cataract 0.2.4 → 0.3.0

data/lib/cataract/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Cataract
-  VERSION = '0.2.4'
+  VERSION = '0.3.0'
 end

data/lib/cataract.rb

@@ -1,6 +1,7 @@
 # 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)

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: cataract
 version: !ruby/object:Gem::Version
-  version: 0.2.4
+  version: 0.3.0
 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
@@ -67,22 +66,13 @@ 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
 - lib/cataract/constants.rb
 - lib/cataract/declaration.rb
 - lib/cataract/declarations.rb
+- lib/cataract/error.rb
 - lib/cataract/import_resolver.rb
 - lib/cataract/import_statement.rb
 - lib/cataract/media_query.rb
@@ -121,7 +111,7 @@ 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;
-}

data/ext/cataract_old/cataract.c

@@ -1,393 +0,0 @@
-#include <ruby.h>
-#include <stdio.h>
-#include "cataract.h"
-
-// Global struct class definitions (declared extern in cataract.h)
-VALUE cDeclaration;
-VALUE cRule;
-
-// Error class definitions (declared extern in cataract.h)
-VALUE eCataractError;
-VALUE eParseError;
-VALUE eDepthError;
-VALUE eSizeError;
-
-// ============================================================================
-// Ruby Bindings and Public API
-// ============================================================================
-
-static VALUE parse_css_internal(VALUE self, VALUE css_string, int depth) {
-    // Check recursion depth to prevent stack overflow and memory exhaustion
-    if (depth > MAX_PARSE_DEPTH) {
-        rb_raise(eDepthError,
-                 "CSS nesting too deep: exceeded maximum depth of %d",
-                 MAX_PARSE_DEPTH);
-    }
-
-    Check_Type(css_string, T_STRING);
-
-    // Extract @charset if present (must be at very start per W3C spec)
-    // Handled separately because @charset must be at the absolute start
-    // and can be processed with simple string operations
-    VALUE charset = Qnil;
-    const char *css_start = RSTRING_PTR(css_string);
-    long css_len = RSTRING_LEN(css_string);
-
-    // Check for @charset at very start: @charset "UTF-8";
-    // Per spec: exact syntax with double quotes required
-    if (css_len > 10 && strncmp(css_start, "@charset ", 9) == 0) {
-        // Find opening quote
-        char *quote_start = strchr(css_start + 9, '"');
-        if (quote_start != NULL) {
-            // Find closing quote and semicolon
-            char *quote_end = strchr(quote_start + 1, '"');
-            if (quote_end != NULL) {
-                char *semicolon = quote_end + 1;
-                // Skip whitespace between quote and semicolon
-                while (semicolon < css_start + css_len && IS_WHITESPACE(*semicolon)) {
-                    semicolon++;
-                }
-                if (semicolon < css_start + css_len && *semicolon == ';') {
-                    // Valid @charset rule found
-                    charset = rb_str_new(quote_start + 1, quote_end - quote_start - 1);
-                    DEBUG_PRINTF("[@charset] Extracted: '%s'\n", RSTRING_PTR(charset));
-                }
-            }
-        }
-    }
-
-    // Parse CSS using our C parser implementation
-    // Returns hash: {query_string => [rules]} already grouped
-    VALUE rules_by_media = parse_css_impl(css_string, depth, Qnil);
-
-    // GC Guard: Protect Ruby objects from garbage collection
-    RB_GC_GUARD(css_string);
-    RB_GC_GUARD(rules_by_media);
-    RB_GC_GUARD(charset);
-
-    // At depth 0 (top-level parse), return hash with rules and charset
-    // Nested parses (depth > 0) return the hash directly
-    if (depth == 0) {
-        VALUE result = rb_hash_new();
-        rb_hash_aset(result, ID2SYM(rb_intern("rules")), rules_by_media);
-        rb_hash_aset(result, ID2SYM(rb_intern("charset")), charset);
-        return result;
-    }
-    return rules_by_media;
-}
-
-/*
- * Ruby-facing wrapper for parse_declarations
- *
- * @param declarations_string [String] CSS declarations like "color: red; margin: 10px"
