Changelog

All notable changes will be documented in this file

Release 0.15.2 2026-06-08

• Improved internal resilience, build and packaging workflows, documentation, and test reliability

Release 0.15.1 2026-06-07

Fixed

• Fixed a terminal output issue where rare file names containing control characters could make later scan messages disappear from the screen. Such bytes are now printed as readable \xNN escapes, while normal Unicode file names remain readable

Release 0.15.0 2026-06-06

Fixed

• File system scans are more resilient: precizer now continues working when an access check fails for one file or directory, treating that path as inaccessible instead of stopping the whole scan

Release 0.14.0 2026-06-04

Changed

• --lock-checksum now protects locked records from being silently removed when files disappear or become unreadable
• Locked checksum protection now takes priority over --ignore, --include, --db-drop-ignored, and --db-drop-inaccessible
• --rehash-locked now verifies locked files even when they also match --ignore
• Inaccessible files are handled more safely and are kept in the database unless --db-drop-inaccessible is explicitly used
• Path and database handling is more reliable, with safer SQL parameter binding and transactional root path saving
• PCRE2 filters can now use JIT compilation for faster matching on large file trees
• Most file and path operations in the main application tests now use internal C helpers instead of external shell commands
• Build and CI support is broader, including dynamic tests, Docker matrix checks, Clang selection, and improved cleanup

Documentation

• Documentation now explains locked checksum behavior, deep locked-file verification, and inaccessible-file cleanup more clearly

Release 0.13.0 2026-05-10

Removed

• CMocka is no longer required to build precizer from source or to run the test suite

Release 0.12.0 2026-03-19

Added

--compare now supports the --ignore and --include path filters, so database comparisons can be limited to the paths that matter. Both options can be repeated, and --include takes precedence when a path matches both filters

Documentation

Clarified in the README and CLI help that --compare reports, summaries, and equality messages are evaluated against the filtered comparison scope

Release 0.11.0 2026-03-16

Changed

• Refined --silent behavior in --compare mode: compare results remain visible
• Renamed --compare-filter values first-source-only and second-source-only to first-source and second-source to make combined filter usage clearer. The old values are no longer accepted

Release 0.10.0 2026-03-14

Changed

• All releases are now signed and will be distributed with digital signatures

Release 0.9.0 2026-03-12

Changed

• Improved --compare output so differences between databases are easier to understand, including clearer diagnostics when --compare-filter is used
• Messages related to files matched by --lock-checksum are now repeated at the end of the run, making important scan results easier to notice in long output
• Updated the bundled SQLite library to a newer upstream version

Fixed

• Improved compatibility when building precizer on macOS

v0.8.0 2026-03-05

Changed

• precizer is now distributed under the GPLv3 license

v0.7.0 2026-02-28

This version mainly improves test reliability (including complex --ignore/--include, interruption, and file metadata change scenarios) and includes minor output and documentation polish without noticeable user-facing CLI behavior changes

v0.6.0 2026-02-10

Added

• Added reporting for filesystem traversal time and the average per second throughput for reading files and calculating checksums.
• Added --dry-run=with-checksums mode for --dry-run. In this mode, files are read and checksums are calculated without writing to the database.
• Added --compare-filter for --compare output scoping (checksum-mismatch, first-source-only, second-source-only) with support for repeated combined filters.

Changed

• Disk usage calculation now uses allocated block count (st_blocks) instead of logical file length (st_size). This reflects actual on‑disk space usage rather than apparent size.
• Database format upgraded to version 4 to support the new accounting model; the migration is fully automatic and transparent for users.

Documentation

• Added a TROUBLESHOOTING section to README.md with guidance for diagnosing slow filesystem walks, checksum computation, and SQLite .db write performance.

v0.5.0 2026-01-29 Pull Request #45

Added

• New CLI option --quiet-ignored to suppress log lines for files filtered out via --ignore/--include.
• New CLI option to drop database records for inaccessible files during --update:
  • Canonical: --db-drop-inaccessible
  • Backward-compatible alias: --drop-inaccessible
• Introduced/expanded CMocka-based testing infrastructure.

Changed

• When --progress is enabled, critical warnings/errors collected during filesystem traversal are saved and printed as a single block near the end of the run (to avoid losing important messages in noisy output).
• --dry-run can now work with read-only databases by switching away from ATTACH and using temporary TEMP tables instead.
• Directory traversal approach was reworked.
• Database-related option naming was standardized to the --db-* style:
  • Canonical: --db-drop-ignored
  • Backward-compatible alias: --db-clean-ignored

Fixed

• File read errors during SHA512 calculation are now treated as warnings rather than hard failures (improves robustness when some files can’t be read).
• Miscellaneous output / logging fixes.

Tests

• Improved tests around --rehash-locked, --lock-checksum, and --watch-timestamps, plus internal test harness adjustments.

Documentation

• Added a stable permalink to README for downloading the latest release