Skip to content

4.1.2

Latest

Choose a tag to compare

@facelessuser facelessuser released this 18 Nov 17:20

4.1.2

  • FIX: Fix :lang() taking time quadratic in the number of top level elements of a fragment that have an html
    or head child, such as many documents concatenated and parsed with html.parser, as the content language
    found in the header of each one was cached in a list that was searched for every new top level element.
  • FIX: Fix :lang() taking time quadratic in the size of the document when many elements share a long language
    tag, such as one inherited from <html lang>, as the tag was checked and matched again for every element.
  • FIX: Fix :-soup-contains() taking time quadratic in the depth of the document, such as a long list of
    unclosed <li> parsed with html.parser, as the text of every element, which includes that of all its
    descendants, was gathered and searched again for every element.
  • FIX: NOCACHE now turns off the caches of the content language in a document's header, of the forms of a
    document, and of the selected options of a select, so :lang(), :default, :indeterminate and :checked
    see changes made to the tree during a lazy iteration such as iselect().
  • FIX: Slimmed relation caches so they now take roughly half the memory.
  • FIX: Fix nth selectors with very large An+B values raising an OverflowError when matching, and
    matching the wrong elements when a value is larger than 2^53, as the position of an element was checked with a
    float division. A value with more digits than Python will convert to an integer now raises a
    SelectorSyntaxError instead of a ValueError.
  • FIX: filter() given a list of tags now matches the tags under the same top level element together, sharing
    the caches that keep matching linear, instead of matching each tag separately, which took time quadratic in the
    depth of a nested document for any selector. A selector that uses :scope or & still matches each tag
    separately, as each tag is its own scope.
  • FIX: The ignore option now disallows a pseudo-class however its name is written, as names were compared
    without resolving CSS escapes, so a name written with an escape, such as :h\61 s(), was not disallowed. Document
    that each alias of a pseudo-class, such as :contains() for :-soup-contains(), must be listed as well.