charlock_holmes 0.7.8 → 0.7.9

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 63323878cbd6ae16be524e46cd02c593003b954a26f8b381595fdd8e4355d651
-  data.tar.gz: 7f8a02a33b2d9c3d0e3800ff2728a9f95b7a2e07692b2ca01b44953ae9cb54ed
+  metadata.gz: 11393fb92c8ecb1d18d4741c9915c0293fde4381ccfeae3e929c84970b6779f4
+  data.tar.gz: addcb3bdf4fc04e53f7a287c483d6510205b317d980f3a8b21a26940dc5a198d
 SHA512:
-  metadata.gz: d1e8235025a81e5dd289e52bcf1696c86ceaa7f39e6fa627fb2625a24a613e86868247bf2caedd5e837b1aea1db73f2259031093c668710cdd7498fea9b93b65
-  data.tar.gz: 7e60283edcbd9795185ea95717407361e02ec7b98114759a32924217231439edb45048890e210d2bda7a5408559ab05d65af9df13569be4a3699858aa2b1fcf8
+  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

@@ -49,4 +49,93 @@ have_library 'icudata' or abort 'libicudata missing'
 $CFLAGS << ' -Wall -funroll-loops'
 $CFLAGS << ' -Wextra -O0 -ggdb3' if ENV['DEBUG']
 
+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

@@ -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 = "0.7.8"
+  VERSION = "0.7.9"
 end

metadata

@@ -1,14 +1,15 @@
 --- !ruby/object:Gem::Specification
 name: charlock_holmes
 version: !ruby/object:Gem::Version
-  version: 0.7.8
+  version: 0.7.9
 platform: ruby
 authors:
 - Brian Lopez
 - Vicent Martí
+autorequire: 
 bindir: bin
 cert_chain: []
-date: 2024-06-28 00:00:00.000000000 Z
+date: 2024-07-10 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake-compiler
@@ -74,6 +75,7 @@ homepage: https://github.com/brianmario/charlock_holmes
 licenses:
 - MIT
 metadata: {}
+post_install_message: 
 rdoc_options:
 - "--charset=UTF-8"
 require_paths:
@@ -89,7 +91,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.6.0.dev
+rubygems_version: 3.0.3.1
+signing_key: 
 specification_version: 4
 summary: Character encoding detection, brought to you by ICU
 test_files: []