Gzip responses not decompressed — `invalid character '\x1f'` errors

[tmchow]

The do() method in internal/client/eightsleep.go sets Accept-Encoding: gzip (line 284) but never decompresses the response body. Since the header is set explicitly, Go's http.Client does not auto-decompress. This causes invalid character '\x1f' looking for beginning of value errors on most endpoints (device info, metrics intervals, status, etc.).

$ eightctl device info --quiet
Error: invalid character '\x1f' looking for beginning of value

Fix: Check resp.Header.Get("Content-Encoding") == "gzip" and wrap resp.Body with gzip.NewReader() before decoding:

var respBody io.Reader = resp.Body
if resp.Header.Get("Content-Encoding") == "gzip" {
    gr, err := gzip.NewReader(resp.Body)
    if err == nil {
        defer gr.Close()
        respBody = gr
    }
}

Then use respBody instead of resp.Body for io.ReadAll and json.NewDecoder calls in the same function.

Environment

• Go 1.25.6, Linux x86_64
• eightctl built from latest main (Feb 2026)