PyHardLinkBackup 1.8.4__tar.gz → 1.9.1__tar.gz

{pyhardlinkbackup-1.8.4 → pyhardlinkbackup-1.9.1}/.pre-commit-config.yaml

@@ -2,6 +2,6 @@
 # See https://pre-commit.com for more information
 repos:
   - repo: https://github.com/jedie/cli-base-utilities
-    rev: v0.29.1
+    rev: v0.30.0
     hooks:
       - id: update-readme-history

pyhardlinkbackup-1.9.1/.venv-app/lib/python3.14/site-packages/cli_base/tests/shell_complete_snapshots/.gitignore

@@ -0,0 +1 @@
+!.*

{pyhardlinkbackup-1.8.4 → pyhardlinkbackup-1.9.1}/PKG-INFO

@@ -1,15 +1,15 @@
 Metadata-Version: 2.4
 Name: PyHardLinkBackup
-Version: 1.8.4
+Version: 1.9.1
 Summary: HardLink/Deduplication Backups with Python
 Project-URL: Documentation, https://github.com/jedie/PyHardLinkBackup
 Project-URL: Source, https://github.com/jedie/PyHardLinkBackup
 Author-email: Jens Diemer <PyHardLinkBackup@jensdiemer.de>
 License: GPL-3.0-or-later
 License-File: LICENSE
-Requires-Python: >=3.12
+Requires-Python: >=3.13
 Requires-Dist: bx-py-utils
-Requires-Dist: cli-base-utilities>=0.27.1
+Requires-Dist: cli-base-utilities>=0.30.0
 Requires-Dist: rich
 Requires-Dist: tyro
 Description-Content-Type: text/markdown
@@ -298,6 +298,11 @@ Overview of main changes:
 
 [comment]: <> (✂✂✂ auto generated history start ✂✂✂)
 
+* [v1.9.1](https://github.com/jedie/PyHardLinkBackup/compare/v1.9.0...v1.9.1)
+  * 2026-04-14 - fix pipx installation and set requires-python to ">=3.13"
+* [v1.9.0](https://github.com/jedie/PyHardLinkBackup/compare/v1.8.4...v1.9.0)
+  * 2026-04-14 - Update existing links in has database
+  * 2026-04-14 - Update requirements
 * [v1.8.4](https://github.com/jedie/PyHardLinkBackup/compare/v1.8.3...v1.8.4)
   * 2026-04-09 - Update requirements
   * 2026-04-09 - Apply project updates
@@ -305,6 +310,9 @@ Overview of main changes:
 * [v1.8.3](https://github.com/jedie/PyHardLinkBackup/compare/v1.8.2...v1.8.3)
   * 2026-04-01 - Update requirements
   * 2026-03-30 - Bump pygments from 2.19.2 to 2.20.0
+
+<details><summary>Expand older history entries ...</summary>
+
 * [v1.8.2](https://github.com/jedie/PyHardLinkBackup/compare/v1.8.1...v1.8.2)
   * 2026-03-28 - Update tests
   * 2026-03-28 - fix code style
@@ -320,9 +328,6 @@ Overview of main changes:
   * 2026-01-22 - rebuid command: skip hashing same files by check the inode uniqueness
   * 2026-01-22 - Add "fs-info" in dev cli
   * 2026-01-22 - rebuild command: fix wrong progress bar
-
-<details><summary>Expand older history entries ...</summary>
-
 * [v1.8.0](https://github.com/jedie/PyHardLinkBackup/compare/v1.7.3...v1.8.0)
   * 2026-01-22 - Add optional "--name" to enforce a name for the backup sub directory
   * 2026-01-22 - Do not cross filesystem boundaries as default

{pyhardlinkbackup-1.8.4 → pyhardlinkbackup-1.9.1}/PyHardLinkBackup/__init__.py

@@ -3,5 +3,5 @@
 """
 
 # See https://packaging.python.org/en/latest/specifications/version-specifiers/
-__version__ = '1.8.4'
+__version__ = '1.9.1'
 __author__ = 'Jens Diemer <PyHardLinkBackup@jensdiemer.de>'

