Annotated Diff Format

Models are bad at counting lines from a @@ -41,6 +41,7 @@ header, so Vizzini doesn’t ask them to. Every line of every hunk carries both line numbers in a fixed-width gutter:

### internal/store/users.go (modified)
@@ -41,6 +41,7 @@ func (s *Store) FindUser(ctx context.Context, name string) (*User, error) {
41 41       if name == "" {
42 42           return nil, ErrEmptyName
43 43       }
44    -     row := s.db.QueryRowContext(ctx, "SELECT id, name FROM users WHERE name = $1", name)
   44 +     q := "SELECT id, name FROM users WHERE name = '" + name + "'"
   45 +     row := s.db.QueryRowContext(ctx, q)
45 46       var u User
46 47       return &u, row.Scan(&u.ID, &u.Name)

Line layout

<old> <new> <op> <text>

Column	Content
old	Line number in the old file, blank for an added line
new	Line number in the new file, blank for a deleted line
op	+ added, - deleted, blank for context

Findings use the new column. For a problem on a deleted line, the model is told to point at the nearest surviving new line.

File headers

Each file starts with one line describing the change:

Header	When
### path (modified)	Content changed
### path (added) / (deleted)	New or removed file
### path (renamed from old/path)	Renamed, possibly with edits
### path (copied from old/path)	Copied
…, mode 100644 -> 100755)	Appended when the file mode changed
…, binary; content omitted)	Binary file. The header is sent, the content is not
### path (modified) [continued]	A later part of a file split across chunks