merge revision(s) 58745,58780,59040,60743: [Backport #13863]

rb_w32_ugetcwd: UTF-8 version getcwd

	* dir.c (rb_dir_getwd): convert from UTF-8.

	* win32/win32.c (w32_getcwd): codepage aware getcwd using
	  GetCurrentDirectoryW.
	potential memory leak

	* dir.c (rb_dir_getwd): get rid of potential memory leak.

	* util.c (ruby_getcwd): ditto.
	file.c: realpath in OS path encoding

	* dir.c (rb_dir_getwd_ospath): return cwd path in the OS path
	  encoding.

	* file.c (rb_realpath_internal): work in the OS path encoding
	load.c: cwd encoding

	* load.c (rb_get_expanded_load_path): save cwd cache in OS path
	  encoding, to get rid of unnecessary conversion and infinite
	  loading when it needs encoding conversion.
	  [ruby-dev:50221] [Bug #13863]

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_4@62784 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

dir.c

@@ -74,6 +74,7 @@ char *strchr(char*,char);
 #define rmdir(p) rb_w32_urmdir(p)
 #undef opendir
 #define opendir(p) rb_w32_uopendir(p)
+#define ruby_getcwd() rb_w32_ugetcwd(NULL, 0)
 #define IS_WIN32 1
 #else
 #define IS_WIN32 0
@@ -1048,26 +1049,52 @@ dir_s_chdir(int argc, VALUE *argv, VALUE obj)
 }
 
 VALUE
-rb_dir_getwd(void)
+rb_dir_getwd_ospath(void)
 {
     char *path;
     VALUE cwd;
-    int fsenc = rb_enc_to_index(rb_filesystem_encoding());
+    VALUE path_guard;
 
-    if (fsenc == ENCINDEX_US_ASCII) fsenc = ENCINDEX_ASCII;
+#undef RUBY_UNTYPED_DATA_WARNING
+#define RUBY_UNTYPED_DATA_WARNING 0
+    path_guard = Data_Wrap_Struct((VALUE)0, NULL, RUBY_DEFAULT_FREE, NULL);
     path = my_getcwd();
-#ifdef __APPLE__
+    DATA_PTR(path_guard) = path;
+#ifdef _WIN32
+    cwd = rb_utf8_str_new_cstr(path);
+    OBJ_TAINT(cwd);
+#elif defined __APPLE__
     cwd = rb_str_normalize_ospath(path, strlen(path));
     OBJ_TAINT(cwd);
 #else
     cwd = rb_tainted_str_new2(path);
 #endif
-    rb_enc_associate_index(cwd, fsenc);
+    DATA_PTR(path_guard) = 0;
 
     xfree(path);
     return cwd;
 }
 
+VALUE
+rb_dir_getwd(void)
+{
+    rb_encoding *fs = rb_filesystem_encoding();
+    int fsenc = rb_enc_to_index(fs);
+    VALUE cwd = rb_dir_getwd_ospath();
+
+    switch (fsenc) {
+      case ENCINDEX_US_ASCII:
+	fsenc = ENCINDEX_ASCII;
+      case ENCINDEX_ASCII:
+	break;
+#if defined _WIN32 || defined __APPLE__
+      default:
+	return rb_str_conv_enc(cwd, NULL, fs);
+#endif
+    }
+    return rb_enc_associate_index(cwd, fsenc);
+}
+
 /*
  *  call-seq:
  *     Dir.getwd -> string

file.c

@@ -126,6 +126,14 @@ int flock(int, int);
 #define STAT(p, s)	stat((p), (s))
 #endif
 
+#if defined _WIN32 || defined __APPLE__
+# define USE_OSPATH 1
+# define TO_OSPATH(str) rb_str_encode_ospath(str)
+#else
+# define USE_OSPATH 0
+# define TO_OSPATH(str) (str)
+#endif
+
 VALUE rb_cFile;
 VALUE rb_mFileTest;
 VALUE rb_cStat;
@@ -222,7 +230,7 @@ rb_get_path(VALUE obj)
 VALUE
 rb_str_encode_ospath(VALUE path)
 {
-#if defined _WIN32 || defined __APPLE__
+#if USE_OSPATH
     int encidx = ENCODING_GET(path);
 #ifdef _WIN32
     if (encidx == ENCINDEX_ASCII) {
@@ -3833,11 +3841,10 @@ realpath_rec(long *prefixlenp, VALUE *resolvedp, const char *unresolved,
             else {
                 struct stat sbuf;
                 int ret;
-                VALUE testpath2 = rb_str_encode_ospath(testpath);
 #ifdef __native_client__
-                ret = stat(RSTRING_PTR(testpath2), &sbuf);
+                ret = stat(RSTRING_PTR(testpath), &sbuf);
 #else
-                ret = lstat(RSTRING_PTR(testpath2), &sbuf);
+                ret = lstat(RSTRING_PTR(testpath), &sbuf);
 #endif
                 if (ret == -1) {
 		    int e = errno;
@@ -3909,9 +3916,12 @@ rb_check_realpath_internal(VALUE basedir, VALUE path, enum rb_realpath_mode mode
 
     if (!NIL_P(basedir)) {
         FilePathValue(basedir);
-        basedir = rb_str_dup_frozen(basedir);
+        basedir = TO_OSPATH(rb_str_dup_frozen(basedir));
     }
 
