Simplify char16_t implementation

5 files changed, 43 insertions, 228 deletions

diff --git a/wcsmbs/c16rtomb.c b/wcsmbs/c16rtomb.c
index 3fed0b5..5374c75 100644
--- a/wcsmbs/c16rtomb.c
+++ b/wcsmbs/c16rtomb.c
@@ -17,25 +17,8 @@
    Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
    02111-1307 USA.  */
 
-#include <assert.h>
-#include <dlfcn.h>
-#include <errno.h>
-#include <gconv.h>
-#include <stdlib.h>
 #include <uchar.h>
-#include <wcsmbsload.h>
-
-#include <sysdep.h>
-
-#ifndef EILSEQ
-# define EILSEQ EINVAL
-#endif
-
-#if __STDC__ >= 201000L
-# define u(c) U##c
-#else
-# define u(c) L##c
-#endif
+#include <wchar.h>
 
 
 /* This is the private state used if PS is NULL.  */
@@ -44,85 +27,7 @@ static mbstate_t state;
 size_t
 c16rtomb (char *s, char16_t c16, mbstate_t *ps)
 {
-#if 1
   // XXX The ISO C 11 spec I have does not say anything about handling
   // XXX surrogates in this interface.
   return wcrtomb (s, c16, ps ?: &state);
-#else
-  char buf[MB_LEN_MAX];
-  struct __gconv_step_data data;
-  int status;
-  size_t result;
-  size_t dummy;
-  const struct gconv_fcts *fcts;
-
-  /* Set information for this step.  */
-  data.__invocation_counter = 0;
-  data.__internal_use = 1;
-  data.__flags = __GCONV_IS_LAST;
-  data.__statep = ps ?: &state;
-  data.__trans = NULL;
-
-  /* A first special case is if S is NULL.  This means put PS in the
-     initial state.  */
-  if (s == NULL)
-    {
-      s = buf;
-      c16 = u('\0');
-    }
-
-  /* Tell where we want to have the result.  */
-  data.__outbuf = (unsigned char *) s;
-  data.__outbufend = (unsigned char *) s + MB_CUR_MAX;
-
-  /* Get the conversion functions.  */
-  fcts = get_gconv_fcts (_NL_CURRENT_DATA (LC_CTYPE));
-  __gconv_fct fct = fcts->fromc16->__fct;
-#ifdef PTR_DEMANGLE
-  if (fcts->tomb->__shlib_handle != NULL)
-    PTR_DEMANGLE (fct);
-#endif
-
-  /* If C16 is the NUL character we write into the output buffer
-     the byte sequence necessary for PS to get into the initial
-     state, followed by a NUL byte.  */
-  if (c16 == L'\0')
-    {
-      status = DL_CALL_FCT (fct, (fcts->fromc16, &data, NULL, NULL,
-				  NULL, &dummy, 1, 1));
-
-      if (status == __GCONV_OK || status == __GCONV_EMPTY_INPUT)
-	*data.__outbuf++ = '\0';
-    }
-  else
-    {
-      /* Do a normal conversion.  */
-      const unsigned char *inbuf = (const unsigned char *) &c16;
-
-      status = DL_CALL_FCT (fct,
-			    (fcts->fromc16, &data, &inbuf,
-			     inbuf + sizeof (char16_t), NULL, &dummy,
-			     0, 1));
-    }
-
-  /* There must not be any problems with the conversion but illegal input
-     characters.  The output buffer must be large enough, otherwise the
-     definition of MB_CUR_MAX is not correct.  All the other possible
-     errors also must not happen.  */
-  assert (status == __GCONV_OK || status == __GCONV_EMPTY_INPUT
-	  || status == __GCONV_ILLEGAL_INPUT
-	  || status == __GCONV_INCOMPLETE_INPUT
-	  || status == __GCONV_FULL_OUTPUT);
-
-  if (status == __GCONV_OK || status == __GCONV_EMPTY_INPUT
-      || status == __GCONV_FULL_OUTPUT)
-    result = data.__outbuf - (unsigned char *) s;
-  else
-    {
-      result = (size_t) -1;
-      __set_errno (EILSEQ);
-    }
-
-  return result;
-#endif
 }
diff --git a/wcsmbs/mbrtoc16.c b/wcsmbs/mbrtoc16.c
index df970fb..f5ed2b4 100644
--- a/wcsmbs/mbrtoc16.c
+++ b/wcsmbs/mbrtoc16.c
@@ -30,12 +30,6 @@
 # define EILSEQ EINVAL
 #endif
 
-#if __STDC__ >= 201000L
-# define U(c) U##c
-#else
-# define U(c) L##c
-#endif
-
 