- * @return [Array<Declaration>] Array of parsed declaration structs
- */
-static VALUE parse_declarations(VALUE self, VALUE declarations_string) {
-    Check_Type(declarations_string, T_STRING);
-
-    const char *input = RSTRING_PTR(declarations_string);
-    long input_len = RSTRING_LEN(declarations_string);
-
-    // Strip outer braces and whitespace (css_parser compatibility)
-    const char *start = input;
-    const char *end = input + input_len;
-
-    while (start < end && (IS_WHITESPACE(*start) || *start == '{')) start++;
-    while (end > start && (IS_WHITESPACE(*(end-1)) || *(end-1) == '}')) end--;
-
-    VALUE result = parse_declarations_string(start, end);
-
-    RB_GC_GUARD(result);
-    return result;
-}
-
-// Public wrapper for Ruby - starts at depth 0
-static VALUE parse_css(VALUE self, VALUE css_string) {
-    // Verify that cRule was initialized in Init_cataract
-    if (cRule == Qnil || cRule == 0) {
-        rb_raise(rb_eRuntimeError, "cRule struct class not initialized - Init_cataract may have failed");
-    }
-    return parse_css_internal(self, css_string, 0);
-}
-
-/*
- * Convert array of Rule structs to full CSS string
- * Format: "selector { prop: value; }\nselector2 { prop: value; }"
- */
-static VALUE rules_to_s(VALUE self, VALUE rules_array) {
-    Check_Type(rules_array, T_ARRAY);
-
-    long len = RARRAY_LEN(rules_array);
-    if (len == 0) {
-        return rb_str_new_cstr("");
-    }
-
-    // Estimate: ~100 chars per rule (selector + declarations)
-    VALUE result = rb_str_buf_new(len * 100);
-
-    for (long i = 0; i < len; i++) {
-        VALUE rule = rb_ary_entry(rules_array, i);
-
-        // Validate this is a Rule struct
-        if (!RB_TYPE_P(rule, T_STRUCT)) {
-            rb_raise(rb_eTypeError,
-                     "Expected array of Rule structs, got %s at index %ld",
-                     rb_obj_classname(rule), i);
-        }
-
-        // Extract: selector, declarations, specificity, media_query
-        VALUE selector = rb_struct_aref(rule, INT2FIX(RULE_SELECTOR));
-        VALUE declarations = rb_struct_aref(rule, INT2FIX(RULE_DECLARATIONS));
-
-        // Append selector
-        rb_str_buf_append(result, selector);
-        rb_str_buf_cat2(result, " { ");
-
-        // Serialize each declaration
-        long decl_len = RARRAY_LEN(declarations);
-        for (long j = 0; j < decl_len; j++) {
-            VALUE decl = rb_ary_entry(declarations, j);
-
-            VALUE property = rb_struct_aref(decl, INT2FIX(DECL_PROPERTY));
-            VALUE value = rb_struct_aref(decl, INT2FIX(DECL_VALUE));
-            VALUE important = rb_struct_aref(decl, INT2FIX(DECL_IMPORTANT));
-
-            rb_str_buf_append(result, property);
-            rb_str_buf_cat2(result, ": ");
-            rb_str_buf_append(result, value);
-
-            if (RTEST(important)) {
-                rb_str_buf_cat2(result, " !important");
-            }
-
-            rb_str_buf_cat2(result, "; ");
-        }
-
-        rb_str_buf_cat2(result, "}\n");
-
-        RB_GC_GUARD(rule);
-        RB_GC_GUARD(selector);
-        RB_GC_GUARD(declarations);
-    }
-
-    RB_GC_GUARD(result);
-    return result;
-}
-
-/*
- * Convert array of Declaration structs to CSS string
- * Format: "prop: value; prop2: value2 !important; "
- *
- * This is the core serialization logic used by both:
- * - Declarations#to_s (instance method)
- * - Internal C serialization (stylesheet.c)
- *
- * Exported (non-static) so stylesheet.c can call it
- */
-VALUE declarations_array_to_s(VALUE declarations_array) {
-    Check_Type(declarations_array, T_ARRAY);
-
-    long len = RARRAY_LEN(declarations_array);
-    if (len == 0) {
-        return rb_str_new_cstr("");
-    }
-
-    // Use rb_str_buf_new for efficient string building
-    VALUE result = rb_str_buf_new(len * 32); // Estimate 32 chars per declaration
-
-    for (long i = 0; i < len; i++) {
-        VALUE decl = rb_ary_entry(declarations_array, i);
-
-        // Validate this is a Declaration struct
-        if (!RB_TYPE_P(decl, T_STRUCT) || rb_obj_class(decl) != cDeclaration) {