+    enc = rb_enc_get(unresolved_path);
+    origenc = enc;
+    unresolved_path = TO_OSPATH(unresolved_path);
     RSTRING_GETMEM(unresolved_path, ptr, len);
     path_names = skipprefixroot(ptr, ptr + len, rb_enc_get(unresolved_path));
     if (ptr != path_names) {
@@ -3928,7 +3938,7 @@ rb_check_realpath_internal(VALUE basedir, VALUE path, enum rb_realpath_mode mode
         }
     }
 
-    curdir = rb_dir_getwd();
+    curdir = rb_dir_getwd_ospath();
     RSTRING_GETMEM(curdir, ptr, len);
     curdir_names = skipprefixroot(ptr, ptr + len, rb_enc_get(curdir));
     resolved = rb_str_subseq(curdir, 0, curdir_names - ptr);
@@ -3936,7 +3946,6 @@ rb_check_realpath_internal(VALUE basedir, VALUE path, enum rb_realpath_mode mode
   root_found:
     RSTRING_GETMEM(resolved, prefixptr, prefixlen);
     pend = prefixptr + prefixlen;
-    enc = rb_enc_get(resolved);
     ptr = chompdirsep(prefixptr, pend, enc);
     if (ptr < pend) {
         prefixlen = ++ptr - prefixptr;
@@ -3951,7 +3960,6 @@ rb_check_realpath_internal(VALUE basedir, VALUE path, enum rb_realpath_mode mode
     }
 #endif
 
-    origenc = enc;
     switch (rb_enc_to_index(enc)) {
       case ENCINDEX_ASCII:
       case ENCINDEX_US_ASCII:
@@ -3970,8 +3978,14 @@ rb_check_realpath_internal(VALUE basedir, VALUE path, enum rb_realpath_mode mode
     if (realpath_rec(&prefixlen, &resolved, path_names, loopcheck, mode, 1))
 	return Qnil;
 
-    if (origenc != enc && rb_enc_str_asciionly_p(resolved))
+    if (origenc != rb_enc_get(resolved)) {
-	rb_enc_associate(resolved, origenc);
+	if (rb_enc_str_asciionly_p(resolved)) {
+	    rb_enc_associate(resolved, origenc);
+	}
+	else {
+	    resolved = rb_str_conv_enc(resolved, NULL, origenc);
+	}
+    }
 
     OBJ_TAINT(resolved);
     return resolved;

internal.h

@@ -982,6 +982,9 @@ void ruby_register_rollback_func_for_ensure(VALUE (*ensure_func)(ANYARGS), VALUE
 /* debug.c */
 PRINTF_ARGS(void ruby_debug_printf(const char*, ...), 1, 2);
 
+/* dir.c */
+VALUE rb_dir_getwd_ospath(void);
+
 /* dmyext.c */
 void Init_enc(void);
 void Init_ext(void);

load.c

@@ -95,15 +95,6 @@ rb_construct_expanded_load_path(enum expand_type type, int *has_relative, int *h
     rb_ary_replace(vm->load_path_snapshot, vm->load_path);
 }
 
-static VALUE
-load_path_getcwd(void)
-{
-    char *cwd = my_getcwd();
-    VALUE cwd_str = rb_filesystem_str_new_cstr(cwd);
-    xfree(cwd);
-    return cwd_str;
-}
-
 VALUE
 rb_get_expanded_load_path(void)
 {
@@ -115,7 +106,7 @@ rb_get_expanded_load_path(void)
 	int has_relative = 0, has_non_cache = 0;
 	rb_construct_expanded_load_path(EXPAND_ALL, &has_relative, &has_non_cache);
 	if (has_relative) {
-	    vm->load_path_check_cache = load_path_getcwd();
+	    vm->load_path_check_cache = rb_dir_getwd_ospath();
 	}
 	else if (has_non_cache) {
 	    /* Non string object. */
@@ -133,7 +124,7 @@ rb_get_expanded_load_path(void)
     }
     else if (vm->load_path_check_cache) {
 	int has_relative = 1, has_non_cache = 1;
-	VALUE cwd = load_path_getcwd();
+	VALUE cwd = rb_dir_getwd_ospath();
 	if (!rb_str_equal(vm->load_path_check_cache, cwd)) {
 	    /* Current working directory or filesystem encoding was changed.
 	       Expand relative load path and non-cacheable objects again. */

