charlock_holmes 0.7.5 → 0.7.9

checksums.yaml

@@ -1,7 +1,7 @@
 ---
-SHA1:
-  metadata.gz: 57e54315a83df12c843ea67de7213128f74f2a7e
-  data.tar.gz: f773e478f37f51fe159787964aadf794c25d347a
+SHA256:
+  metadata.gz: 11393fb92c8ecb1d18d4741c9915c0293fde4381ccfeae3e929c84970b6779f4
+  data.tar.gz: addcb3bdf4fc04e53f7a287c483d6510205b317d980f3a8b21a26940dc5a198d
 SHA512:
-  metadata.gz: cc59faffae9eacb3515fbc653b326f26fde8b607177f11ba64fcbcdd4033f06b03531c2f3e388d4d66ce09433f39164583b024beb189b8b37adc9a26eaf2586d
-  data.tar.gz: 7e7cbe1bb7d7bc8bfbae5417ce1c8047aa9e324fdfa528952c848f3180ac90c3ea3c81bfe74829352a1e81ac3648ac195d0f02f38ad4935876a5af20615396a3
+  metadata.gz: 1729db533d5d41aa4708888763270e3f346552a01a8f29371a010253db3bb0cc4f58a7ed628b735829e7f5d9763d35de20a17855dc9c393d4244d1acf30f0dbe
+  data.tar.gz: 5033611bfe43c98fd8d88a7141b94676750a2c7fe49a44f5e621d61a4bfc4eef8e627b478471c0cce7dc128b4fc3ad6922129a73c48274681522ff639e82dd75

data/ext/charlock_holmes/common.h

@@ -38,4 +38,19 @@ static inline VALUE charlock_new_str2(const char *str)
 #endif
 }
 
+
+#ifdef __cplusplus
+extern "C"
+{
+#endif
+
+extern void Init_charlock_holmes(void);
+extern void _init_charlock_encoding_detector(void);
+extern void _init_charlock_converter(void);
+extern void _init_charlock_transliterator(void);
+
+#ifdef __cplusplus
+}
+#endif
+
 #endif

data/ext/charlock_holmes/converter.c

@@ -20,7 +20,7 @@ static VALUE rb_converter_convert(VALUE self, VALUE rb_txt, VALUE rb_src_enc, VA
 	Check_Type(rb_dst_enc, T_STRING);
 
 	src_txt = RSTRING_PTR(rb_txt);
-	src_len = RSTRING_LEN(rb_txt);
+	src_len = (int32_t)RSTRING_LEN(rb_txt);
 	src_enc = RSTRING_PTR(rb_src_enc);
 	dst_enc = RSTRING_PTR(rb_dst_enc);
 
@@ -29,7 +29,7 @@ static VALUE rb_converter_convert(VALUE self, VALUE rb_txt, VALUE rb_src_enc, VA
 	if (status != U_BUFFER_OVERFLOW_ERROR) {
 		rb_raise(rb_eArgError, "%s", u_errorName(status));
 	}
-	out_buf = malloc(out_len);
+	out_buf = (char *) malloc(out_len);
 
 	// now do the actual conversion
 	status = U_ZERO_ERROR;
@@ -50,7 +50,7 @@ static VALUE rb_converter_convert(VALUE self, VALUE rb_txt, VALUE rb_src_enc, VA
 	return rb_out;
 }
 
-void _init_charlock_converter() {
+void _init_charlock_converter(void) {
 	rb_cConverter = rb_define_class_under(rb_mCharlockHolmes, "Converter", rb_cObject);
 
 	rb_define_singleton_method(rb_cConverter, "convert", rb_converter_convert, 3);

data/ext/charlock_holmes/encoding_detector.c

@@ -8,6 +8,25 @@ typedef struct {
 	UCharsetDetector *csd;
 } charlock_detector_t;
 