{pyhardlinkbackup-1.8.4 → pyhardlinkbackup-1.9.1}/PyHardLinkBackup/backup.py

@@ -130,7 +130,6 @@ def backup_one_file(
                 else:
                     logger.info('Store unique file: %s to %s', src_path, dst_path)
                     dst_path.write_bytes(file_content)
-                    hash_db[file_hash] = dst_path
                     backup_result.copied_files += 1
                     backup_result.copied_size += size
 
@@ -146,10 +145,12 @@ def backup_one_file(
                 else:
                     logger.info('Copy unique file: %s to %s', src_path, dst_path)
                     copy_with_progress(src_path, dst_path, progress=progress, total_size=size)
-                    hash_db[file_hash] = dst_path
                     backup_result.copied_files += 1
                     backup_result.copied_size += size
 
+            # Store new file in hash database or update existing entry to latest backuped file:
+            hash_db[file_hash] = dst_path
+
             # Keep original file metadata (permission bits, time stamps, and flags)
             shutil.copystat(src_path, dst_path)
         else:

{pyhardlinkbackup-1.8.4 → pyhardlinkbackup-1.9.1}/PyHardLinkBackup/cli_app/__init__.py

@@ -31,9 +31,10 @@ def version():
 
 
 def main(args: Sequence[str] | None = None):
-    print_version(PyHardLinkBackup)
+    project_name = 'phlb'  # Enforce program name if pipx used
+    print_version(PyHardLinkBackup, project_name=project_name)
     app.cli(
-        prog='phlb',  # Enforce program name if pipx used
+        prog=project_name,
         description=constants.CLI_EPILOG,
         use_underscores=False,  # use hyphens instead of underscores
         sort_subcommands=True,

{pyhardlinkbackup-1.8.4 → pyhardlinkbackup-1.9.1}/PyHardLinkBackup/tests/test_backup.py

@@ -274,6 +274,7 @@ class BackupTreeTestCase(
                 'wb backups/source/2026-01-01-123456/min_sized_file1.bin',
                 'w backups/.phlb/hash-lookup/bb/c4/bbc4de2ca238d1ec41fb622b75b5cf7d31a6d2ac92405043dd8f8220364fefc8',
                 'a backups/source/2026-01-01-123456/SHA256SUMS',
+                'w backups/.phlb/hash-lookup/bb/c4/bbc4de2ca238d1ec41fb622b75b5cf7d31a6d2ac92405043dd8f8220364fefc8',
                 'a backups/source/2026-01-01-123456/SHA256SUMS',
                 'w backups/source/2026-01-01-123456-summary.txt',
             ],
@@ -320,7 +321,7 @@ class BackupTreeTestCase(
             assert_hash_db_info(
                 backup_root=self.backup_root,
                 expected="""
-                    bb/c4/bbc4de2ca238d1… -> source/2026-01-01-123456/min_sized_file1.bin
+                    bb/c4/bbc4de2ca238d1… -> source/2026-01-01-123456/min_sized_file2.bin
                     e3/71/e3711d0eacddeb… -> source/2026-01-01-123456/large_file1.bin
                 """,
             )
@@ -416,15 +417,15 @@ class BackupTreeTestCase(
             redirected_out.stdout,
         )
 
-        # The FileHashDatabase remains the same:
+        # The FileHashDatabase always points to the latest backed-up files:
         with self.assertLogs('PyHardLinkBackup', level=logging.DEBUG):
             assert_hash_db_info(
                 backup_root=self.backup_root,
                 expected="""
                     23/d2/23d2ce40d26211… -> source/2026-01-02-123456/min_sized_file_newA.bin
                     9a/56/9a567077114134… -> source/2026-01-02-123456/min_sized_file_newB.bin
-                    bb/c4/bbc4de2ca238d1… -> source/2026-01-01-123456/min_sized_file1.bin
-                    e3/71/e3711d0eacddeb… -> source/2026-01-01-123456/large_file1.bin
+                    bb/c4/bbc4de2ca238d1… -> source/2026-01-02-123456/min_sized_file2.bin
+                    e3/71/e3711d0eacddeb… -> source/2026-01-02-123456/large_file2.bin
                 """,
             )
 
