Add symlink traversal to support PDM cached envs (#237)

* Add symlink traversal to support PDM cached envs

Fixes #236

Co-authored-by: Dmitrii Sutiagin <[email protected]>

setup.cfg

@@ -24,7 +24,7 @@ sections = FUTURE,STDLIB,THIRDPARTY,FIRSTPARTY,LOCALFOLDER
 
 [metadata]
 name = shiv
-version = 1.0.3
+version = 1.0.4
 description = A command line utility for building fully self contained Python zipapps.
 long_description = file: README.md
 long_description_content_type = text/markdown

src/shiv/builder.py

@@ -12,9 +12,10 @@
 import zipfile
 
 from datetime import datetime, timezone
+from itertools import chain
 from pathlib import Path
 from stat import S_IFMT, S_IMODE, S_IXGRP, S_IXOTH, S_IXUSR
-from typing import IO, Any, List, Optional, Tuple
+from typing import Generator, IO, Any, List, Optional, Tuple
 
 from . import bootstrap
 from .bootstrap.environment import Environment
@@ -69,6 +70,15 @@ def write_to_zipapp(
     archive.writestr(zinfo, data)
 
 
+def rglob_follow_symlinks(path: Path, glob: str) -> Generator[Path, None, None]:
+    """Path.rglob extended to follow symlinks, while we wait for Python 3.13."""
+    for p in path.rglob('*'):
+        if p.is_symlink() and p.is_dir():
+            yield from chain([p], rglob_follow_symlinks(p, glob))
+        else:
+            yield p
+
+
 def create_archive(
     sources: List[Path], target: Path, interpreter: str, main: str, env: Environment, compressed: bool = True
 ) -> None:
@@ -110,7 +120,11 @@ def create_archive(
                 # Glob is known to return results in non-deterministic order.
                 # We need to sort them by in-archive paths to ensure
                 # that archive contents are reproducible.
-                for path in sorted(source.rglob("*"), key=str):
+                #
+                # NOTE: https://github.com/linkedin/shiv/issues/236
+                # this special rglob function can be replaced with "rglob('*', follow_symlinks=True)"
+                # when Python 3.13 becomes the lowest supported version
+                for path in sorted(rglob_follow_symlinks(source, "*"), key=str):
 
                     # Skip compiled files and directories (as they are not required to be present in the zip).
                     if path.suffix == ".pyc" or path.is_dir():

test/test_builder.py

@@ -9,7 +9,7 @@
 
 import pytest
 
-from shiv.builder import create_archive, write_file_prefix
+from shiv.builder import create_archive, rglob_follow_symlinks, write_file_prefix
 
 UGOX = stat.S_IXUSR | stat.S_IXGRP | stat.S_IXOTH
 
@@ -39,6 +39,16 @@ def test_binprm_error(self):
         with pytest.raises(SystemExit):
             tmp_write_prefix(f"/{'c' * 200}/python")
 
+    def test_rglob_follow_symlinks(self, tmp_path):
+        real_dir = tmp_path / 'real_dir'
+        real_dir.mkdir()
+        real_file = real_dir / 'real_file'
+        real_file.touch()
+        sym_dir = tmp_path / 'sym_dir'
+        sym_dir.symlink_to(real_dir)
+        sym_file = sym_dir / real_file.name
+        assert sorted(rglob_follow_symlinks(tmp_path, '*'), key=str) == [real_dir, real_file, sym_dir, sym_file]
+
     def test_create_archive(self, sp, env):
         with tempfile.TemporaryDirectory() as tmpdir:
             target = Path(tmpdir, "test.zip")