Merge pull request #252 from rrthomas/master

Fix Hunspell behaviour with apostrophes and some code clean-up

9 files changed, 50 insertions, 58 deletions

diff --git a/configure.ac b/configure.ac
index af388ca..18d988a 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1,4 +1,4 @@
-AC_INIT([enchant],[2.2.9])
+AC_INIT([enchant],[2.2.10])
 AC_CONFIG_SRCDIR(src/enchant.h)
 AC_CONFIG_AUX_DIR([build-aux])
 AM_INIT_AUTOMAKE([subdir-objects])
diff --git a/gnulib b/gnulib
-Subproject 28098428d3d371a238837f338739283cf19dc65
+Subproject b58bf6ee39a6a114550a6bb68e7db5262c17f8b
diff --git a/m4/.gitignore b/m4/.gitignore
index fc96477..9df6a16 100644
--- a/m4/.gitignore
+++ b/m4/.gitignore
@@ -59,3 +59,5 @@
 /manywarnings-c++.m4
 /gnu-make.m4
 /getopt.m4
+/pid_t.m4
+/zzgnulib.m4
diff --git a/providers/enchant_hunspell.cpp b/providers/enchant_hunspell.cpp
index 91fdce0..206e1e6 100644
--- a/providers/enchant_hunspell.cpp
+++ b/providers/enchant_hunspell.cpp
@@ -65,6 +65,7 @@ public:
 	bool checkWord (const char *word, size_t len);
 	char **suggestWord (const char* const word, size_t len, size_t *out_n_suggs);
 	const char *getWordchars ();
+	bool apostropheIsWordChar;
 
 	bool requestDictionary (const char * szLang);
 
@@ -83,7 +84,7 @@ g_iconv_is_valid(GIConv i)
 }
 
 HunspellChecker::HunspellChecker()
-: m_translate_in(nullptr), m_translate_out(nullptr), hunspell(nullptr)
+: apostropheIsWordChar(false), m_translate_in(nullptr), m_translate_out(nullptr), hunspell(nullptr)
 {
 }
 
@@ -114,7 +115,7 @@ HunspellChecker::checkWord(const char *utf8Word, size_t len)
 	if (static_cast<size_t>(-1) == result)
 		return false;
 	*out = '\0';
-	if (hunspell->spell(word8))
+	if (hunspell->spell(std::string(word8)))
 		return true;
 	else
 		return false;
@@ -141,29 +142,23 @@ HunspellChecker::suggestWord(const char* const utf8Word, size_t len, size_t *nsu
 		return nullptr;
 
 	*out = '\0';
-	char **sugMS;
-	*nsug = hunspell->suggest(&sugMS, word8);
+	std::vector<std::string> sugMS = hunspell->suggest(word8);
+	*nsug = sugMS.size();
 	if (*nsug > 0) {
 		char **sug = g_new0 (char *, *nsug + 1);
 		for (size_t i=0; i<*nsug; i++) {
-			in = sugMS[i];
+			in = const_cast<char *>(sugMS[i].c_str());
 			len_in = strlen(in);
 			len_out = MAXWORDLEN;
 			char *word = g_new0(char, len_out + 1);
 			out = word;
 			if (static_cast<size_t>(-1) == g_iconv(m_translate_out, &in, &len_in, &out, &len_out)) {
-				for (size_t j = i; j < *nsug; j++)
-					free(sugMS[j]);
-				free(sugMS);
-
 				*nsug = i;
-				return sug;
+				break;
 			}
-			*(out) = 0;
+			*out = '\0';
 			sug[i] = word;
-			free(sugMS[i]);
 		}
-		free(sugMS);
 		return sug;
 	}
 	else
@@ -328,6 +323,10 @@ HunspellChecker::requestDictionary(const char *szLang)
 	m_translate_in = g_iconv_open(enc, "UTF-8");
 	m_translate_out = g_iconv_open("UTF-8", enc);
 
+	const char *word_chars = hunspell->get_wordchars();
+	apostropheIsWordChar = g_utf8_strchr(word_chars, -1, g_utf8_get_char("'")) ||
+		g_utf8_strchr(word_chars, -1, g_utf8_get_char("’"));
+
 	return true;
 }
 
@@ -364,8 +363,12 @@ hunspell_dict_get_extra_word_characters (EnchantDict *me)
 static int
 hunspell_dict_is_word_character (EnchantDict *me, uint32_t uc, size_t n)
 {
-	(void)n;
 	HunspellChecker * checker = static_cast<HunspellChecker *>(me->user_data);
+	/* Accept quote marks anywhere except at the end of a word, as per
+	   hunspell's textparser.cxx/TextParser::next_token */
+	if ((uc == g_utf8_get_char("'") || uc == g_utf8_get_char("’")) && checker->apostropheIsWordChar) {
+		return n < 2;
+	}
 	return g_unichar_isalpha(uc) || g_utf8_strchr(checker->getWordchars(), -1, uc);
 }
 