@@ -463,9 +464,13 @@ class BackupTreeTestCase(
                 'a backups/source/2026-01-02-123456/SHA256SUMS',
                 'wb backups/source/2026-01-02-123456/hardlink2file1',
                 'a backups/source/2026-01-02-123456/SHA256SUMS',
+                'w backups/.phlb/hash-lookup/e3/71/e3711d0eacddeb105af4ad9b0d63069d759acf32e49712663419e68dc294a94a',
                 'a backups/source/2026-01-02-123456/SHA256SUMS',
+                'w backups/.phlb/hash-lookup/e3/71/e3711d0eacddeb105af4ad9b0d63069d759acf32e49712663419e68dc294a94a',
                 'a backups/source/2026-01-02-123456/SHA256SUMS',
+                'w backups/.phlb/hash-lookup/bb/c4/bbc4de2ca238d1ec41fb622b75b5cf7d31a6d2ac92405043dd8f8220364fefc8',
                 'a backups/source/2026-01-02-123456/SHA256SUMS',
+                'w backups/.phlb/hash-lookup/bb/c4/bbc4de2ca238d1ec41fb622b75b5cf7d31a6d2ac92405043dd8f8220364fefc8',
                 'a backups/source/2026-01-02-123456/SHA256SUMS',
                 'wb backups/source/2026-01-02-123456/min_sized_file_newA.bin',
                 'w backups/.phlb/hash-lookup/23/d2/23d2ce40d26211a9ffe8096fd1f927f2abd094691839d24f88440f7c5168d500',
@@ -499,8 +504,8 @@ class BackupTreeTestCase(
         # Don't create broken hardlinks!
 
         """DocWrite: README.md ## FileHashDatabase - Missing hardlink target file
-        If a hardlink source from a old backup is missing, we cannot create a hardlink to it.
-        But it still works to hardlink same files within the current backup.
+        Deleting files from old backups is safe: the hash DB entry always points to the
+        most recently backed-up file, so subsequent backups can still create hardlinks.
         """
 
         # Let's remove one of the files used for hardlinking from the first backup:
@@ -515,8 +520,8 @@ class BackupTreeTestCase(
         self.assertIn('Backup complete', redirected_out.stdout)
         backup_dir = result.backup_dir
 
-        # Note: min_sized_file1.bin and min_sized_file2.bin are hardlinked,
-        # but not with the first backup anymore! So it's only nlink=2 now!
+        # Note: min_sized_file1.bin and min_sized_file2.bin accumulate hardlinks
+        # because hash_db always points to the latest backup file.
         with self.assertLogs('PyHardLinkBackup', level=logging.DEBUG):
             assert_fs_tree_overview(
                 root=backup_dir,
@@ -527,8 +532,8 @@ class BackupTreeTestCase(
                     hardlink2file1           12:00:00     file            1      14  8a11514a
                     large_file1.bin          12:00:00     hardlink        5    1001  fb3014ff
                     large_file2.bin          12:00:00     hardlink        5    1001  fb3014ff
-                    min_sized_file1.bin      12:00:00     hardlink        2    1000  f0d93de4
-                    min_sized_file2.bin      12:00:00     hardlink        2    1000  f0d93de4
+                    min_sized_file1.bin      12:00:00     hardlink        5    1000  f0d93de4
+                    min_sized_file2.bin      12:00:00     hardlink        5    1000  f0d93de4
                     min_sized_file_newA.bin  12:00:00     hardlink        2    1001  a48f0e33
                     min_sized_file_newB.bin  12:00:00     hardlink        2    1000  7d9c564d
                     small_file_newA.txt      12:00:00     file            1      10  76d1acf1
@@ -547,26 +552,26 @@ class BackupTreeTestCase(
                 backup_count=12,
                 backup_size=6091,
                 symlink_files=1,
-                hardlinked_files=5,
-                hardlinked_size=5003,
-                copied_files=6,
-                copied_size=1074,
+                hardlinked_files=6,
+                hardlinked_size=6003,
+                copied_files=5,
+                copied_size=74,
                 copied_small_files=5,
                 copied_small_size=74,
                 error_count=0,
             ),
         )
 
