cataract 0.2.5 → 0.4.0

data/lib/cataract/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Cataract
-  VERSION = '0.2.5'
+  VERSION = '0.4.0'
 end

data/lib/cataract.rb

@@ -1,34 +1,18 @@
 # frozen_string_literal: true
 
-require_relative 'cataract/version'
-require_relative 'cataract/error'
-require_relative 'cataract/constants'
-
-# Load struct definitions first (before C extension or pure Ruby)
-require_relative 'cataract/declaration'
-require_relative 'cataract/rule'
-require_relative 'cataract/at_rule'
-require_relative 'cataract/media_query'
-require_relative 'cataract/import_statement'
-
-# Load pure Ruby or C extension based on ENV var
-if %w[1 true].include?(ENV.fetch('CATARACT_PURE', nil)) || RUBY_ENGINE == 'jruby'
-  require_relative 'cataract/pure'
-else
-  require_relative 'cataract/native_extension'
-end
-
-# Load supporting Ruby files (used by both implementations)
-require_relative 'cataract/stylesheet_scope'
-require_relative 'cataract/stylesheet'
-require_relative 'cataract/declarations'
-require_relative 'cataract/import_resolver'
-
 # Cataract is a high-performance CSS parser written in C with a Ruby interface.
 #
 # It provides fast CSS parsing, rule querying, cascade merging, and serialization.
 # Designed for performance-critical applications that need to process large amounts of CSS.
 #
+# Pure Ruby and native each live entirely under their own Backends::Pure /
+# Backends::Native namespace (no shared method/constant is ever defined by
+# a backend directly on Cataract or on a shared value type), so both can be
+# required in the same process - e.g. to compare their output directly.
+# Backends.active is the one this process picked by default (below); it's
+# what shared value types (Stylesheet, Declarations, Rule) fall back to
+# when they aren't otherwise told which backend produced them.
+#
 # @example Basic usage
 #   require 'cataract'
 #
@@ -44,6 +28,33 @@ require_relative 'cataract/import_resolver'
 # @see Stylesheet Main class for working with parsed CSS
 # @see Rule Represents individual CSS rules
 module Cataract
+  module Backends
+    class << self
+      attr_accessor :active
+    end
+  end
+end
+
+if %w[1 true].include?(ENV.fetch('CATARACT_PURE', nil)) || RUBY_ENGINE == 'jruby'
+  require_relative 'cataract/pure'
+  Cataract::Backends.active = Cataract::Backends::Pure
+else
+  require_relative 'cataract/native'
+  Cataract::Backends.active = Cataract::Backends::Native
+end
+
+module Cataract
+  # Mirror the active backend's identity constants at the top level.
+  #
+  # Native is a Module (IMPLEMENTATION/COMPILE_FLAGS live on it directly);
+  # Pure is a frozen instance (they live on its class instead), so reach
+  # through .class only when active isn't already a Module itself.
+  backend_const_holder = Backends.active.is_a?(Module) ? Backends.active : Backends.active.class
+  IMPLEMENTATION = backend_const_holder::IMPLEMENTATION
+  COMPILE_FLAGS = backend_const_holder::COMPILE_FLAGS
+  NATIVE_EXTENSION_LOADED = true if defined?(Backends::Native) && Backends.active == Backends::Native
+  PURE_RUBY_LOADED = true if defined?(Backends::Pure) && Backends.active == Backends::Pure
+
   class << self
     # Parse a CSS string into a Stylesheet object.
     #
@@ -75,10 +86,8 @@ module Cataract
     #
     # @see Stylesheet#parse
     # @see Stylesheet.parse
-    unless method_defined?(:parse_css)
-      def parse_css(css, **options)
-        Stylesheet.parse(css, **options)
-      end
+    def parse_css(css, **options)
+      Stylesheet.parse(css, **options)
     end
 
     # Flatten CSS rules according to CSS cascade rules.
@@ -113,9 +122,11 @@ module Cataract
     #
     # @note This is a module-level convenience method. The same functionality is available
     #   as an instance method: `stylesheet.flatten`
-    # @note Implemented in C (see ext/cataract/flatten.c)
     #
     # @see Stylesheet#flatten
-    # Cataract.flatten is defined in C via rb_define_module_function
+    def flatten(stylesheet_or_css)
+      stylesheet_or_css = Stylesheet.parse(stylesheet_or_css) if stylesheet_or_css.is_a?(String)
+      stylesheet_or_css.flatten
+    end
   end
 end

data/lib/tasks/profile.rake

@@ -35,7 +35,7 @@ namespace :profile do
     puts "CSS size: #{css_content.bytesize} bytes"
     puts
 
