Merge branch 'wcstombs'

This topic branch fixes the problem where a UTF-16 command-line was
converted to UTF-8 in an incorrect way (because Cygwin treated it as if
it was a file name and applied some magic that is intended to allow for
otherwise invalid file names on Windows).

Signed-off-by: Johannes Schindelin <[email protected]>

Author: dscho

newlib/libc/locale/lctype.c

@@ -25,7 +25,12 @@
 
 #define LCCTYPE_SIZE (sizeof(struct lc_ctype_T) / sizeof(char *))
 
+#ifdef __CYGWIN__
+/* Cygwin uses __utf8_mbtowc() by default, therefore mb_cur_max := 6 */
+static char	numone[] = { '\x06', '\0'};
+#else
 static char	numone[] = { '\1', '\0'};
+#endif
 
 const struct lc_ctype_T _C_ctype_locale = {
 	"ASCII",			/* codeset */

winsup/cygwin/dcrt0.cc

@@ -912,9 +912,9 @@ dll_crt0_1 (void *)
   if (!__argc)
     {
       PWCHAR wline = GetCommandLineW ();
-      size_t size = sys_wcstombs_no_path (NULL, 0, wline) + 1;
+      size_t size = sys_wcstombs (NULL, 0, wline) + 1;
       char *line = (char *) alloca (size);
-      sys_wcstombs_no_path (line, size, wline);
+      sys_wcstombs (line, size, wline);
 
       /* Scan the command line and build argv.  Expand wildcards if not
 	 called from another cygwin process. */

winsup/cygwin/dtable.cc

@@ -1028,7 +1028,7 @@ handle_to_fn (HANDLE h, char *posix_fn)
       if (wcsncasecmp (w32, DEVICE_PREFIX, DEVICE_PREFIX_LEN) != 0
 	  || !QueryDosDeviceW (NULL, fnbuf, sizeof (fnbuf) / sizeof (WCHAR)))
 	{
-	  sys_wcstombs (posix_fn, NT_MAX_PATH, w32, w32len);
+	  sys_wcstombs_path (posix_fn, NT_MAX_PATH, w32, w32len);
 	  return false;
 	}
 

winsup/cygwin/environ.cc

@@ -918,7 +918,7 @@ win32env_to_cygenv (PWCHAR rawenv, bool posify)
      eventually want to use them).  */
   for (i = 0, w = rawenv; *w != L'\0'; w = wcschr (w, L'\0') + 1, i++)
     {
-      sys_wcstombs_alloc_no_path (&newp, HEAP_NOTHEAP, w);
+      sys_wcstombs_alloc (&newp, HEAP_NOTHEAP, w);
       if (i >= envc)
         envp = (char **) realloc (envp, (4 + (envc += 100)) * sizeof (char *));
       envp[i] = newp;
@@ -978,15 +978,15 @@ getwinenveq (const char *name, size_t namelen, int x)
   int totlen = GetEnvironmentVariableW (name0, valbuf, 32768);
   if (totlen > 0)
     {
-      totlen = sys_wcstombs_no_path (NULL, 0, valbuf) + 1;
+      totlen = sys_wcstombs (NULL, 0, valbuf) + 1;
       if (x == HEAP_1_STR)
 	totlen += namelen;
       else
 	namelen = 0;
       char *p = (char *) cmalloc_abort ((cygheap_types) x, totlen);
       if (namelen)
 	strcpy (p, name);
-      sys_wcstombs_no_path (p + namelen, totlen, valbuf);
+      sys_wcstombs (p + namelen, totlen, valbuf);
       debug_printf ("using value from GetEnvironmentVariable for '%W'", name0);
       return p;
     }
@@ -1144,7 +1144,7 @@ build_env (const char * const *envp, PWCHAR &envblock, int &envc,
 	  for (winnum = 0, var = cwinenv;
 	       *var;
 	       ++winnum, var = wcschr (var, L'\0') + 1)
-	    sys_wcstombs_alloc_no_path (&winenv[winnum], HEAP_NOTHEAP, var);
+	    sys_wcstombs_alloc (&winenv[winnum], HEAP_NOTHEAP, var);
 	}
       DestroyEnvironmentBlock (cwinenv);
       /* Eliminate variables which are already available in envp, as well as

