From 69bb29a2914fe072fdc19e029447170cc410ac19 Mon Sep 17 00:00:00 2001
From: Valentin Haudiquet <valentin.haudiquet@canonical.com>
Date: Fri, 2 Oct 2026 01:18:48 +0200
Subject: [PATCH] pull: ignore the pax global header when detecting a tarball
 wrapper

Git-archive tarballs open with a pax global header record
('pax_global_header'): GNU tar consumes it and never lists or
extracts it, but the tar crate yields it as an entry. Counting it
for the wrapper detection made every git-archive orig look
multi-rooted - linux-riscv's orig stayed nested under 'linux-6.8/'
and its .diff.gz failed to apply with 'can't find file to patch'.

Skip the record outright: unpack would no-op on it anyway, so this
only keeps it out of the wrapper detection (and the tree).
---
 src/pull.rs | 47 +++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 47 insertions(+)

diff --git a/src/pull.rs b/src/pull.rs
index 32c02cf..47a455b 100644
--- a/src/pull.rs
+++ b/src/pull.rs
@@ -263,6 +263,19 @@ where
 
     for entry in archive.entries()? {
         let mut entry = entry?;
+
+        // A pax global header record ('pax_global_header', the opening
+        // entry of git-archive tarballs) is metadata about the archive,
+        // not content: GNU tar consumes it and never lists or extracts
+        // it, but the tar crate yields it as an entry. It must not be
+        // extracted (unpack would no-op) and must not count for the
+        // wrapper detection below — it used to make every git-archive
+        // orig look multi-rooted, skipping the unwrap and leaving the
+        // sources nested one level too deep.
+        if entry.header().entry_type().is_pax_global_extensions() {
+            continue;
+        }
+
         let path = entry.path()?.to_path_buf();
 
         // Archives built with './'-prefixed entries (common in third-party
@@ -1235,6 +1248,40 @@ mod tests {
         );
     }
 
+    #[test]
+    fn test_extract_tar_skips_pax_global_header() {
+        // Regression: git-archive tarballs open with a pax global header
+        // record ('pax_global_header'): GNU tar consumes it and never
+        // lists or extracts it, but the tar crate yields it as an entry.
+        // Counting it for the wrapper detection made every git-archive
+        // orig look multi-rooted: the unwrap was skipped, the sources
+        // stayed nested one level too deep and the package's .diff.gz
+        // did not apply (linux-riscv).
+        let temp_dir = tempfile::tempdir().unwrap();
+        let tar_path = temp_dir.path().join("git.orig.tar");
+        let mut tar_bytes = Vec::new();
+        push_tar_entry(
+            &mut tar_bytes,
+            "pax_global_header",
+            b'g',
+            b"52 comment=d0e1c2d",
+            None,
+        );
+        push_tar_entry(&mut tar_bytes, "pkg-1.0/Makefile", b'0', b"all:", None);
+        tar_bytes.extend_from_slice(&[0u8; 1024]);
+        std::fs::write(&tar_path, tar_bytes).unwrap();
+
+        let tree = temp_dir.path().join("tree");
+        let extracted = extract_tar_archive(&tar_path, &tree, None, |f| f).unwrap();
+        unwrap_wrapper(&tree, &extracted.wrapper.unwrap()).unwrap();
+
+        assert!(!tree.join("pax_global_header").exists());
+        assert_eq!(
+            std::fs::read_to_string(tree.join("Makefile")).unwrap(),
+            "all:"
+        );
+    }
+
     #[test]
     fn test_extract_tar_rejects_hardlink_escape() {
         let temp_dir = tempfile::tempdir().unwrap();