-    require_relative '../../lib/cataract/pure'
+    require_relative '../../lib/cataract'
     require 'json'
 
     # Use higher sampling rate (interval in microseconds, default is 1000)
@@ -74,6 +74,9 @@ namespace :profile do
       abort('stackprof gem not found. Install with: gem install stackprof')
     end
 
+    # Ensure we're using pure Ruby implementation
+    ENV['CATARACT_PURE'] = '1'
+
     fixture_path = File.expand_path('../../test/fixtures/bootstrap.css', __dir__)
     unless File.exist?(fixture_path)
       abort("Fixture not found: #{fixture_path}")
@@ -94,7 +97,7 @@ namespace :profile do
     puts "CSS size: #{css_content.bytesize} bytes"
     puts
 
-    require_relative '../../lib/cataract/pure'
+    require_relative '../../lib/cataract'
     require 'json'
 
     # Parse once outside profiling to get stylesheet
@@ -162,7 +165,7 @@ namespace :profile do
     puts "CSS size: #{css_content.bytesize} bytes"
     puts
 
-    require_relative '../../lib/cataract/pure'
+    require_relative '../../lib/cataract'
     require 'json'
 
     # Parse once outside profiling to get stylesheet

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: cataract
 version: !ruby/object:Gem::Version
-  version: 0.2.5
+  version: 0.4.0
 platform: ruby
 authors:
 - James Cook
@@ -56,7 +56,6 @@ files:
 - ext/cataract/css_parser.c
 - ext/cataract/extconf.rb
 - ext/cataract/flatten.c
-- ext/cataract/import_scanner.c
 - ext/cataract/shorthand_expander.c
 - ext/cataract/specificity.c
 - ext/cataract/value_splitter.c
@@ -67,16 +66,6 @@ files:
 - ext/cataract_color/color_conversion_named.c
 - ext/cataract_color/color_conversion_oklab.c
 - ext/cataract_color/extconf.rb
-- ext/cataract_old/cataract.c
-- ext/cataract_old/cataract.h
-- ext/cataract_old/css_parser.c
-- ext/cataract_old/extconf.rb
-- ext/cataract_old/import_scanner.c
-- ext/cataract_old/merge.c
-- ext/cataract_old/shorthand_expander.c
-- ext/cataract_old/specificity.c
-- ext/cataract_old/stylesheet.c
-- ext/cataract_old/value_splitter.c
 - lib/cataract.rb
 - lib/cataract/at_rule.rb
 - lib/cataract/color_conversion.rb
@@ -87,10 +76,11 @@ files:
 - lib/cataract/import_resolver.rb
 - lib/cataract/import_statement.rb
 - lib/cataract/media_query.rb
+- lib/cataract/native.rb
 - lib/cataract/pure.rb
 - lib/cataract/pure/byte_constants.rb
+- lib/cataract/pure/declarations.rb
 - lib/cataract/pure/flatten.rb
-- lib/cataract/pure/helpers.rb
 - lib/cataract/pure/parser.rb
 - lib/cataract/pure/serializer.rb
 - lib/cataract/pure/specificity.rb
@@ -122,7 +112,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.6.9
+rubygems_version: 4.0.10
 specification_version: 4
 summary: High-performance CSS parser with C extensions
 test_files: []