+static void rb_encdec__free(void *obj)
+{
+	charlock_detector_t *detector;
+
+	detector = (charlock_detector_t *)obj;
+
+	if (detector->csd)
+		ucsdet_close(detector->csd);
+
+	free(detector);
+}
+
+static const rb_data_type_t charlock_detector_type = {
+    "Charlock/Detector",
+    { 0, rb_encdec__free, 0, },
+    0, 0,
+    RUBY_TYPED_FREE_IMMEDIATELY,
+};
+
 static VALUE rb_encdec_buildmatch(const UCharsetMatch *match)
 {
 	UErrorCode status = U_ZERO_ERROR;
@@ -47,7 +66,7 @@ static VALUE rb_encdec_buildmatch(const UCharsetMatch *match)
 	return rb_match;
 }
 
-static VALUE rb_encdec_binarymatch() {
+static VALUE rb_encdec_binarymatch(void) {
 	VALUE rb_match;
 
 	rb_match = rb_hash_new();
@@ -167,7 +186,7 @@ static VALUE rb_encdec_detect(int argc, VALUE *argv, VALUE self)
 	rb_scan_args(argc, argv, "11", &rb_str, &rb_enc_hint);
 
 	Check_Type(rb_str, T_STRING);
-	Data_Get_Struct(self, charlock_detector_t, detector);
+	TypedData_Get_Struct(self, charlock_detector_t, &charlock_detector_type, detector);
 
 	// first lets see if this is binary content
 	if (detect_binary_content(self, rb_str)) {
@@ -180,7 +199,7 @@ static VALUE rb_encdec_detect(int argc, VALUE *argv, VALUE self)
 
 	if (!NIL_P(rb_enc_hint)) {
 		Check_Type(rb_enc_hint, T_STRING);
-		ucsdet_setDeclaredEncoding(detector->csd, RSTRING_PTR(rb_enc_hint), RSTRING_LEN(rb_enc_hint), &status);
+		ucsdet_setDeclaredEncoding(detector->csd, RSTRING_PTR(rb_enc_hint), (int32_t)RSTRING_LEN(rb_enc_hint), &status);
 	}
 
 	return rb_encdec_buildmatch(ucsdet_detect(detector->csd, &status));
@@ -215,7 +234,7 @@ static VALUE rb_encdec_detect_all(int argc, VALUE *argv, VALUE self)
 	rb_scan_args(argc, argv, "11", &rb_str, &rb_enc_hint);
 
 	Check_Type(rb_str, T_STRING);
-	Data_Get_Struct(self, charlock_detector_t, detector);
+	TypedData_Get_Struct(self, charlock_detector_t, &charlock_detector_type, detector);
 
 	rb_ret = rb_ary_new();
 
@@ -229,7 +248,7 @@ static VALUE rb_encdec_detect_all(int argc, VALUE *argv, VALUE self)
 
 	if (!NIL_P(rb_enc_hint)) {
 		Check_Type(rb_enc_hint, T_STRING);
-		ucsdet_setDeclaredEncoding(detector->csd, RSTRING_PTR(rb_enc_hint), RSTRING_LEN(rb_enc_hint), &status);
+		ucsdet_setDeclaredEncoding(detector->csd, RSTRING_PTR(rb_enc_hint), (int32_t)RSTRING_LEN(rb_enc_hint), &status);
 	}
 
 	csm = ucsdet_detectAll(detector->csd, &match_count, &status);
@@ -257,7 +276,7 @@ static VALUE rb_get_strip_tags(VALUE self)
 	UBool val;
 	VALUE rb_val;
 
-	Data_Get_Struct(self, charlock_detector_t, detector);
+	TypedData_Get_Struct(self, charlock_detector_t, &charlock_detector_type, detector);
 
 	val = ucsdet_isInputFilterEnabled(detector->csd);
 
@@ -279,7 +298,7 @@ static VALUE rb_set_strip_tags(VALUE self, VALUE rb_val)
 	charlock_detector_t *detector;
 	UBool val;
 
-	Data_Get_Struct(self, charlock_detector_t, detector);
+	TypedData_Get_Struct(self, charlock_detector_t, &charlock_detector_type, detector);
 
 	val = rb_val == Qtrue ? 1 : 0;
 
@@ -334,26 +353,14 @@ static VALUE rb_get_supported_encodings(VALUE klass)
 	return rb_encoding_list;
 }
 