 /* This is the private state used if PS is NULL.  */
 static mbstate_t state;
@@ -46,6 +40,11 @@ mbrtoc16 (char16_t *pc16, const char *s, size_t n, mbstate_t *ps)
   if (ps == NULL)
     ps = &state;
 
+  /* The standard text does not say that S being NULL means the state
+     is reset even if the second half of a surrogate still have to be
+     returned.  In fact, the error code description indicates
+     otherwise.  Therefore always first try to return a second
+     half.  */
   if (ps->__count & 0x80000000)
     {
       /* We have to return the second word for a surrogate.  */
@@ -55,13 +54,13 @@ mbrtoc16 (char16_t *pc16, const char *s, size_t n, mbstate_t *ps)
       return (size_t) -3;
     }
 
-  char16_t buf[2];
+  wchar_t wc;
   struct __gconv_step_data data;
   int status;
   size_t result;
   size_t dummy;
   const unsigned char *inbuf, *endbuf;
-  unsigned char *outbuf = (unsigned char *) buf;
+  unsigned char *outbuf = (unsigned char *) &wc;
   const struct gconv_fcts *fcts;
 
   /* Set information for this step.  */
@@ -75,14 +74,14 @@ mbrtoc16 (char16_t *pc16, const char *s, size_t n, mbstate_t *ps)
      initial state.  */
   if (s == NULL)
     {
-      outbuf = (unsigned char *) buf;
+      pc16 = NULL;
       s = "";
       n = 1;
     }
 
   /* Tell where we want the result.  */
   data.__outbuf = outbuf;
-  data.__outbufend = outbuf + sizeof (char16_t);
+  data.__outbufend = outbuf + sizeof (wchar_t);
 
   /* Get the conversion functions.  */
   fcts = get_gconv_fcts (_NL_CURRENT_DATA (LC_CTYPE));
@@ -91,28 +90,20 @@ mbrtoc16 (char16_t *pc16, const char *s, size_t n, mbstate_t *ps)
   inbuf = (const unsigned char *) s;
   endbuf = inbuf + n;
   if (__builtin_expect (endbuf < inbuf, 0))
-    endbuf = (const unsigned char *) ~(uintptr_t) 0;
-  __gconv_fct fct = fcts->toc16->__fct;
+    {
+      endbuf = (const unsigned char *) ~(uintptr_t) 0;
+      if (endbuf == inbuf)
+	goto ilseq;
+    }
+  __gconv_fct fct = fcts->towc->__fct;
 #ifdef PTR_DEMANGLE
-  if (fcts->toc16->__shlib_handle != NULL)
+  if (fcts->towc->__shlib_handle != NULL)
     PTR_DEMANGLE (fct);
 #endif
 
-  /* We first have to check whether the character can be represented
-     without a surrogate.  If we immediately pass in a buffer large
-     enough to hold two char16_t values and the first character does
-     not require a surrogate the routine will try to convert more
-     input if N is larger then needed for the first character.  */
-  status = DL_CALL_FCT (fct, (fcts->toc16, &data, &inbuf, endbuf,
+  status = DL_CALL_FCT (fct, (fcts->towc, &data, &inbuf, endbuf,
 			      NULL, &dummy, 0, 1));
 
-  if (status == __GCONV_FULL_OUTPUT && data.__outbuf == outbuf)
-    {
-      data.__outbufend = outbuf + 2 * sizeof (char16_t);
-      status = DL_CALL_FCT (fct, (fcts->toc16, &data, &inbuf, endbuf,
-				  NULL, &dummy, 0, 1));
-    }
-
   /* There must not be any problems with the conversion but illegal input
      characters.  The output buffer must be large enough, otherwise the
      definition of MB_CUR_MAX is not correct.  All the other possible
@@ -125,33 +116,35 @@ mbrtoc16 (char16_t *pc16, const char *s, size_t n, mbstate_t *ps)
   if (status == __GCONV_OK || status == __GCONV_EMPTY_INPUT
       || status == __GCONV_FULL_OUTPUT)
     {
-      if (pc16 != NULL)
-	*pc16 = buf[0];
+      result = inbuf - (const unsigned char *) s;
 
-      if (data.__outbuf != outbuf && *(char16_t *) outbuf == U('\0'))
+      if (wc < 0x10000)
 	{
-	  /* The converted character is the NUL character.  */
-	  assert (__mbsinit (data.__statep));
-	  result = 0;
+	  if (pc16 != NULL)
+	    *pc16 = wc;
+
+	  if (data.__outbuf != outbuf && wc == L'\0')
+	    {
+	      /* The converted character is the NUL character.  */
+	      assert (__mbsinit (data.__statep));
+	      result = 0;
+	    }
 	}
       else
 	{
-	  result = inbuf - (const unsigned char *) s;
+	  /* This is a surrogate.  */
+	  if (pc16 != NULL)
+	    *pc16 = 0xd7c0 + (wc >> 10);
 
