Version 5.3.1 (2024-03-??)

   Renamed check_urls.py to check_urls_db.py.

   Renamed check_url.py to check_urls.py.

   Stop splitting and un-splitting URLs. Pass bookmark.href as is.

Version 5.3.0 (2024-03-06)

   Added get_url.py: a script to get one file from an URL.

   Renamed set-URLs -> set-urls.

Version 5.2.5 (2024-03-05)

   Feat(Robots/bkmk_rrequests): Ignore all problems with certificates.

   Fix(Robots/bkmk_robot_base): Pass query part.

Version 5.2.4 (2024-03-04)

   Fix(Robots/bkmk_rrequests): No need to re-check error 404 via proxy.

Version 5.2.3 (2024-03-03)

   Feat(Robots/bkmk_rrequests): Report 40x and 50x errors.

   Fix HTML pasrer based on Bs4: Find "shortcut icon".

Version 5.2.2 (2024-03-03)

   Robots/bkmk_rrequests: Add request headers.

   Robots/bkmk_robot_base: Process "data:image/" icons.

Version 5.2.1 (2024-03-02)

   Speedup second access through proxy.

Version 5.2.0 (2024-03-02)

   For the robot based on requests allow to use a proxy.

Version 5.1.0 (2024-03-01)

   Robot based on requests.

Version 5.0.0 (2023-11-22)

   Python 3.

   Report redirects and set URLs.

   Delete URLs.

   Remove BeautifulSoup.py (use globally installed).

Version 4.6.0 (2014-07-06)

   Split simple robot: separate network operations and
   URL handling/HTML parsing.

   Change parse_html to parse strings, not files.

   Split parse_html/__init__.py into __main__.py.

   Adapt JSON storage to recent Mozilla export format.

   Add ChangeLog.

   Allow parameters in BKMK_* environment variables; for example,
   BKMK_ROBOT=forking:subproc=urllib or
   BKMK_STORAGE=json:filename=bookmarks_db.json.

   Pass subproc parameter to the subprocess to allow different robots.

   Add a new robot based on urllib2.

Version 4.5.6 (2014-01-14)

   Remove absolute directory ~/lib to make it portable.

Version 4.5.5 (2013-12-05)

   Parse <meta charset="...">.

Version 4.5.4 (2013-11-23)

   Published through git/gitweb.

Version 4.5.3 (2013-07-26)

   Minor tweak in Makefile.

   Switched to git.

Version 4.5.2 (2012-09-24)

   Removed svn:keywords.

   Handle redirects with codes 303 and 307.

   Fixed a bug in handling place: URIs (do not append '//').

Version 4.5.1 (2011-12-28).

   Read/write mozilla-specific date/time format in json storage.

Version 4.5.0 (2011-12-18).

   Encode international domain names with IDNA encoding.

   Adapted to different Mozilla 'place' URIs.

Version 4.4.0 (2011-01-07).

   Moved BeautifulSoup.py and subproc.py from Robots/ to the top-level
   directory.

   Moved parse_html.py and its submodules to a separate parse_html package.

   Added statistics code to parse_html, got a statistics on parser
   success/failrure rate, reordered parsers.

   Removed old cruft.

Version 4.3.1 (2011-01-03).

   Get favicon before HTML redirect (refresh).

Version 4.3.0 (2011-01-01).

   Robots no longer have one global temporary file - there are at least two
   (html and favicon), and in the future there will be more for asynchronous
   robot(s) that would test many URLs in parallel.

Version 4.2.2.

    Added HTML Parser based on lxml.

Version 4.2.1 (2010-08-12).

   Added HTML Parser based on html5 library.

Version 4.2.0 (2010-08-11).

   New storage: json; it allows to load and store Mozilla (Firefox) backup
   files.

Version 4.1.2

   Process http error 307 as a temporary redirect.

Version 4.1.1 (2008-03-10)

   Catch and report all errors.

   Consider application/xhtml+xml as HTML.

   Better handling of exceptions while looking up the icon.

   Recode HTML entities.

   Always use utf-8 as the default encoding.

Version 4.1.0 (2008-01-14)

   Parser for HTML based on BeautifulSoup.

   Changed User-agent header: I saw a number of sites that forbid
   "Mozilla compatible" browsers. Added a number of fake headers to pretend
   this is a real web-browser - there are still stupid sites
   that are trying to protect themselves from robots by analyzing headers.

   Handle redirects while looking for the icon.

   Handle float timeouts in HTML redirects.

   Minimal required version of Python is 2.5 now.

Version 4.0.0 (2007-10-20)

   Extended support for Mozilla: charset and icon in bookmarks.
   Use the charset to add Accept-Charset header.
   Retrieve favicon.ico (or whatever <link> points to) and store it.

   The project celebrates 10th anniversary!

Version 3.4.1 (2005-01-29)

   Updated to Python 2.4. Switched from CVS to Subversion.

Version 3.4.0 (2004-09-23)

   Extended support for Mozilla: keywords in bookmarks.
   Updated to m_lib version 1.2.

Version 3.3.2

   parse_html.py can now recode unicode entities in titles.

Version 3.3.0

   Required Python 2.2.

   HTML parser. If the protocol is HTTP, and there is Content-Type header, and
   content type is text/html, the object is parsed to extract its title; if
   the Content-Type header has charset, or if the HTML has <META> with
   charset, the title is converted from the given charset to the default
   charset. The <HEADER> is also parsed to extract <META> tag with redirect,
   if any.

Version 3.0

   Complete rewrite from scratch. Created mechanism for pluggable storage
   managers, writers (DB dumpers/exporters) and robots.