From 339a1791b1f6930c1bbe0c58bb670a0a62496f1b Mon Sep 17 00:00:00 2001
From: Danny McClanahan <1305167+cosmicexplorer@users.noreply.github.com>
Date: Sun, 6 Aug 2023 09:29:28 -0400
Subject: [PATCH] fix reproducibility

---
 pex/common.py                          | 4 +++-
 pex/pex_builder.py                     | 6 +++---
 tests/integration/test_reproducible.py | 3 +++
 3 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/pex/common.py b/pex/common.py
index 898d71964..622eb6c3d 100644
--- a/pex/common.py
+++ b/pex/common.py
@@ -627,11 +627,13 @@ def zip(
                     # If labels are provided, respect the given ordering, but still sort the files
                     # within each label to get deterministic output.
                     sorted(self.filesets.get(label, ()))
+                    # NB: An iterable of labels with non-deterministic order is not reproducible!
                     for label in labels
                 )
             )
         else:
-            selected_files = OrderedSet(self.files())
+            # Otherwise, sort the files to get reproducible output by default.
+            selected_files = OrderedSet(sorted(self.files()))
 
         compression = zipfile.ZIP_DEFLATED if compress else zipfile.ZIP_STORED
         with open_zip(filename, mode, compression) as zf:
diff --git a/pex/pex_builder.py b/pex/pex_builder.py
index 5290e3a4d..926796b46 100644
--- a/pex/pex_builder.py
+++ b/pex/pex_builder.py
@@ -719,7 +719,7 @@ def iter_bootstrap_script_labels(cls):
     def iter_metadata_labels(cls):
         # type: () -> Iterator[str]
         """``Chroot`` labels covering metadata files."""
-        # PEX-INFO: This is accessed after unpacking the zip.
+        # PEX-INFO
         yield "manifest"
 
     @classmethod
@@ -733,8 +733,8 @@ def iter_bootstrap_libs_labels(cls):
     def iter_deps_libs_labels(cls, pex_info):
         # type: (PexInfo) -> Iterator[str]
         """``Chroot`` labels covering the third-party code that was resolved into dists."""
-        # Subdirectories of .deps:
-        for dist_label in pex_info.distributions.keys():
+        # Subdirectories of .deps/: Keys need to be sorted for deterministic output.
+        for dist_label in sorted(pex_info.distributions.keys()):
             yield dist_label
 
     @classmethod
diff --git a/tests/integration/test_reproducible.py b/tests/integration/test_reproducible.py
index aab05c84b..b43970e7b 100644
--- a/tests/integration/test_reproducible.py
+++ b/tests/integration/test_reproducible.py
@@ -97,6 +97,9 @@ def explode_pex(path):
                     ), "{} and {} have different content.".format(member1, member2)
                 # Check that the entire file is equal, including metadata.
                 assert filecmp.cmp(member1, member2, shallow=False)
+            # Check that the file list is identical.
+            with ZipFile(pex1) as zfp1, ZipFile(pex2) as zfp2:
+                assert zfp1.namelist() == zfp2.namelist()
             # Finally, check that the .pex files are byte-for-byte identical.
             assert filecmp.cmp(pex1, pex2, shallow=False)