you much good to know "the message is about 4096 bytes". However, if you compress first, then you know the approximate length of the compressed data, which means you know something about the content of the plaintext.
As one of several possible attacks: imagine the attacker could supply data that will get fed back to them in the encrypted blob, along with your own data, and all compressed together. The blob will compress better if your data matches their data. So, repeatedly feed in your data and get back blobs, watching the total size of the compressed-then-encrypted blob, and you can predict enough about their data to guess it.
As one of several possible attacks: imagine the attacker could supply data that will get fed back to them in the encrypted blob, along with your own data, and all compressed together. The blob will compress better if your data matches their data. So, repeatedly feed in your data and get back blobs, watching the total size of the compressed-then-encrypted blob, and you can predict enough about their data to guess it.