-        # Note: min_sized_file1.bin is now from the 2026-01-03 backup!
+        # All files points now to "2026-01-03" and non of them to the first "2026-01-01" backup:
         self.assertEqual(backup_dir.name, '2026-01-03-123456')  # Latest backup dir name
         with self.assertLogs('PyHardLinkBackup', level=logging.DEBUG):
             assert_hash_db_info(
                 backup_root=self.backup_root,
                 expected="""
-                    23/d2/23d2ce40d26211… -> source/2026-01-02-123456/min_sized_file_newA.bin
-                    9a/56/9a567077114134… -> source/2026-01-02-123456/min_sized_file_newB.bin
-                    bb/c4/bbc4de2ca238d1… -> source/2026-01-03-123456/min_sized_file1.bin
-                    e3/71/e3711d0eacddeb… -> source/2026-01-01-123456/large_file1.bin
+                    23/d2/23d2ce40d26211… -> source/2026-01-03-123456/min_sized_file_newA.bin
+                    9a/56/9a567077114134… -> source/2026-01-03-123456/min_sized_file_newB.bin
+                    bb/c4/bbc4de2ca238d1… -> source/2026-01-03-123456/min_sized_file2.bin
+                    e3/71/e3711d0eacddeb… -> source/2026-01-03-123456/large_file2.bin
                 """,
             )
 
@@ -916,6 +921,7 @@ class BackupTreeTestCase(
             [
                 'w backups/.phlb_test',
                 'a backups/source/2026-02-22-123456-backup.log',
+                'w backups/.phlb/hash-lookup/23/d2/23d2ce40d26211a9ffe8096fd1f927f2abd094691839d24f88440f7c5168d500',
                 'a backups/source/2026-02-22-123456/SHA256SUMS',
                 'wb backups/source/2026-02-22-123456/large_fileB.txt',
                 'w backups/.phlb/hash-lookup/2a/92/2a925556d3ec9e4258624a324cd9300a9a3d9c86dac6bbbb63071bdb7787afd2',

{pyhardlinkbackup-1.8.4 → pyhardlinkbackup-1.9.1}/PyHardLinkBackup/tests/test_project_setup.py

@@ -22,7 +22,7 @@ class ProjectSetupTestCase(TestCase):
         assert_is_file(cli_bin)
 
         output = subprocess.check_output([cli_bin, 'version'], text=True)
-        self.assertIn(f'PyHardLinkBackup v{__version__}', output)
+        self.assertIn(f'phlb v{__version__}', output)
 
         dev_cli_bin = PACKAGE_ROOT / 'dev-cli.py'
         assert_is_file(dev_cli_bin)

{pyhardlinkbackup-1.8.4 → pyhardlinkbackup-1.9.1}/PyHardLinkBackup/tests/test_readme.py

@@ -47,9 +47,6 @@ class ReadmeTestCase(BaseTestCase):
             ),
         )
 
-        # Installed via pipx is called 'phlb', not 'cli.py':
-        stdout = stdout.replace('./cli.py', 'phlb')
-
         assert_cli_help_in_readme(text_block=stdout, marker='main help')
 
     def test_backup_help(self):
@@ -63,9 +60,6 @@ class ReadmeTestCase(BaseTestCase):
             ),
         )
 
-        # Installed via pipx is called 'phlb', not 'cli.py':
-        stdout = stdout.replace('./cli.py', 'phlb')
-
         assert_cli_help_in_readme(text_block=stdout, marker='backup help')
 
     def test_dev_help(self):

{pyhardlinkbackup-1.8.4 → pyhardlinkbackup-1.9.1}/PyHardLinkBackup/utilities/file_hash_database.py

@@ -5,10 +5,6 @@ from pathlib import Path
 logger = logging.getLogger(__name__)
 