data/ext/cataract/import_scanner.c

@@ -1,174 +0,0 @@
-#include <ruby.h>
-#include <ctype.h>
-#include <string.h>
-#include "cataract.h"
-
-/*
- * Scan CSS for @import statements
- *
- * Matches patterns:
- *   @import url("path");
- *   @import url('path');
- *   @import "path";
- *   @import 'path';
- *   @import url("path") print;  (with media query)
- *
- * Returns array of hashes: [{url: "...", media: "...", full_match: "..."}]
- */
-VALUE extract_imports(VALUE self, VALUE css_string) {
-    Check_Type(css_string, T_STRING);
-
-    const char *css = RSTRING_PTR(css_string);
-    long css_len = RSTRING_LEN(css_string);
-
-    VALUE imports = rb_ary_new();
-
-    const char *p = css;
-    const char *end = css + css_len;
-
-    while (p < end) {
-        // Skip whitespace and comments
-        while (p < end) {
-            if (IS_WHITESPACE(*p)) {
-                p++;
-            } else if (p + 2 <= end && p[0] == '/' && p[1] == '*') {
-                // Skip /* */ comment
-                p += 2;
-                while (p + 1 < end && !(p[0] == '*' && p[1] == '/')) {
-                    p++;
-                }
-                if (p + 1 < end) p += 2; // Skip */
-            } else {
-                break;
-            }
-        }
-
-        // Check for @import
-        if (p + 7 <= end && strncasecmp(p, "@import", 7) == 0) {
-            const char *import_start = p;
-            p += 7;
-
-            // Skip whitespace after @import
-            while (p < end && IS_WHITESPACE(*p)) p++;
-
-            // Check for optional url(
-            int has_url_function = 0;
-            if (p + 4 <= end && strncasecmp(p, "url(", 4) == 0) {
-                has_url_function = 1;
-                p += 4;
-                while (p < end && IS_WHITESPACE(*p)) p++;
-            }
-
-            // Find opening quote
-            if (p >= end || (*p != '"' && *p != '\'')) {
-                // Invalid @import, skip to next semicolon
-                while (p < end && *p != ';') p++;
-                if (p < end) p++; // Skip semicolon
-                continue;
-            }
-
-            char quote_char = *p;
-            p++; // Skip opening quote
-
-            const char *url_start = p;
-
-            // Find closing quote (handle escaped quotes)
-            while (p < end && *p != quote_char) {
-                if (*p == '\\' && p + 1 < end) {
-                    p += 2; // Skip escaped character
-                } else {
-                    p++;
-                }
-            }
-
-            if (p >= end) {
-                // Unterminated string
-                break;
-            }
-
-            const char *url_end = p;
-            p++; // Skip closing quote
-
-            // Skip closing paren if we had url(
-            if (has_url_function) {
-                while (p < end && IS_WHITESPACE(*p)) p++;
-                if (p < end && *p == ')') {
-                    p++;
-                }
-            }
-
-            // Skip whitespace before optional media query or semicolon
-            while (p < end && IS_WHITESPACE(*p)) p++;
-
-            // Check for optional media query (everything until semicolon)
-            const char *media_start = NULL;
-            const char *media_end = NULL;
-
-            if (p < end && *p != ';') {
-                media_start = p;
-
-                // Find semicolon
-                while (p < end && *p != ';') p++;
-
-                media_end = p;
-
-                // Trim trailing whitespace from media query
-                while (media_end > media_start && IS_WHITESPACE(*(media_end - 1))) {
-                    media_end--;
-                }
-            }
-
-            // Skip semicolon
-            if (p < end && *p == ';') p++;
-
-            const char *import_end = p;
-
-            // Build result hash
-            VALUE import_hash = rb_hash_new();
-
-            // Extract URL
-            VALUE url = rb_str_new(url_start, url_end - url_start);
-            rb_hash_aset(import_hash, ID2SYM(rb_intern("url")), url);
-
-            // Extract media query (or nil)
-            VALUE media = Qnil;
-            if (media_start && media_end > media_start) {
-                media = rb_str_new(media_start, media_end - media_start);
-            }
-            rb_hash_aset(import_hash, ID2SYM(rb_intern("media")), media);
-
-            // Extract full match
-            VALUE full_match = rb_str_new(import_start, import_end - import_start);
-            rb_hash_aset(import_hash, ID2SYM(rb_intern("full_match")), full_match);
-
-            rb_ary_push(imports, import_hash);
-
-            RB_GC_GUARD(url);
-            RB_GC_GUARD(media);
-            RB_GC_GUARD(full_match);
-            RB_GC_GUARD(import_hash);
-        } else {
-            // Not an @import, skip to next line or rule
-            // Once we hit a non-@import rule (except @charset), stop looking
-            // Per CSS spec, @import must be at the top
-
-            // Skip @charset if present
-            if (p + 8 <= end && strncasecmp(p, "@charset", 8) == 0) {
-                // Skip to semicolon
-                while (p < end && *p != ';') p++;
-                if (p < end) p++; // Skip semicolon
-                continue;
-            }
-
-            // If we hit any other content, stop scanning for imports
-            if (p < end && !IS_WHITESPACE(*p)) {
-                break;
-            }
-
-            p++;
-        }
-    }
-
-    RB_GC_GUARD(imports);
-    return imports;
-}