mittens 0.3.2 → 0.3.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: b5d5216e3e6a0e86aa92ab80ec8876e66be4d766d500d82609f166452f28fb08
-  data.tar.gz: f5db98b8a25cc9fb2183a2f0010695239a66c82faf00e1ddb34355739087c48b
+  metadata.gz: d7e35e680454076f43e5973db399645697e0e98732555293d6b5e3bae115308d
+  data.tar.gz: 6c619fe9518ba5ddd42001ee0e99b4883b0fa703d5bf7b45e9022a039c78fb7d
 SHA512:
-  metadata.gz: 886dac65964a1b37c5156ff27b7703a71d6acde7f01c9525d1a564636b1fdfa99bf8871b98e8628ed511cdda536e4d28d7e5bebbefdb9b0a67454346e1f571c7
-  data.tar.gz: 62a9a20fca62714a28826626183dba6816d11863fb169da43c0fa7d35d2a7978704f4764be63bce5de1986b188566f15f7ec5be5a0d415d9a85fad7285b2f1ac
+  metadata.gz: 1148ef35cf15d3027b51934844763437f1a606b2eed92dec9ccb17c86a78e544d664a51e6470da7d117c3a30fb1e0638119f3f1091fb203b29630e11513d5672
+  data.tar.gz: c1cc877ce40bf0c2fc3dd9a7946b4cf0165f870392ff87024c3a717cfe742420b8de6b1bbe8ecf5645e0566f81ed906f1ce87a9072ba645c4ca3096702b253d8

data/CHANGELOG.md

@@ -1,3 +1,11 @@
+## 0.3.4 (2026-08-06)
+
+- Fixed `dup` and `clone` methods
+
+## 0.3.3 (2026-08-05)
+
+- Fixed issue with GC compaction
+
 ## 0.3.2 (2026-07-10)
 
 - Updated Snowball to 3.1.1

data/ext/mittens/ext.c

@@ -1,21 +1,37 @@
+#include <limits.h>
+
 #include "libstemmer.h"
 #include "ruby/ruby.h"
 
+#define GetAlgorithm(language) (NIL_P(language) ? "english" : StringValueCStr(language))
+#define GetStemmer(obj, ptr) TypedData_Get_Struct((obj), stemmer_t, &stemmer_data_type, (ptr))
+
 typedef struct stemmer {
-    struct sb_stemmer * stemmer;
+    struct sb_stemmer* stemmer;
+    VALUE language;
 } stemmer_t;
 
-static void stemmer_free(void *ptr)
+static void stemmer_mark(void* ptr)
 {
-    stemmer_t *stemmer = (stemmer_t *)ptr;
-    // safe to pass null pointer according to docs
-    sb_stemmer_delete(stemmer->stemmer);
+    stemmer_t* stemmer = (stemmer_t*) ptr;
+    rb_gc_mark(stemmer->language);
+}
+
+static void stemmer_free(void* ptr)
+{
+    stemmer_t* stemmer = (stemmer_t*) ptr;
+
+    // safe to pass null pointer, but check anyways
+    if (stemmer->stemmer != NULL)
+        sb_stemmer_delete(stemmer->stemmer);
+
     xfree(ptr);
 }
 
 const rb_data_type_t stemmer_data_type = {
-    .wrap_struct_name = "stemmer",
+    .wrap_struct_name = "Mittens::Stemmer",
     .function = {
+        .dmark = stemmer_mark,
         .dfree = stemmer_free,
     },
     .flags = RUBY_TYPED_FREE_IMMEDIATELY
@@ -23,74 +39,104 @@ const rb_data_type_t stemmer_data_type = {
 
 static VALUE stemmer_allocate(VALUE klass)
 {
-    stemmer_t *stemmer;
+    stemmer_t* stemmer;
     VALUE obj = TypedData_Make_Struct(klass, stemmer_t, &stemmer_data_type, stemmer);
     stemmer->stemmer = NULL;
+    stemmer->language = Qnil;
     return obj;
 }
 
 static VALUE stemmer_initialize(int argc, VALUE* argv, VALUE self)
 {
-    VALUE opts;
-    rb_scan_args(argc, argv, ":", &opts);
+    stemmer_t* stemmer;
+    GetStemmer(self, stemmer);
 
-    const char * algorithm = "english";
-    if (!NIL_P(opts)) {
-        VALUE language = rb_hash_aref(opts, ID2SYM(rb_intern("language")));
-        if (!NIL_P(language)) {
-            Check_Type(language, T_STRING);
-            algorithm = RSTRING_PTR(language);
-        }
-    }
+    if (stemmer->stemmer != NULL)
+        rb_raise(rb_eRuntimeError, "stemmer already initialized");
 
-    stemmer_t *stemmer;
-    TypedData_Get_Struct(self, stemmer_t, &stemmer_data_type, stemmer);
+    VALUE opts;
+    rb_scan_args(argc, argv, ":", &opts);
 
-    // in case called multiple times
-    sb_stemmer_delete(stemmer->stemmer);
+    VALUE language = Qnil;
+    if (!NIL_P(opts))
+        language = rb_hash_aref(opts, ID2SYM(rb_intern("language")));
 