 
-class HashAlreadyExistsError(ValueError):
-    pass
-
-
 class FileHashDatabase:
     """DocWrite: README.md ## FileHashDatabase
     A simple "database" to store file content hash <-> relative path mappings.
@@ -54,12 +50,9 @@ class FileHashDatabase:
             return abs_file_path
 
     def __setitem__(self, hash: str, abs_file_path: Path):
+        """
+        Create or update the hash entry with the given absolute file path.
+        """
         hash_path = self._get_hash_path(hash)
         hash_path.parent.mkdir(parents=True, exist_ok=True)
-
-        # File should be found before and results in hardlink creation!
-        # So deny change of existing hashes:
-        if hash_path.exists():
-            raise HashAlreadyExistsError(f'Hash {hash} already exists in the database!')
-
         hash_path.write_text(str(abs_file_path.relative_to(self.backup_root)))

{pyhardlinkbackup-1.8.4 → pyhardlinkbackup-1.9.1}/PyHardLinkBackup/utilities/tests/test_file_hash_database.py

@@ -8,7 +8,7 @@ from bx_py_utils.test_utils.assertion import assert_text_equal
 from bx_py_utils.test_utils.log_utils import NoLogs
 from cli_base.cli_tools.test_utils.base_testcases import BaseTestCase
 
-from PyHardLinkBackup.utilities.file_hash_database import FileHashDatabase, HashAlreadyExistsError
+from PyHardLinkBackup.utilities.file_hash_database import FileHashDatabase
 from PyHardLinkBackup.utilities.filesystem import iter_scandir_files
 
 
@@ -127,10 +127,28 @@ class FileHashDatabaseTestCase(BaseTestCase):
                 )
 
             ########################################################################################
-            # Deny "overwrite" of existing hash:
+            # Update existing hash to point to a newer file:
 
-            with self.assertRaises(HashAlreadyExistsError):
-                hash_db['12abcd345678abcdef'] = 'foo/bar/baz'  # already exists!
+            """DocWrite: README.md ## FileHashDatabase
+            The entry for each hash is always updated to point to the most recently backed-up file.
+            This means you can safely delete old backups: the hash DB will still point to a valid
+            file in the most recent backup, so deduplication continues to work correctly.
+            """
+
+            file_c_path = backup_root_path / 'rel/path/to/file-C'
+            file_c_path.parent.mkdir(parents=True, exist_ok=True)
+            file_c_path.touch()
+
+            hash_db['12abcd345678abcdef'] = file_c_path
+            self.assertEqual(hash_db.get('12abcd345678abcdef'), file_c_path)
+            with self.assertLogs('PyHardLinkBackup', level=logging.DEBUG):
+                assert_hash_db_info(
+                    backup_root=hash_db.backup_root,
+                    expected="""
+                        12/34/12345678abcdef… -> rel/path/to/file-A
+                        12/ab/12abcd345678ab… -> rel/path/to/file-C
+                    """,
+                )
 
             ########################################################################################
             # Don't use stale entries pointing to missing files:
@@ -139,8 +157,10 @@ class FileHashDatabaseTestCase(BaseTestCase):
             file_a_path.unlink()
 
             """DocWrite: README.md ## FileHashDatabase - Missing hardlink target file
-            We check if the hardlink source file still exists. If not, we remove the hash entry from the database.
-            A warning is logged in this case."""
+            The `get()` method checks whether the referenced file still exists.
+            If not, the stale entry is removed and a warning is logged.
+            On the next backup run, the file is then copied fresh instead of hardlinked.
+            """
             with self.assertLogs(level=logging.WARNING) as logs:
                 self.assertIs(hash_db.get('12345678abcdef'), None)
             self.assertIn('Hash database entry found, but file does not exist', ''.join(logs.output))
@@ -148,6 +168,6 @@ class FileHashDatabaseTestCase(BaseTestCase):
                 assert_hash_db_info(
                     backup_root=hash_db.backup_root,
                     expected="""
-                        12/ab/12abcd345678ab… -> rel/path/to/file-B
+                        12/ab/12abcd345678ab… -> rel/path/to/file-C
                     """,
                 )

{pyhardlinkbackup-1.8.4 → pyhardlinkbackup-1.9.1}/README.md

@@ -282,6 +282,11 @@ Overview of main changes:
 
 [comment]: <> (✂✂✂ auto generated history start ✂✂✂)
 