-            rb_raise(rb_eTypeError,
-                     "Expected array of Declaration structs, got %s at index %ld",
-                     rb_obj_classname(decl), i);
-        }
-
-        // Extract struct fields
-        VALUE property = rb_struct_aref(decl, INT2FIX(DECL_PROPERTY));
-        VALUE value = rb_struct_aref(decl, INT2FIX(DECL_VALUE));
-        VALUE important = rb_struct_aref(decl, INT2FIX(DECL_IMPORTANT));
-
-        // Append: "property: value"
-        rb_str_buf_append(result, property);
-        rb_str_buf_cat2(result, ": ");
-        rb_str_buf_append(result, value);
-
-        // Append " !important" if needed
-        if (RTEST(important)) {
-            rb_str_buf_cat2(result, " !important");
-        }
-
-        rb_str_buf_cat2(result, "; ");
-
-        RB_GC_GUARD(decl);
-        RB_GC_GUARD(property);
-        RB_GC_GUARD(value);
-        RB_GC_GUARD(important);
-    }
-
-    // Strip trailing space
-    rb_str_set_len(result, RSTRING_LEN(result) - 1);
-
-    RB_GC_GUARD(result);
-    return result;
-}
-
-/*
- * Instance method: Declarations#to_s
- * Converts declarations to CSS string
- *
- * @return [String] CSS declarations like "color: red; margin: 10px !important;"
- */
-static VALUE declarations_to_s_method(VALUE self) {
-    // Get @values instance variable (array of Declaration structs)
-    VALUE values = rb_ivar_get(self, rb_intern("@values"));
-
-    // Call core serialization function
-    return declarations_array_to_s(values);
-}
-
-void Init_cataract() {
-    VALUE module = rb_define_module("Cataract");
-
-    // Initialize merge constants (cached strings and symbol IDs)
-    init_merge_constants();
-
-    // Define error class hierarchy
-    eCataractError = rb_define_class_under(module, "Error", rb_eStandardError);
-    eParseError = rb_define_class_under(module, "ParseError", eCataractError);
-    eDepthError = rb_define_class_under(module, "DepthError", eCataractError);
-    eSizeError = rb_define_class_under(module, "SizeError", eCataractError);
-
-    // Define Cataract::Declarations class (Ruby side will add methods)
-    VALUE cDeclarations = rb_define_class_under(module, "Declarations", rb_cObject);
-
-    // Define Cataract::Declaration = Struct.new(:property, :value, :important)
-    cDeclaration = rb_struct_define_under(
-        module,
-        "Declaration",
-        "property",
-        "value",
-        "important",
-        NULL
-    );
-
-    // Add methods to Declarations class
-    rb_define_method(cDeclarations, "to_s", declarations_to_s_method, 0);
-
-    // Define Cataract::Rule = Struct.new(:selector, :declarations, :specificity)
-    // Note: media_query removed - media info now stored at group level in hash structure
-    cRule = rb_struct_define_under(
-        module,
-        "Rule",
-        "selector",
-        "declarations",
-        "specificity",
-        NULL
-    );
-
-    // Define Cataract::Stylesheet class (Ruby side will reopen and add methods)
-    rb_define_class_under(module, "Stylesheet", rb_cObject);
-
-    rb_define_module_function(module, "parse_css", parse_css, 1);
-    rb_define_module_function(module, "parse_declarations", parse_declarations, 1);
-    rb_define_module_function(module, "calculate_specificity", calculate_specificity, 1);
-    rb_define_module_function(module, "merge_rules", cataract_merge_wrapper, 1);
-    rb_define_module_function(module, "apply_cascade", cataract_merge_wrapper, 1);  // Alias with better name
-    /* @api private */
-    rb_define_module_function(module, "_rules_to_s", rules_to_s, 1);
-
-    /* @api private */
-    rb_define_module_function(module, "_split_value", cataract_split_value, 1);
-    /* @api private */
-    rb_define_module_function(module, "_expand_margin", cataract_expand_margin, 1);
-    /* @api private */
-    rb_define_module_function(module, "_expand_padding", cataract_expand_padding, 1);
-    /* @api private */
-    rb_define_module_function(module, "_expand_border_color", cataract_expand_border_color, 1);
-    /* @api private */