-static void rb_encdec__free(void *obj)
-{
-	charlock_detector_t *detector;
-
-	detector = (charlock_detector_t *)obj;
-
-	if (detector->csd)
-		ucsdet_close(detector->csd);
-
-	free(detector);
-}
-
 static VALUE rb_encdec__alloc(VALUE klass)
 {
 	charlock_detector_t *detector;
 	UErrorCode status = U_ZERO_ERROR;
 	VALUE obj;
 
-	detector = calloc(1, sizeof(charlock_detector_t));
-	obj = Data_Wrap_Struct(klass, NULL, rb_encdec__free, (void *)detector);
+	detector = (charlock_detector_t *) calloc(1, sizeof(charlock_detector_t));
+        obj = TypedData_Wrap_Struct(klass, &charlock_detector_type, (void *)detector);
 
 	detector->csd = ucsdet_open(&status);
 	if (U_FAILURE(status)) {
@@ -363,7 +370,7 @@ static VALUE rb_encdec__alloc(VALUE klass)
 	return obj;
 }
 
-void _init_charlock_encoding_detector()
+void _init_charlock_encoding_detector(void)
 {
 	rb_cEncodingDetector = rb_define_class_under(rb_mCharlockHolmes, "EncodingDetector", rb_cObject);
 	rb_define_alloc_func(rb_cEncodingDetector, rb_encdec__alloc);

data/ext/charlock_holmes/ext.c

@@ -1,15 +1,11 @@
 #include "common.h"
 
-extern void _init_charlock_encoding_detector();
-extern void _init_charlock_converter();
-extern void _init_charlock_transliterator();
-
 VALUE rb_mCharlockHolmes;
 
-void Init_charlock_holmes() {
+void Init_charlock_holmes(void) {
 	rb_mCharlockHolmes = rb_define_module("CharlockHolmes");
 
 	_init_charlock_encoding_detector();
 	_init_charlock_converter();
 	_init_charlock_transliterator();
-}
+}

data/ext/charlock_holmes/extconf.rb

@@ -1,14 +1,5 @@
 require 'mkmf'
 
-CWD = File.expand_path(File.dirname(__FILE__))
-def sys(cmd)
-  puts "  -- #{cmd}"
-  unless ret = xsystem(cmd)
-    raise "#{cmd} failed, please report issue on https://github.com/brianmario/charlock_holmes"
-  end
-  ret
-end
-
 if `which make`.strip.empty?
   STDERR.puts "\n\n"
   STDERR.puts "***************************************************************************************"
@@ -21,25 +12,28 @@ end
 # ICU dependency
 #
 
-dir_config 'icu'
-
-rubyopt = ENV.delete("RUBYOPT")
-
-icu4c = "/usr"
-# detect homebrew installs
-if !have_library 'icui18n'
-  base = if !`which brew`.empty?
-    `brew --prefix`.strip
-  elsif File.exists?("/usr/local/Cellar/icu4c")
-    '/usr/local/Cellar'
-  end
+ldflags = cppflags = nil
 
-  if base and icu4c = Dir[File.join(base, 'Cellar/icu4c/*')].sort.last
-    $INCFLAGS << " -I#{icu4c}/include "
-    $LDFLAGS  << " -L#{icu4c}/lib "
+if RbConfig::CONFIG["host_os"] =~ /darwin/
+  begin
+    brew_prefix = `brew --prefix icu4c`.chomp
+    ldflags   = "#{brew_prefix}/lib"
+    cppflags  = "#{brew_prefix}/include"
+    pkg_conf  = "#{brew_prefix}/lib/pkgconfig"
+    # pkg_config should be less error prone than parsing compiler
+    # commandline options, but we need to set default ldflags and cpp flags
+    # in case the user doesn't have pkg-config installed
+    ENV['PKG_CONFIG_PATH'] ||= pkg_conf
+  rescue
   end
 end
 
