v1.0.2

What's Changed

Improvements

• Improved per-sitemap output filename readability so exported .txt files are easier to identify against their source sitemap URL or local sitemap file
• Remote sitemap outputs now include cleaner host/path/query hints in the readable filename prefix
• Local sitemap outputs now include parent-directory and file-stem hints in the readable filename prefix
• Kept the trailing short hash derived from the full original source string, so collision safety is preserved

Docs

• Updated README output filename examples and descriptions to reflect the more identifiable naming format

Notes

• This release changes the human-readable prefix of per-sitemap output filenames
• The hash basis is unchanged, but any downstream scripts that match the old filename prefix format may need to be updated

Full Changelog: v1.0.1...v1.0.2

v1.0.1

What's Changed

Bug Fixes

• Fixed per-sitemap filename collisions for child sitemap URLs sharing the same host/path but differing by query string; output filenames now use a sanitized readable base plus a short hash of the full source URL
• Ensured each per-sitemap file contains only the URLs extracted from that specific sitemap source
• Corrected the README clone URL to phase3dev/sitemap-extract
• Fixed local sitemap handling so directory-based inputs and nested local sitemap references resolve correctly
• Replaced the bare except: in get_current_ip() with except Exception: so KeyboardInterrupt and SystemExit are not swallowed

Improvements

• Added all_extracted_urls.txt, always written at the end of a run from the deduplicated union of all extracted page URLs
• Deduplicated per-sitemap URL files before writing
• Tightened directory scanning so --directory matches only .xml and .xml.gz, not suffix variants like .xml.bak
• Normalized save_dir once before processor construction so the processor always receives a canonical path
• Made --stealth behavior consistent across runtime, CLI help, and README by forcing max_workers=1 instead of warning only
• Replaced non-interruptible retry sleeps with a shared interruptible sleep helper so Ctrl+C responds promptly during retry backoff and stagger delays
• Preserved global pacing semantics under multithreading by locking the shared request clock around delay calculation and sleep
• Removed request-scoped proxy and user-agent state from shared instance fields, keeping them local per request to eliminate cross-thread races

Tests

• Added minimal stdlib regression tests covering interruptible sleep, proxy/IP formatting, interrupt propagation, concurrent stats/failure tracking, and a threaded local sitemap run

Docs & Maintenance

• Documented supported Python as 3.9+ in the README
• Added a minimal SECURITY.md
• Removed unused lxml from requirements.txt

Full Changelog: v1.0.0...v1.0.1

v1.0.0

Initial Release