-    rb_define_module_function(module, "_expand_border_style", cataract_expand_border_style, 1);
-    /* @api private */
-    rb_define_module_function(module, "_expand_border_width", cataract_expand_border_width, 1);
-    /* @api private */
-    rb_define_module_function(module, "_expand_border", cataract_expand_border, 1);
-    /* @api private */
-    rb_define_module_function(module, "_expand_border_side", cataract_expand_border_side, 2);
-    /* @api private */
-    rb_define_module_function(module, "_expand_font", cataract_expand_font, 1);
-    /* @api private */
-    rb_define_module_function(module, "_expand_list_style", cataract_expand_list_style, 1);
-    /* @api private */
-    rb_define_module_function(module, "_expand_background", cataract_expand_background, 1);
-
-    // Shorthand creation (inverse of expansion)
-    /* @api private */
-    rb_define_module_function(module, "_create_margin_shorthand", cataract_create_margin_shorthand, 1);
-    /* @api private */
-    rb_define_module_function(module, "_create_padding_shorthand", cataract_create_padding_shorthand, 1);
-    /* @api private */
-    rb_define_module_function(module, "_create_border_width_shorthand", cataract_create_border_width_shorthand, 1);
-    /* @api private */
-    rb_define_module_function(module, "_create_border_style_shorthand", cataract_create_border_style_shorthand, 1);
-    /* @api private */
-    rb_define_module_function(module, "_create_border_color_shorthand", cataract_create_border_color_shorthand, 1);
-    /* @api private */
-    rb_define_module_function(module, "_create_border_shorthand", cataract_create_border_shorthand, 1);
-    /* @api private */
-    rb_define_module_function(module, "_create_background_shorthand", cataract_create_background_shorthand, 1);
-    /* @api private */
-    rb_define_module_function(module, "_create_font_shorthand", cataract_create_font_shorthand, 1);
-    /* @api private */
-    rb_define_module_function(module, "_create_list_style_shorthand", cataract_create_list_style_shorthand, 1);
-
-    // Serialization
-    /* @api private */
-    rb_define_module_function(module, "_stylesheet_to_s_c", stylesheet_to_s_c, 2);
-    /* @api private */
-    rb_define_module_function(module, "_stylesheet_to_formatted_s_c", stylesheet_to_formatted_s_c, 2);
-
-    // Import scanning
-    rb_define_module_function(module, "extract_imports", extract_imports, 1);
-
-    // Export string allocation mode as a constant for verification in benchmarks
-    #ifdef DISABLE_STR_BUF_OPTIMIZATION
-        rb_define_const(module, "STRING_ALLOC_MODE", ID2SYM(rb_intern("dynamic")));
-    #else
-        rb_define_const(module, "STRING_ALLOC_MODE", ID2SYM(rb_intern("buffer")));
-    #endif
-
-    // Export compile-time optimization flags as a hash for runtime introspection
-    VALUE compile_flags = rb_hash_new();
-
-    #ifdef DISABLE_STR_BUF_OPTIMIZATION
-        rb_hash_aset(compile_flags, ID2SYM(rb_intern("str_buf_optimization")), Qfalse);
-    #else
-        rb_hash_aset(compile_flags, ID2SYM(rb_intern("str_buf_optimization")), Qtrue);
-    #endif
-
-    #ifdef CATARACT_DEBUG
-        rb_hash_aset(compile_flags, ID2SYM(rb_intern("debug")), Qtrue);
-    #else
-        rb_hash_aset(compile_flags, ID2SYM(rb_intern("debug")), Qfalse);
-    #endif
-
-    #ifdef DISABLE_LOOP_UNROLL
-        rb_hash_aset(compile_flags, ID2SYM(rb_intern("loop_unroll")), Qfalse);
-    #else
-        rb_hash_aset(compile_flags, ID2SYM(rb_intern("loop_unroll")), Qtrue);
-    #endif
-
-    // Note: Compiler flags like -O3, -march=native, -funroll-loops don't have
-    // preprocessor defines, so we can't detect them at runtime. They're purely
-    // compiler optimizations that affect the generated code.
-
-    rb_define_const(module, "COMPILE_FLAGS", compile_flags);
-
-    // NOTE: Color conversion is now a separate extension (cataract_color)
-    // It's initialized when you require 'cataract/color_conversion'
-}
-
-// NOTE: shorthand_expander.c and value_splitter.c are now compiled separately (not included)
-