     // if adding support for encoding, may want to change encoding returned from stem
-    stemmer->stemmer = sb_stemmer_new(algorithm, NULL);
-    if (stemmer->stemmer == NULL) {
-        rb_raise(rb_eArgError, "unknown language: %s", algorithm);
+    stemmer->stemmer = sb_stemmer_new(GetAlgorithm(language), NULL);
+    if (stemmer->stemmer == NULL)
+    {
+        if (!NIL_P(language))
+            rb_raise(rb_eArgError, "unknown language: %" PRIsVALUE, language);
+
+        rb_raise(rb_eNoMemError, "failed to allocate memory");
     }
 
+    // copy (after GetAlgorithm) in case string is modified in future
+    stemmer->language = rb_obj_dup(language);
+
     return self;
 }
 
 static VALUE stemmer_stem(VALUE self, VALUE value)
 {
-    stemmer_t *stemmer;
-    TypedData_Get_Struct(self, stemmer_t, &stemmer_data_type, stemmer);
+    stemmer_t* stemmer;
+    GetStemmer(self, stemmer);
+
+    if (stemmer->stemmer == NULL)
+        rb_raise(rb_eRuntimeError, "stemmer not initialized");
 
-    Check_Type(value, T_STRING);
+    const char* word = StringValuePtr(value);
+    long size = RSTRING_LEN(value);
+    if (size > INT_MAX)
+        rb_raise(rb_eArgError, "string exceeds max length");
 
-    const sb_symbol * word = (const sb_symbol *) RSTRING_PTR(value);
-    int size = (int) RSTRING_LEN(value);
-    const sb_symbol * pointer_out = sb_stemmer_stem(stemmer->stemmer, word, size);
+    const sb_symbol* pointer_out = sb_stemmer_stem(stemmer->stemmer, (const sb_symbol*) word, (int) size);
+    int length_out = sb_stemmer_length(stemmer->stemmer);
 
-    return rb_utf8_str_new_cstr((char *) pointer_out);
+    // place after last use of word
+    RB_GC_GUARD(value);
+
+    return rb_utf8_str_new((char*) pointer_out, (long) length_out);
+}
+
+static VALUE stemmer_initialize_copy(VALUE self, VALUE orig)
+{
+    stemmer_t* stemmer;
+    GetStemmer(self, stemmer);
+
+    stemmer_t* orig_stemmer;
+    GetStemmer(orig, orig_stemmer);
+
+    if (stemmer->stemmer != NULL)
+        rb_raise(rb_eRuntimeError, "stemmer already initialized");
+
+    stemmer->stemmer = sb_stemmer_new(GetAlgorithm(orig_stemmer->language), NULL);
+    if (stemmer->stemmer == NULL)
+        rb_raise(rb_eNoMemError, "failed to allocate memory");
+
+    stemmer->language = orig_stemmer->language;
+
+    return self;
 }
 
 static VALUE stemmer_languages(VALUE klass)
 {
     VALUE out = rb_ary_new();
 
-    const char **language = sb_stemmer_list();
-    while (*language != NULL) {
-        rb_ary_push(out, rb_utf8_str_new_cstr(*language));
-        language++;
-    }
+    const char** language = sb_stemmer_list();
+    while (*language != NULL)
+        rb_ary_push(out, rb_utf8_str_new_cstr(*language++));
 
     return out;
 }
 
 void Init_ext(void)
 {
-    VALUE rb_mMittens = rb_define_module("Mittens");
-    VALUE rb_cStemmer = rb_define_class_under(rb_mMittens, "Stemmer", rb_cObject);
-    rb_define_alloc_func(rb_cStemmer, stemmer_allocate);
-    rb_define_method(rb_cStemmer, "initialize", stemmer_initialize, -1);
-    rb_define_method(rb_cStemmer, "stem", stemmer_stem, 1);
-    rb_define_singleton_method(rb_cStemmer, "languages", stemmer_languages, 0);
+    VALUE mMittens = rb_define_module("Mittens");
+    VALUE cStemmer = rb_define_class_under(mMittens, "Stemmer", rb_cObject);
+    rb_define_alloc_func(cStemmer, stemmer_allocate);
+    rb_define_method(cStemmer, "initialize", stemmer_initialize, -1);
+    rb_define_method(cStemmer, "stem", stemmer_stem, 1);
+    rb_define_method(cStemmer, "initialize_copy", stemmer_initialize_copy, 1);
+    rb_define_singleton_method(cStemmer, "languages", stemmer_languages, 0);
 }

data/ext/mittens/extconf.rb

@@ -10,6 +10,8 @@ output, status = Open3.capture2(make_cmd, "CFLAGS=#{cflags}", chdir: vendor)
 puts output
 raise "Command failed" unless status.success?
 
+append_cflags("-Wall -Wextra -Wpedantic -Wconversion")
+
 $INCFLAGS += " -I$(srcdir)/../../vendor/snowball/include"
 $LDFLAGS += " $(srcdir)/../../vendor/snowball/libstemmer.a"
 

data/lib/mittens/version.rb

@@ -1,3 +1,3 @@
 module Mittens
-  VERSION = "0.3.2"
+  VERSION = "0.3.4"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: mittens
 version: !ruby/object:Gem::Version
-  version: 0.3.2
+  version: 0.3.4
 platform: ruby
 authors:
 - Andrew Kane