diff --git a/src/enchant.c b/src/enchant.c
index bf00d8a..579b6a0 100644
--- a/src/enchant.c
+++ b/src/enchant.c
@@ -1,7 +1,7 @@
 /* enchant
  * Copyright (C) 2003 Dom Lachowicz
  *               2007 Hannu Väisänen
- *               2016-2017 Reuben Thomas
+ *               2016-2020 Reuben Thomas
  *
  * This library is free software; you can redistribute it and/or
  * modify it under the terms of the GNU Lesser General Public
@@ -135,9 +135,6 @@ print_utf (const char * str)
 static void
 do_mode_a (EnchantDict * dict, GString * word, size_t start_pos, size_t lineCount, gboolean terse_mode)
 {
-	size_t n_suggs;
-	char ** suggs;	
-
 	if (word->len <= MIN_WORD_LENGTH || enchant_dict_check (dict, word->str, word->len) == 0) {
 		if (!terse_mode) {
 			if (lineCount)
@@ -145,27 +142,23 @@ do_mode_a (EnchantDict * dict, GString * word, size_t start_pos, size_t lineCoun
 			else
 				printf ("*\n");
 		}
-	}
-	else {
-		suggs = enchant_dict_suggest (dict, word->str, 
-					      word->len, &n_suggs);
+	} else {
+		size_t n_suggs;
+		char ** suggs = enchant_dict_suggest (dict, word->str, word->len, &n_suggs);
 		if (!n_suggs || !suggs) {
 			printf ("# ");
 			if (lineCount)
 				printf ("%u ", (unsigned int)lineCount);
 			print_utf (word->str);
 			printf (" %u\n", (unsigned int)start_pos);
-		}
-		else {
-			size_t i = 0;
-			
+		} else {
 			printf ("& ");
 			if (lineCount)
 				printf ("%u ", (unsigned int)lineCount);
 			print_utf (word->str);
 			printf (" %u %u:", (unsigned int)n_suggs, (unsigned int)start_pos);
 			
-			for (i = 0; i < n_suggs; i++) {
+			for (size_t i = 0; i < n_suggs; i++) {
 				putchar (' ');
 				print_utf (suggs[i]);
 
diff --git a/src/lib.c b/src/lib.c
index dae11d9..5a5cb5a 100644
--- a/src/lib.c
+++ b/src/lib.c
@@ -1358,7 +1358,7 @@ enchant_dict_is_word_character (EnchantDict * dict, uint32_t uc_in, size_t n)
 	case G_UNICODE_TITLECASE_LETTER:
 	case G_UNICODE_UPPERCASE_LETTER:
 	case G_UNICODE_OTHER_LETTER:
-	case G_UNICODE_COMBINING_MARK: /* Older name for G_UNICODE_SPACING_MARK; deprecated since glib 2.30 */
+	case G_UNICODE_SPACING_MARK:
 	case G_UNICODE_ENCLOSING_MARK:
 	case G_UNICODE_NON_SPACING_MARK:
 	case G_UNICODE_DECIMAL_NUMBER:
diff --git a/src/pwl.c b/src/pwl.c
index ecd7eed..4f118a1 100644
--- a/src/pwl.c
+++ b/src/pwl.c
@@ -461,7 +461,7 @@ static int enchant_is_all_caps(const char*const word, size_t len)
 				case G_UNICODE_SURROGATE:
 				case G_UNICODE_MODIFIER_LETTER:
 				case G_UNICODE_OTHER_LETTER:
-				case G_UNICODE_COMBINING_MARK:
+				case G_UNICODE_SPACING_MARK:
 				case G_UNICODE_ENCLOSING_MARK:
 				case G_UNICODE_NON_SPACING_MARK:
 				case G_UNICODE_DECIMAL_NUMBER:
diff --git a/tests/asan-suppressions.txt b/tests/asan-suppressions.txt
index 89dc7cd..f066ce2 100644
--- a/tests/asan-suppressions.txt
+++ b/tests/asan-suppressions.txt
@@ -3,3 +3,4 @@ leak:getdelim
 
 # gnulib
 leak:set_this_relocation_prefix
+leak:set_relocation_prefix
diff --git a/tests/enchant_providers/providers.test.cpp b/tests/enchant_providers/providers.test.cpp
index 16f7f49..7d72faf 100644
--- a/tests/enchant_providers/providers.test.cpp
+++ b/tests/enchant_providers/providers.test.cpp
@@ -95,36 +95,29 @@ int Test(char* path)
 

 int TestProvidersInDirectory(char * dir_name)

 {

-    GDir *dir;

-	G_CONST_RETURN char *dir_entry;

-	size_t entry_len, g_module_suffix_len;

+    GDir *dir = g_dir_open (dir_name, 0, NULL);

+    if (!dir) 

+        return 0;

 	

-	char * filename;

-    int result = 0;

-	

-	dir = g_dir_open (dir_name, 0, NULL);

-	if (!dir) 

-		return 0;

-	

-	g_module_suffix_len = strlen (G_MODULE_SUFFIX);

+    size_t g_module_suffix_len = strlen (G_MODULE_SUFFIX);

 

-	while ((dir_entry = g_dir_read_name (dir)) != NULL)

-		{

-			entry_len = strlen (dir_entry);

-			if ((entry_len > g_module_suffix_len) && 

-				!strcmp(dir_entry+(entry_len-g_module_suffix_len), G_MODULE_SUFFIX))

-				{

-					filename = g_build_filename (dir_name, dir_entry, NULL);

-                                        int resultT = Test(filename);

-                                        if(resultT != 0)

-                                            {

-                                                result = resultT;

-                                            }

-                                        g_free (filename);

-				}

-		}

-	

-	g_dir_close (dir);

+    int result = 0;

+    const char *dir_entry;

+    while ((dir_entry = g_dir_read_name (dir)) != NULL)

+        {

+            size_t entry_len = strlen (dir_entry);

+            if ((entry_len > g_module_suffix_len) && 

+                !strcmp(dir_entry + (entry_len - g_module_suffix_len), G_MODULE_SUFFIX))

+                {

+                    char *filename = g_build_filename (dir_name, dir_entry, NULL);

+                    int resultT = Test(filename);

+                    if(resultT != 0)

+                        result = resultT;

+                    g_free (filename);

+                }

+        }

+    

+    g_dir_close (dir);

     return result;

 }