From f63e6000f9f9cc6eca309be525aeafea9e9a8cd5 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Padez?= <andre.padez@gmail.com>
Date: Sat, 29 Aug 2026 01:59:22 +0100
Subject: [PATCH] slim-embedded-art: the audio hash excludes a trailing ID3v1
 tag

mutagen updates an existing v1 tag on save - it normalised the genre
byte 0xFF to 0x0F on Shamrocks and Shenanigans and the guard refused a
whole album of good rewrites over one metadata byte. The v1 block is
metadata, not audio; the guard now hashes strictly the frames between
the v2 tag and any trailing TAG block.
---
 scripts/slim-embedded-art.py | 25 ++++++++++++++++++-------
 1 file changed, 18 insertions(+), 7 deletions(-)

diff --git a/scripts/slim-embedded-art.py b/scripts/slim-embedded-art.py
index 1cee63e..bb8f0b2 100755
--- a/scripts/slim-embedded-art.py
+++ b/scripts/slim-embedded-art.py
@@ -42,13 +42,12 @@ PROCESS_OVER = 100_000
 
 
 def audio_region_hash(path, offset=None):
-    """SHA-256 of everything after the ID3v2 tag - the audio, byte-for-byte.
+    """SHA-256 of the audio: after the ID3v2 tag, before any trailing ID3v1 tag.
 
-    The offset comes from mutagen's OWN parse of the tag when available. Reading the header's
-    syncsafe size field ourselves disagreed with mutagen on files whose tagger wrote a
-    non-standard size (Shamrocks and Shenanigans, 2026-08-29): mutagen preserved the audio from
-    where IT parsed the tag to end, our hash read from where the header claimed, and the guard
-    refused a perfectly good rewrite.
+    The v2 offset comes from mutagen's own parse. The trailing 128-byte ID3v1 block (starts
+    "TAG") is EXCLUDED: mutagen updates an existing v1 tag on save - measured on Shamrocks and
+    Shenanigans (2026-08-29), it normalised the v1 genre byte 0xFF to 0x0F (Rap) and the guard
+    refused a whole album of perfectly good rewrites over that one metadata byte.
     """
     if offset is None:
         try:
@@ -62,11 +61,23 @@ def audio_region_hash(path, offset=None):
                     offset = 10 + ((head[6] << 21) | (head[7] << 14) | (head[8] << 7) | head[9])
                     if head[5] & 0x10:  # v2.4 footer
                         offset += 10
+    size = os.path.getsize(path)
+    end = size
+    if size - offset >= 128:
+        with open(path, "rb") as f:
+            f.seek(size - 128)
+            if f.read(3) == b"TAG":
+                end = size - 128
     with open(path, "rb") as f:
         f.seek(offset)
         h = hashlib.sha256()
-        while chunk := f.read(1 << 20):
+        remaining = end - offset
+        while remaining > 0:
+            chunk = f.read(min(1 << 20, remaining))
+            if not chunk:
+                break
             h.update(chunk)
+            remaining -= len(chunk)
         return h.hexdigest()