-	  if (data.__outbuf != outbuf + 2)
-	    {
-	      /* This is a surrogate.  */
-	      assert (buf[0] >= 0xd800 && buf[0] <= 0xdfff);
-	      assert (buf[1] >= 0xdc00 && buf[1] <= 0xdfff);
-	      ps->__count |= 0x80000000;
-	      ps->__value.__wch = buf[1];
-	    }
+	  ps->__count |= 0x80000000;
+	  ps->__value.__wch = 0xdc00 + (wc & 0x3ff);
 	}
     }
   else if (status == __GCONV_INCOMPLETE_INPUT)
     result = (size_t) -2;
   else
     {
+    ilseq:
       result = (size_t) -1;
       __set_errno (EILSEQ);
     }
diff --git a/wcsmbs/mbrtowc.c b/wcsmbs/mbrtowc.c
index 03b8348..0c99b74 100644
--- a/wcsmbs/mbrtowc.c
+++ b/wcsmbs/mbrtowc.c
@@ -1,4 +1,4 @@
-/* Copyright (C) 1996, 1997, 1998, 1999, 2000, 2002, 2004, 2005, 2011
+/* Copyright (C) 1996, 1997, 1998, 1999, 2000, 2002, 2004, 2005, 2011, 2012
    Free Software Foundation, Inc.
    This file is part of the GNU C Library.
    Contributed by Ulrich Drepper <drepper@gnu.org>, 1996.
@@ -73,7 +73,11 @@ __mbrtowc (wchar_t *pwc, const char *s, size_t n, mbstate_t *ps)
   inbuf = (const unsigned char *) s;
   endbuf = inbuf + n;
   if (__builtin_expect (endbuf < inbuf, 0))
-    endbuf = (const unsigned char *) ~(uintptr_t) 0;
+    {
+      endbuf = (const unsigned char *) ~(uintptr_t) 0;
+      if (endbuf == inbuf)
+	goto ilseq;
+    }
   __gconv_fct fct = fcts->towc->__fct;
 #ifdef PTR_DEMANGLE
   if (fcts->towc->__shlib_handle != NULL)
@@ -108,6 +112,7 @@ __mbrtowc (wchar_t *pwc, const char *s, size_t n, mbstate_t *ps)
     result = (size_t) -2;
   else
     {
+    ilseq:
       result = (size_t) -1;
       __set_errno (EILSEQ);
     }
diff --git a/wcsmbs/wcsmbsload.c b/wcsmbs/wcsmbsload.c
index 9ce26f1..27ea442 100644
--- a/wcsmbs/wcsmbsload.c
+++ b/wcsmbs/wcsmbsload.c
@@ -68,44 +68,6 @@ static const struct __gconv_step to_mb =
   .__data = NULL
 };
 
-static const struct __gconv_step to_c16 =
-{
-  .__shlib_handle = NULL,
-  .__modname = NULL,
-  .__counter = INT_MAX,
-  .__from_name = (char *) "ANSI_X3.4-1968//TRANSLIT",
-  .__to_name = (char *) "UTF-16//",
-  .__fct = __gconv_transform_ascii_char16,
-  .__btowc_fct = NULL,
-  .__init_fct = NULL,
-  .__end_fct = NULL,
-  .__min_needed_from = 1,
-  .__max_needed_from = 1,
-  .__min_needed_to = 4,
-  .__max_needed_to = 4,
-  .__stateful = 0,
-  .__data = NULL
-};
-
-static const struct __gconv_step from_c16 =
-{
-  .__shlib_handle = NULL,
-  .__modname = NULL,
-  .__counter = INT_MAX,
-  .__from_name = (char *) "UTF-16//",
-  .__to_name = (char *) "ANSI_X3.4-1968//TRANSLIT",
-  .__fct = __gconv_transform_char16_ascii,
-  .__btowc_fct = NULL,
-  .__init_fct = NULL,
-  .__end_fct = NULL,
-  .__min_needed_from = 4,
-  .__max_needed_from = 4,
-  .__min_needed_to = 1,
-  .__max_needed_to = 1,
-  .__stateful = 0,
-  .__data = NULL
-};
-
 