+* [v1.9.1](https://github.com/jedie/PyHardLinkBackup/compare/v1.9.0...v1.9.1)
+  * 2026-04-14 - fix pipx installation and set requires-python to ">=3.13"
+* [v1.9.0](https://github.com/jedie/PyHardLinkBackup/compare/v1.8.4...v1.9.0)
+  * 2026-04-14 - Update existing links in has database
+  * 2026-04-14 - Update requirements
 * [v1.8.4](https://github.com/jedie/PyHardLinkBackup/compare/v1.8.3...v1.8.4)
   * 2026-04-09 - Update requirements
   * 2026-04-09 - Apply project updates
@@ -289,6 +294,9 @@ Overview of main changes:
 * [v1.8.3](https://github.com/jedie/PyHardLinkBackup/compare/v1.8.2...v1.8.3)
   * 2026-04-01 - Update requirements
   * 2026-03-30 - Bump pygments from 2.19.2 to 2.20.0
+
+<details><summary>Expand older history entries ...</summary>
+
 * [v1.8.2](https://github.com/jedie/PyHardLinkBackup/compare/v1.8.1...v1.8.2)
   * 2026-03-28 - Update tests
   * 2026-03-28 - fix code style
@@ -304,9 +312,6 @@ Overview of main changes:
   * 2026-01-22 - rebuid command: skip hashing same files by check the inode uniqueness
   * 2026-01-22 - Add "fs-info" in dev cli
   * 2026-01-22 - rebuild command: fix wrong progress bar
-
-<details><summary>Expand older history entries ...</summary>
-
 * [v1.8.0](https://github.com/jedie/PyHardLinkBackup/compare/v1.7.3...v1.8.0)
   * 2026-01-22 - Add optional "--name" to enforce a name for the backup sub directory
   * 2026-01-22 - Do not cross filesystem boundaries as default

{pyhardlinkbackup-1.8.4 → pyhardlinkbackup-1.9.1}/docs/README.md

@@ -36,13 +36,18 @@ Notes:
   * The "relative path" that will be stored is not validated, so it can be any string.
   * We don't "cache" anything in Memory, to avoid high memory consumption for large datasets.
 
+The entry for each hash is always updated to point to the most recently backed-up file.
+This means you can safely delete old backups: the hash DB will still point to a valid
+file in the most recent backup, so deduplication continues to work correctly.
+
 ## FileHashDatabase - Missing hardlink target file
 
-If a hardlink source from a old backup is missing, we cannot create a hardlink to it.
-But it still works to hardlink same files within the current backup.
+Deleting files from old backups is safe: the hash DB entry always points to the
+most recently backed-up file, so subsequent backups can still create hardlinks.
 
-We check if the hardlink source file still exists. If not, we remove the hash entry from the database.
-A warning is logged in this case.
+The `get()` method checks whether the referenced file still exists.
+If not, the stale entry is removed and a warning is logged.
+On the next backup run, the file is then copied fresh instead of hardlinked.
 
 ## FileSizeDatabase
 

{pyhardlinkbackup-1.8.4 → pyhardlinkbackup-1.9.1}/pyproject.toml

@@ -7,9 +7,9 @@ readme = "README.md"
 authors = [
     {name = 'Jens Diemer', email = 'PyHardLinkBackup@jensdiemer.de'}
 ]
-requires-python = ">=3.12"
+requires-python = ">=3.13"
 dependencies = [
-    "cli-base-utilities>=0.27.1",  # https://github.com/jedie/cli-base-utilities
+    "cli-base-utilities>=0.30.0",  # https://github.com/jedie/cli-base-utilities
     "bx_py_utils",  # https://github.com/boxine/bx_py_utils
     "tyro",  # https://github.com/brentyi/tyro
     "rich",  # https://github.com/Textualize/rich
@@ -43,8 +43,9 @@ exclude-newer = "1 week"
 [tool.uv.exclude-newer-package]
 # Exclude own packages from the "exclude-newer" rule and
 # add external packages temporarily to fix known issues or current CVEs
+uv = "2026-04-13T12:00:00Z"
+cli-base-utilities = "2026-04-13T12:00:00Z"
 cryptography = "2026-04-08T12:00:00Z"
-django = "2026-04-08T12:00:00Z"
 
 
 [tool.cli_base.pip_audit]