Scrapbook X 1.13.0b14
=====================

Compatibility notes
-------------------

 * Some user preferences are modified.  Those set in previous versions
   would be unavailable and need a reset.


Changes since v1.13.0b13
------------------------

Bug fixes:

 * An error can occur when capturing a special page such as an "about:"
   page.

 * The favicon of a page captured by link or deep capture can sometimes be
   different to that captured from direct browse.

 * Elements with a special protocol (such as chrome, moz-icon, or about)
   that are set to be not captured are still captured.

 * "Capture As" dialog becomes collapsed when toggling the script option.

 * "Loading charset" option only works for one deep-captured page.

 * Invoking "Abort" in the Capture dialog during a countdown can lead to an
   incomplete abort and a further error.

 * Invoking "Abort" in the Capture dialog when an item is being captured
   causes the wrong item be marked.

 * Adding a duplicated filter in the Capture dialog is not avoided correctly
   in certain cases.

 * Modifying a non-first filter in the Capture dialog triggers an error.

 * A deep capture can freeze up in certain cases if "Capture Selection"
   is run and there is a link referencing the main page.

 * An embeded media may not be included correctly if a page put into
   the Combine Wizard contains a <base> element.

 * Copy Item may not work correctly on some platforms.

Functionality changes:

 * Added capture options to determine whether to record some detail
   during a capture.

   "capture.default.recordSkippedUrl", "capture.default.recordRemovedAttr",
   and "capture.default.recordInDepthLink" user preferences are now
   available to determine whether to record the skipped URLs, the removed
   attributes, and mark deep capture page links during a capture.  They are
   disabled by default since such information are more for debugging and
   are less likely needed for general users.

 * Removed the "Reorganize Styles" capture option.

   All stylesheets are now saved in their original <link> or <style> tags
   after tidied up.  We made this change since the previous behavior of
   merging all styles into a single .css file fails to preserve the
   original data structure and causes some unresolvable issues, especially
   when javascript is involved.

   The corresponding user preference "capture.default.rewriteStyles"
   is removed.  A similar user preference "capture.default.tidyCss" which
   defaults to true and can be used to determine whether to remove
   non-effective styles is added.

 * Removed the "Keep linking to the source for not-captured objects"
   capture option.

   This option could cause some unresolvable issues and thus we decided
   to remove it.  We'll consider implementing this feature in another way
   in the future.

 * <noscript> tags are now always saved when capturing.

 * Every non-captured <link> or <script> tag now writes a non-functional
   URI containing the source rather then be removed when capturing.

 * A capture now removes the integrity check related attributes from
   web page elements (especially stylesheets) so that they won't be
   blocked due to a modification made by ScrapBook X or a context
   change from online to offline.

   A new user preference "capture.default.removeIntegrity" can be used
   to disable this behavior is added.

 * "Download linked files" option now uses the previously used value
   in a Capture As.

 * Moved "Timeout" and "Loading charset" options to the capture dialog.

   These options only work before a page is loaded. Previously they
   have been placed in the Capture As dialog, which show up after the
   page is loaded, causing them be ignored when capturing the main page.

 * The script that saves the status of a canvas now uses double quotes
   instead of single quotes.

 * Combine Wizard now doesn't produce the redundant linefeeds between
   the wrapping div elements of each item combined.

 * Combine and Internalize now always preserve all images, media, fonts,
   frames, styles, and scripts no matter how the capture detail options
   are set.

User interface changes:

 * The default document title is now selectable for a "Capture As"
   via dragging.

 * "Pause" in Capture dialog can now be invoked at any time during
   the batch capture.

 * An invalid RegExp is now forbidden in Capture dialog to prevent
   an error on filtering URLs.

 * Capture dialog now shows real-time status for each entry.

 * Capture dialog now automatically detects an entry with circular
   redirection and fails out on it.

 * Double left click in Manage window now opens the content of a
   non-folder-type folder.

 * Fixed the inconsistent zh-tw localization about RegExp.

Also includes lots of internal code optimization and clean-up.