+dir_config 'icu', cppflags, ldflags
+
+pkg_config("icu-i18n")
+pkg_config("icu-io")
+pkg_config("icu-uc")
+
 unless have_library 'icui18n' and have_header 'unicode/ucnv.h'
   STDERR.puts "\n\n"
   STDERR.puts "***************************************************************************************"
@@ -52,15 +46,96 @@ have_library 'z' or abort 'libz missing'
 have_library 'icuuc' or abort 'libicuuc missing'
 have_library 'icudata' or abort 'libicudata missing'
 
-# icu4c might be built in C++11 mode, but it also might not have been
-icuconfig = `which icu-config`.chomp
-icuconfig = "#{icu4c}/bin/icu-config" if icuconfig.empty?
-if File.exist?(icuconfig) && `#{icuconfig} --cxxflags`.include?("c++11")
-  $CXXFLAGS << ' -std=c++11'
-end
-
 $CFLAGS << ' -Wall -funroll-loops'
 $CFLAGS << ' -Wextra -O0 -ggdb3' if ENV['DEBUG']
 
-ENV['RUBYOPT'] = rubyopt
+minimal_program = <<~SRC
+  #include <unicode/translit.h>
+  int main() { return 0; }
+SRC
+
+# Pass -x c++ to force gcc to compile the test program
+# as C++ (as it will end in .c by default).
+compile_options = +"-x c++"
+
+icu_requires_version_flag = checking_for("icu that requires explicit C++ version flag") do
+  !try_compile(minimal_program, compile_options)
+end
+
+if icu_requires_version_flag
+  abort "Cannot compile icu with your compiler: recent versions require C++17 support." unless %w[c++20 c++17 c++11 c++0x].any? do |std|
+    checking_for("icu that compiles with #{std} standard") do
+      flags = compile_options + " -std=#{std}"
+      if try_compile(minimal_program, flags)
+        $CPPFLAGS << flags
+
+        true
+      end
+    end
+  end
+end
+
+def libflag_to_filename(ldflag)
+  case ldflag
+  when /\A-l(.+)/
+    "lib#{Regexp.last_match(1)}.#{$LIBEXT}"
+  end
+end
+
+def resolve_static_library(libflag, dirs)
+  filename = libflag_to_filename(libflag)
+
+  dir = dirs.find { |path| File.exist?(File.join(path, filename)) }
+
+  raise "Unable to find #{filename} in #{dirs}" unless dir
+
+  File.join(dir, filename)
+end
+
+def substitute_static_libs(packages)
+  packages.each do |pkg|
+    unless pkg_config(pkg)
+      message = <<~MSG
+        Unable to run `pkg-config #{pkg}`.
+
+        Check that PKG_CONFIG_PATH includes #{pkg}.pc (or unset it if it's already set).
+
+        Current environment:
+        PKG_CONFIG_PATH=#{ENV['PKG_CONFIG_PATH']}
+      MSG
+
+      raise message
+    end
+  end
+
+  # First, find all the -l<lib> flags added by pkg-config. We want to drop
+  # these dynamically linked libraries and substitute them with the static libraries.
+  libflags = packages.map do |pkg|
+    pkg_config(pkg, 'libs-only-l')&.strip&.split(' ')
+  end.flatten.uniq
+
+  # To find where the static libraries live, we need to search the
+  # library paths given by the -L flag from pkg-config.
+  lib_paths = packages.map do |pkg|
+    include_path = pkg_config(pkg, 'libs-only-L')&.strip
+    include_path&.split(' ')&.map { |lib| lib.gsub(/^-L/, '') }
+  end.flatten.uniq
+
+  # Drop the -l<lib> flags and add in the static libraries.
+  new_libs = $libs.shellsplit
+  new_libs.reject! { |arg| libflags.include?(arg) }
+  libflags.each { |flag| new_libs << resolve_static_library(flag, lib_paths) }
+  $libs = new_libs.uniq.shelljoin
+end
+
+static_p = enable_config('static', false)
+message "Static linking is #{static_p ? 'enabled' : 'disabled'}.\n"
+
+if static_p
+  $CXXFLAGS << ' -fPIC'
+  ENV['PKG_CONFIG_ALLOW_SYSTEM_LIBS'] = '1'
+
+  substitute_static_libs(%w[icu-i18n icu-io icu-uc])
+end
+
 create_makefile 'charlock_holmes/charlock_holmes'