test/ruby/test_rubyoptions.rb

@@ -920,4 +920,16 @@ class TestRubyOptions < Test::Unit::TestCase
       end
     end
   end
+
+  def test_cwd_encoding
+    with_tmpchdir do
+      testdir = "\u30c6\u30b9\u30c8"
+      Dir.mkdir(testdir)
+      Dir.chdir(testdir) do
+        File.write("a.rb", "require './b'")
+        File.write("b.rb", "puts 'ok'")
+        assert_ruby_status([{"RUBYLIB"=>"."}, *%w[-E cp932:utf-8 a.rb]])
+      end
+    end
+  end
 end

util.c

@@ -511,7 +511,10 @@ ruby_getcwd(void)
     char *buf = xmalloc(2);
     strcpy(buf, ".");
 #elif defined HAVE_GETCWD
+# undef RUBY_UNTYPED_DATA_WARNING
+# define RUBY_UNTYPED_DATA_WARNING 0
 # if defined NO_GETCWD_MALLOC
+    VALUE guard = Data_Wrap_Struct((VALUE)0, NULL, RUBY_DEFAULT_FREE, NULL);
     int size = 200;
     char *buf = xmalloc(size);
 
@@ -519,17 +522,22 @@ ruby_getcwd(void)
 	int e = errno;
 	if (e != ERANGE) {
 	    xfree(buf);
+	    DATA_PTR(guard) = NULL;
 	    rb_syserr_fail(e, "getcwd");
 	}
 	size *= 2;
+	DATA_PTR(guard) = buf;
 	buf = xrealloc(buf, size);
     }
 # else
+    VALUE guard = Data_Wrap_Struct((VALUE)0, NULL, free, NULL);
     char *buf, *cwd = getcwd(NULL, 0);
+    DATA_PTR(guard) = cwd;
     if (!cwd) rb_sys_fail("getcwd");
     buf = ruby_strdup(cwd);	/* allocate by xmalloc */
     free(cwd);
 # endif
+    DATA_PTR(RB_GC_GUARD(guard)) = NULL;
 #else
 # ifndef PATH_MAX
 #  define PATH_MAX 8192

version.h

@@ -1,6 +1,6 @@
 #define RUBY_VERSION "2.4.4"
 #define RUBY_RELEASE_DATE "2018-03-17"
-#define RUBY_PATCHLEVEL 259
+#define RUBY_PATCHLEVEL 260
 
 #define RUBY_RELEASE_YEAR 2018
 #define RUBY_RELEASE_MONTH 3

win32/dir.h

@@ -33,6 +33,7 @@ long           rb_w32_telldir(DIR *);
 void           rb_w32_seekdir(DIR *, long);
 void           rb_w32_rewinddir(DIR *);
 void           rb_w32_closedir(DIR *);
+char          *rb_w32_ugetcwd(char *, int);
 
 #define opendir(s)   rb_w32_opendir((s))
 #define readdir(d)   rb_w32_readdir((d), 0)

win32/win32.c

@@ -4654,43 +4654,61 @@ clock_getres(clockid_t clock_id, struct timespec *sp)
 }
 
 /* License: Ruby's */
-char *
+static char *
-rb_w32_getcwd(char *buffer, int size)
+w32_getcwd(char *buffer, int size, UINT cp)
 {
-    char *p = buffer;
+    WCHAR *p;
-    int len;
+    int wlen, len;
 
-    len = GetCurrentDirectory(0, NULL);
+    len = GetCurrentDirectoryW(0, NULL);
     if (!len) {
 	errno = map_errno(GetLastError());
 	return NULL;
     }
 
-    if (p) {
+    if (buffer && size < len) {
+	errno = ERANGE;
+	return NULL;
+    }
+
+    p = ALLOCA_N(WCHAR, len);
+    if (!GetCurrentDirectoryW(len, p)) {
+	errno = map_errno(GetLastError());
+        return NULL;
+    }
+
+    wlen = translate_wchar(p, L'\\', L'/') - p + 1;
+    len = WideCharToMultiByte(cp, 0, p, wlen, NULL, 0, NULL, NULL);
+    if (buffer) {
 	if (size < len) {
 	    errno = ERANGE;
 	    return NULL;
 	}
     }
     else {
-	p = malloc(len);
+	buffer = malloc(len);
-	size = len;
+	if (!buffer) {
-	if (!p) {
 	    errno = ENOMEM;
 	    return NULL;
 	}
     }
+    WideCharToMultiByte(cp, 0, p, wlen, buffer, len, NULL, NULL);
 
-    if (!GetCurrentDirectory(size, p)) {
+    return buffer;
-	errno = map_errno(GetLastError());
+}
-	if (!buffer)
-	    free(p);
-        return NULL;
-    }
 
-    translate_char(p, '\\', '/', filecp());
+/* License: Ruby's */
+char *
+rb_w32_getcwd(char *buffer, int size)
+{
+    return w32_getcwd(buffer, size, filecp());
+}
 
-    return p;
+/* License: Ruby's */
+char *
+rb_w32_ugetcwd(char *buffer, int size)
+{
+    return w32_getcwd(buffer, size, CP_UTF8);
 }
 
 /* License: Artistic or GPL */