Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

HTML5 documents should not require namespaces in CSS selector queries #2403

Merged
merged 13 commits into from Jan 4, 2022
Merged

HTML5 documents should not require namespaces in CSS selector queries #2403

merged 13 commits into from Jan 4, 2022

Commits on Jan 3, 2022

  1. test: comment on custom xpath functions

    @flavorjones
    flavorjones committed Jan 3, 2022
    Configuration menu
    Copy the full SHA
    685403a View commit details
    Browse the repository at this point in the history

  2. CSS::Node#to_xpath arguments are now required 

    This makes it explicit that visitor must be injected.

Although this is technically a breaking change, CSS::Node is
essentially an internal API and I would be extremely surprised if it
was being used directly by anyone.

This commit also ensures that CSS::Node, CSS::Parser, and
CSS::Tokenizer are omitted from API documentation (to reflect their
status as an internal-only and unstable API).
    @flavorjones
    flavorjones committed Jan 3, 2022
    Configuration menu
    Copy the full SHA
    b24efcc View commit details
    Browse the repository at this point in the history

  3. prefactor: allow injection of an XPathVisitor to CSS.xpath_for 

    Note that the method signature of internal-only method
CSS::Parser.xpath_for has changed to prefer required parameters rather
than an options hash with assumed defaults.

Also note that I've introduced some XPath constants for XPath query
prefixes, which we should start using.
    @flavorjones
    flavorjones committed Jan 3, 2022
    Configuration menu
    Copy the full SHA
    4b6bfa9 View commit details
    Browse the repository at this point in the history

  4. fix: CSS cache includes XPathVisitor configuration 

    which prevents incorrect cache results being returned in cases where
different visitor configurations are used.

We also deprecate XPathVisitorAlwaysUseBuiltins and
XPathVisitorOptimallyUseBuiltins in favor of a simple XPathVisitor
with constructor arguments.
    @flavorjones
    flavorjones committed Jan 3, 2022
    Configuration menu
    Copy the full SHA
    b3ffffb View commit details
    Browse the repository at this point in the history

  5. prefactor: CSS parser distinguishes attr names from element names 

    which will allow us to more easily special-case element names in an
upcoming commit.
    @flavorjones
    flavorjones committed Jan 3, 2022
    Configuration menu
    Copy the full SHA
    e005149 View commit details
    Browse the repository at this point in the history

  6. introduce enum CSS::XPathVisitor::BuiltinsConfig 

    and refactor the XPathVisitor#css_class to not rely on aliasing during
object initialization, which feels funny to me
    @flavorjones
    flavorjones committed Jan 3, 2022
    Configuration menu
    Copy the full SHA
    f935e5a View commit details
    Browse the repository at this point in the history

  7. test: CSS integration tests run on XML, HTML4, and HTML5 documents 

    we're about to introduce variations in how CSS selectors are
translated into XPath and I want to ensure we have adequate test coverage.
    @flavorjones
    flavorjones committed Jan 3, 2022
    Configuration menu
    Copy the full SHA
    32f9ab0 View commit details
    Browse the repository at this point in the history

  8. fix: CSS searches in HTML5 documents should not require namespaces 

    In HTML5, foreign elements have namespaces; but those namespaces
should not be considered for the purposes of CSS searches.

Unfortunately, this implementation's use of local-name() is ~10x
slower than using the normal inline element name matching, which
subsequent commits will explore.
    @flavorjones
    flavorjones committed Jan 3, 2022
    Configuration menu
    Copy the full SHA
    85ac956 View commit details
    Browse the repository at this point in the history

  9. doc: document XPathVisitor 

    Also tweak :nodoc: directives to avoid accidentally excluding
important modules (like Nokogiri::CSS)

and avoid including code that's not usually relevant to the user
API (like Nokogiri::XML:PP, Nokogiri::CSS:Tokenizer, and
Nokogiri::HTML4::Document::EncodingReader)
    @flavorjones
    flavorjones committed Jan 3, 2022
    Configuration menu
    Copy the full SHA
    d030381 View commit details
    Browse the repository at this point in the history

  10. doc: document XML::Notation 

    and other small changes to doc strings
    @flavorjones
    flavorjones committed Jan 3, 2022
    Configuration menu
    Copy the full SHA
    42068de View commit details
    Browse the repository at this point in the history

  11. feat: introduce builtin xpath function local-name-is

    @flavorjones
    flavorjones committed Jan 3, 2022
    Configuration menu
    Copy the full SHA
    2096ef9 View commit details
    Browse the repository at this point in the history

  12. dev: introduce some helpers for introspecting on libxml2 patches

    @flavorjones
    flavorjones committed Jan 3, 2022
    Configuration menu
    Copy the full SHA
    0fd4de4 View commit details
    Browse the repository at this point in the history

  13. feat: support wildcard namespaces in xpath queries 

    This is almost as fast as a standard child-axis search, and much
faster than the builtin or using local-name():

  //span
       18.923k (± 8.9%) i/s -     93.906k in   5.010792s
  //*[local-name()='span']
        1.849k (± 2.8%) i/s -      9.261k in   5.011560s
  //*[nokogiri-builtin:local-name-is('span')]
        3.191k (± 2.4%) i/s -     16.150k in   5.064798s
  //*:span
       18.016k (± 4.6%) i/s -     89.900k in   5.003444s

  Comparison:
  //span:
      18922.5 i/s
  //*:span:
      18016.5 i/s - same-ish: difference falls within error
  //*[nokogiri-builtin:local-name-is('span')]:
       3190.6 i/s - 5.93x  (± 0.00) slower
  //*[local-name()='span']:
       1849.4 i/s - 10.23x  (± 0.00) slower
    @flavorjones
    flavorjones committed Jan 3, 2022
    Configuration menu
    Copy the full SHA
    d1a710e View commit details
    Browse the repository at this point in the history