winsup/cygwin/external.cc

@@ -92,7 +92,7 @@ fillout_pinfo (pid_t pid, int winpid)
 	  ep.rusage_self = p->rusage_self;
 	  ep.rusage_children = p->rusage_children;
 	  ep.progname[0] = '\0';
-	  sys_wcstombs(ep.progname, MAX_PATH, p->progname);
+	  sys_wcstombs_path (ep.progname, MAX_PATH, p->progname);
 	  ep.strace_mask = 0;
 	  ep.version = EXTERNAL_PINFO_VERSION;
 

winsup/cygwin/fhandler/disk_file.cc

@@ -2412,7 +2412,7 @@ fhandler_disk_file::readdir_helper (DIR *dir, dirent *de, DWORD w32_err,
 	  char *p = stpcpy (file, pc.get_posix ());
 	  if (p[-1] != '/')
 	    *p++ = '/';
-	  sys_wcstombs (p, NT_MAX_PATH - (p - file),
+	  sys_wcstombs_path (p, NT_MAX_PATH - (p - file),
 			fname->Buffer, fname->Length / sizeof (WCHAR));
 	  path_conv fpath (file, PC_SYM_NOFOLLOW);
 	  if (fpath.issymlink ())
@@ -2433,7 +2433,7 @@ fhandler_disk_file::readdir_helper (DIR *dir, dirent *de, DWORD w32_err,
 	}
     }
 
-  sys_wcstombs (de->d_name, NAME_MAX + 1, fname->Buffer,
+  sys_wcstombs_path (de->d_name, NAME_MAX + 1, fname->Buffer,
 		fname->Length / sizeof (WCHAR));
 
   /* Don't try to optimize relative to dir->__d_position.  On several

winsup/cygwin/fhandler/netdrive.cc

@@ -283,15 +283,15 @@ fhandler_netdrive::readdir (DIR *dir, dirent *de)
 	  tp.u_get (&ds);
 	  RtlInitUnicodeString (&ss, bs);
 	  RtlDowncaseUnicodeString (&ds, &ss, FALSE);
-	  sys_wcstombs (de->d_name, sizeof de->d_name,
+	  sys_wcstombs_path (de->d_name, sizeof de->d_name,
 			ds.Buffer, ds.Length / sizeof (WCHAR));
 	  de->d_ino = hash_path_name (get_ino (), de->d_name);
 	}
       else
 	{
-	  sys_wcstombs (de->d_name, sizeof de->d_name, bs);
+	  sys_wcstombs_path (de->d_name, sizeof de->d_name, bs);
 	  char *rpath = tp.c_get ();
-	  sys_wcstombs (rpath, NT_MAX_PATH, nro->lpRemoteName);
+	  sys_wcstombs_path (rpath, NT_MAX_PATH, nro->lpRemoteName);
 	  de->d_ino = readdir_get_ino (rpath, false);
 	  /* We can't trust remote inode numbers of only 32 bit.  That means,
 	     remote NT4 NTFS, as well as shares of Samba version < 3.0. */

winsup/cygwin/fhandler/process.cc

@@ -574,10 +574,10 @@ static off_t
 format_process_winexename (void *data, char *&destbuf)
 {
   _pinfo *p = (_pinfo *) data;
-  size_t len = sys_wcstombs (NULL, 0, p->progname);
+  size_t len = sys_wcstombs_path (NULL, 0, p->progname);
   destbuf = (char *) crealloc_abort (destbuf, len + 1);
   /* With trailing \0 for backward compat reasons. */
-  sys_wcstombs (destbuf, len + 1, p->progname);
+  sys_wcstombs_path (destbuf, len + 1, p->progname);
   return len;
 }
 
@@ -1051,7 +1051,7 @@ format_process_maps (void *data, char *&destbuf)
 		      drive_maps.fixup_if_match (msi->SectionFileName.Buffer);
 		  if (mount_table->conv_to_posix_path (dosname,
 						       posix_modname, 0))
-		    sys_wcstombs (posix_modname, NT_MAX_PATH, dosname);
+		    sys_wcstombs_path (posix_modname, NT_MAX_PATH, dosname);
 		  stat (posix_modname, &st);
 		}
 	      else if (!threads.fill_if_match (cur.abase, mb.Type,
@@ -1108,7 +1108,7 @@ format_process_stat (void *data, char *&destbuf)
   else
     {
       PWCHAR last_slash = wcsrchr (p->progname, L'\\');
-      sys_wcstombs (cmd, NAME_MAX + 1,
+      sys_wcstombs_path (cmd, NAME_MAX + 1,
 		    last_slash ? last_slash + 1 : p->progname);
       int len = strlen (cmd);
       if (len > 4)
@@ -1215,7 +1215,8 @@ format_process_status (void *data, char *&destbuf)
   bool fetch_siginfo = false;
 