 /* For the default locale we only have to handle ANSI_X3.4-1968.  */
 const struct gconv_fcts __wcsmbs_gconv_fcts_c =
@@ -114,11 +76,6 @@ const struct gconv_fcts __wcsmbs_gconv_fcts_c =
   .towc_nsteps = 1,
   .tomb = (struct __gconv_step *) &to_mb,
   .tomb_nsteps = 1,
-
-  .toc16 = (struct __gconv_step *) &to_c16,
-  .toc16_nsteps = 1,
-  .fromc16 = (struct __gconv_step *) &from_c16,
-  .fromc16_nsteps = 1,
 };
 
 
@@ -234,24 +191,9 @@ __wcsmbs_load_conv (struct __locale_data *new_category)
 	new_fcts->tomb = __wcsmbs_getfct (complete_name, "INTERNAL",
 					  &new_fcts->tomb_nsteps);
 
-      if (new_fcts->tomb != NULL)
-	{
-	  new_fcts->toc16 = __wcsmbs_getfct ("CHAR16", complete_name,
-					     &new_fcts->toc16_nsteps);
-
-	  if (new_fcts->toc16 != NULL)
-	    new_fcts->fromc16 = __wcsmbs_getfct (complete_name, "CHAR16",
-						 &new_fcts->fromc16_nsteps);
-	  else
-	    {
-	      __gconv_close_transform (new_fcts->toc16, new_fcts->toc16_nsteps);
-	      new_fcts->toc16 = NULL;
-	    }
-	}
-
       /* If any of the conversion functions is not available we don't
 	 use any since this would mean we cannot convert back and
-	 forth.*/
+	 forth.  NB: NEW_FCTS was allocated with calloc.  */
       if (new_fcts->tomb == NULL)
 	{
 	  if (new_fcts->towc != NULL)
@@ -264,12 +206,6 @@ __wcsmbs_load_conv (struct __locale_data *new_category)
 	}
       else
 	{
-	  // XXX At least for now we live with the CHAR16 not being available.
-	  if (new_fcts->toc16 == NULL)
-	    new_fcts->toc16 = __wcsmbs_gconv_fcts_c.toc16;
-	  if (new_fcts->fromc16 == NULL)
-	    new_fcts->fromc16 = __wcsmbs_gconv_fcts_c.fromc16;
-
 	  new_category->private.ctype = new_fcts;
 	  new_category->private.cleanup = &_nl_cleanup_ctype;
 	}
@@ -297,10 +233,6 @@ __wcsmbs_clone_conv (struct gconv_fcts *copy)
     ++copy->towc->__counter;
   if (copy->tomb->__shlib_handle != NULL)
     ++copy->tomb->__counter;
-  if (copy->toc16->__shlib_handle != NULL)
-    ++copy->toc16->__counter;
-  if (copy->fromc16->__shlib_handle != NULL)
-    ++copy->fromc16->__counter;
 }
 
 
@@ -320,19 +252,6 @@ __wcsmbs_named_conv (struct gconv_fcts *copy, const char *name)
       return 1;
     }
 
-  copy->fromc16 = __wcsmbs_getfct (name, "CHAR16", &copy->fromc16_nsteps);
-  if (copy->fromc16 == NULL)
-    copy->toc16 = NULL;
-  else
-    {
-      copy->toc16 = __wcsmbs_getfct ("CHAR16", name, &copy->toc16_nsteps);
-      if (copy->toc16 == NULL)
-	{
-	  __gconv_close_transform (copy->fromc16, copy->fromc16_nsteps);
-	  copy->fromc16 = NULL;
-	}
-    }
-
   return 0;
 }
 
@@ -348,8 +267,6 @@ _nl_cleanup_ctype (struct __locale_data *locale)
       /* Free the old conversions.  */
       __gconv_close_transform (data->tomb, data->tomb_nsteps);
       __gconv_close_transform (data->towc, data->towc_nsteps);
-      __gconv_close_transform (data->fromc16, data->fromc16_nsteps);
-      __gconv_close_transform (data->toc16, data->toc16_nsteps);
       free ((char *) data);
     }
 }
diff --git a/wcsmbs/wcsmbsload.h b/wcsmbs/wcsmbsload.h
index 064c41c..98f53bc 100644
--- a/wcsmbs/wcsmbsload.h
+++ b/wcsmbs/wcsmbsload.h
@@ -32,11 +32,6 @@ struct gconv_fcts
     size_t towc_nsteps;
     struct __gconv_step *tomb;
     size_t tomb_nsteps;
-
-    struct __gconv_step *toc16;
-    size_t toc16_nsteps;
-    struct __gconv_step *fromc16;
-    size_t fromc16_nsteps;
   };
 
 /* Set of currently active conversion functions.  */