data/ext/charlock_holmes/transliterator.cpp

@@ -36,7 +36,7 @@ static VALUE rb_cTransliterator;
 
 static VALUE rb_transliterator_id_list(VALUE self) {
   UErrorCode status = U_ZERO_ERROR;
-  StringEnumeration *id_list;
+  icu::StringEnumeration *id_list;
   int32_t id_list_size;
   const char *curr_id;
   int32_t curr_id_len;
@@ -44,7 +44,7 @@ static VALUE rb_transliterator_id_list(VALUE self) {
   VALUE rb_curr_id;
 
   id_list_size = 0;
-  id_list = Transliterator::getAvailableIDs(status);
+  id_list = icu::Transliterator::getAvailableIDs(status);
   if(!U_SUCCESS(status)) {
     rb_raise(rb_eArgError, "%s", u_errorName(status));
   }
@@ -78,12 +78,12 @@ static VALUE rb_transliterator_id_list(VALUE self) {
 static VALUE rb_transliterator_transliterate(VALUE self, VALUE rb_txt, VALUE rb_id) {
   UErrorCode status = U_ZERO_ERROR;
   UParseError p_error;
-  Transliterator *trans;
+  icu::Transliterator *trans;
   const char *txt;
   size_t txt_len;
   const char *id;
   size_t id_len;
-  UnicodeString *u_txt;
+  icu::UnicodeString *u_txt;
   std::string result;
   VALUE rb_out;
 
@@ -98,14 +98,14 @@ static VALUE rb_transliterator_transliterate(VALUE self, VALUE rb_txt, VALUE rb_
   id = RSTRING_PTR(rb_id);
   id_len = RSTRING_LEN(rb_id);
 
-  trans = Transliterator::createInstance(UnicodeString(id, id_len), UTRANS_FORWARD, p_error, status);
+  trans = icu::Transliterator::createInstance(icu::UnicodeString(id, id_len), UTRANS_FORWARD, p_error, status);
   if(!U_SUCCESS(status)) {
     rb_raise(rb_eArgError, "%s", u_errorName(status));
   }
 
-  u_txt = new UnicodeString(txt, txt_len);
+  u_txt = new icu::UnicodeString(txt, txt_len);
   trans->transliterate(*u_txt);
-  StringByteSink<std::string> sink(&result);
+  icu::StringByteSink<std::string> sink(&result);
   u_txt->toUTF8(sink);
 
   delete u_txt;
@@ -116,7 +116,7 @@ static VALUE rb_transliterator_transliterate(VALUE self, VALUE rb_txt, VALUE rb_
   return rb_out;
 }
 
-void _init_charlock_transliterator() {
+void _init_charlock_transliterator(void) {
 #ifdef HAVE_RUBY_ENCODING_H
   rb_eEncodingCompatibilityError = rb_const_get(rb_cEncoding, rb_intern("CompatibilityError"));
 #endif

data/lib/charlock_holmes/version.rb

@@ -1,3 +1,3 @@
 module CharlockHolmes
-  VERSION = "0.7.5"
+  VERSION = "0.7.9"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: charlock_holmes
 version: !ruby/object:Gem::Version
-  version: 0.7.5
+  version: 0.7.9
 platform: ruby
 authors:
 - Brian Lopez
@@ -9,50 +9,50 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2017-08-14 00:00:00.000000000 Z
+date: 2024-07-10 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake-compiler
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.7.5
+        version: '1.0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.7.5
+        version: '1.0'
 - !ruby/object:Gem::Dependency
   name: minitest
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '5.11'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '5.11'
 - !ruby/object:Gem::Dependency
   name: chardet
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '0.9'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '0.9'
 description: charlock_holmes provides binary and text detection as well as text transcoding
   using libicu
 email: seniorlopez@gmail.com
@@ -91,8 +91,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.6.11
+rubygems_version: 3.0.3.1
 signing_key: 
 specification_version: 4
 summary: Character encoding detection, brought to you by ICU