   PWCHAR last_slash = wcsrchr (p->progname, L'\\');
-  sys_wcstombs (cmd, NAME_MAX + 1, last_slash ? last_slash + 1 : p->progname);
+  sys_wcstombs_path (cmd, NAME_MAX + 1,
+		  last_slash ? last_slash + 1 : p->progname);
   int len = strlen (cmd);
   if (len > 4)
     {

winsup/cygwin/fhandler/procsys.cc

@@ -236,10 +236,11 @@ fhandler_procsys::fill_filebuf ()
   NtClose (h);
   if (!NT_SUCCESS (status))
     goto unreadable;
-  len = sys_wcstombs (NULL, 0, target.Buffer, target.Length / sizeof (WCHAR));
+  len = sys_wcstombs_path (NULL, 0,
+		  target.Buffer, target.Length / sizeof (WCHAR));
   filebuf = (char *) crealloc_abort (filebuf, procsys_len + len + 1);
-  sys_wcstombs (fnamep = stpcpy (filebuf, procsys), len + 1, target.Buffer,
-		target.Length / sizeof (WCHAR));
+  sys_wcstombs_path (fnamep = stpcpy (filebuf, procsys), len + 1,
+		  target.Buffer, target.Length / sizeof (WCHAR));
   while ((fnamep = strchr (fnamep, '\\')))
     *fnamep = '/';
   return true;
@@ -377,8 +378,8 @@ fhandler_procsys::readdir (DIR *dir, dirent *de)
 	res = ENMFILE;
       else
 	{
-	  sys_wcstombs (de->d_name, NAME_MAX + 1, dbi->ObjectName.Buffer,
-			dbi->ObjectName.Length / sizeof (WCHAR));
+	  sys_wcstombs_path (de->d_name, NAME_MAX + 1, dbi->ObjectName.Buffer,
+			     dbi->ObjectName.Length / sizeof (WCHAR));
 	  de->d_ino = hash_path_name (get_ino (), de->d_name);
 	  if (RtlEqualUnicodeString (&dbi->ObjectTypeName, &ro_u_natdir, FALSE))
 	    de->d_type = DT_DIR;

winsup/cygwin/local_includes/wchar.h

@@ -173,29 +173,29 @@ extern size_t _sys_wcstombs_alloc (char **dst_p, int type, const wchar_t *src,
 				   size_t nwc, bool is_path);
 
 static inline size_t
-sys_wcstombs (char *dst, size_t len, const wchar_t * src,
-	      size_t nwc = (size_t) -1)
+sys_wcstombs_path (char *dst, size_t len, const wchar_t * src,
+		   size_t nwc = (size_t) -1)
 {
   return _sys_wcstombs (dst, len, src, nwc, true);
 }
 
 static inline size_t
-sys_wcstombs_no_path (char *dst, size_t len, const wchar_t * src,
-		      size_t nwc = (size_t) -1)
+sys_wcstombs (char *dst, size_t len, const wchar_t * src,
+	      size_t nwc = (size_t) -1)
 {
   return _sys_wcstombs (dst, len, src, nwc, false);
 }
 
 static inline size_t
-sys_wcstombs_alloc (char **dst_p, int type, const wchar_t *src,
-		    size_t nwc = (size_t) -1)
+sys_wcstombs_alloc_path (char **dst_p, int type, const wchar_t *src,
+			 size_t nwc = (size_t) -1)
 {
   return _sys_wcstombs_alloc (dst_p, type, src, nwc, true);
 }
 
 static inline size_t
-sys_wcstombs_alloc_no_path (char **dst_p, int type, const wchar_t *src,
-			    size_t nwc = (size_t) -1)
+sys_wcstombs_alloc (char **dst_p, int type, const wchar_t *src,
+		    size_t nwc = (size_t) -1)
 {
   return _sys_wcstombs_alloc (dst_p, type, src, nwc, false);
 }