Index: libcpp/ChangeLog

2005-03-12  Geoffrey Keating  <geoffk@apple.com>

	* directives.c (glue_header_name): Update call to cpp_spell_token.
	* internal.h (_cpp_interpret_identifier): New.
	* charset.c (_cpp_interpret_identifier): New.
	(_cpp_valid_ucn): Allow UCN version of '$'.
	* lex.c (lex_identifier): Add extra parameter to indicate if initial
	character was '$' or '\'.  Support identifiers with UCNs.
	(forms_identifier_p): Allow UCNs.
	(_cpp_lex_direct): Pass extra parameter to lex_identifier.
	(utf8_to_ucn): New.
	(cpp_spell_token): Add FORSTRING parameter.  Use it.
	(cpp_token_as_text): Update call to cpp_spell_token.
	(cpp_output_token): Write UCNs back out.
	(stringify_arg): Update call to cpp_spell_token.
	(paste_tokens): Likewise.
	(cpp_macro_definition): Likewise.
	* macro.c (stringify_arg): Likewise.
	(paste_tokens): Likewise.
	(cpp_macro_definition): Likewise.
	* include/cpplib.h: Add parameter to cpp_spell_token.

Index: gcc/ChangeLog
2005-03-12  Geoffrey Keating  <geoffk@apple.com>

	* c-lex.c (c_lex_with_flags): Add parameter to call to
	cpp_spell_token.

Index: gcc/testsuite/ChangeLog
2005-03-12  Geoffrey Keating  <geoffk@apple.com>

	* gcc.dg/ucnid-1.c: New.
	* gcc.dg/ucnid-2.c: New.
	* gcc.dg/ucnid-3.c: New.
	* gcc.dg/ucnid-4.c: New.
	* gcc.dg/ucnid-5.c: New.
	* gcc.dg/ucnid-6.c: New.
	* gcc.dg/cpp/ucnid-1.c: New.
	* gcc.dg/cpp/ucnid-2.c: New.
	* gcc.dg/cpp/ucnid-3.c: New.
	* g++.dg/other/ucnid-1.C: New.

From-SVN: r96333

7 files changed, 198 insertions, 42 deletions

diff --git a/libcpp/ChangeLog b/libcpp/ChangeLog
index b246de782da..5190599b9b0 100644
--- a/libcpp/ChangeLog
+++ b/libcpp/ChangeLog
@@ -1,3 +1,25 @@
+2005-03-11  Geoffrey Keating  <geoffk@apple.com>
+
+	* directives.c (glue_header_name): Update call to cpp_spell_token.
+	* internal.h (_cpp_interpret_identifier): New.
+	* charset.c (_cpp_interpret_identifier): New.
+	(_cpp_valid_ucn): Allow UCN version of '$'.
+	* lex.c (lex_identifier): Add extra parameter to indicate if initial
+	character was '$' or '\'.  Support identifiers with UCNs.
+	(forms_identifier_p): Allow UCNs.
+	(_cpp_lex_direct): Pass extra parameter to lex_identifier.
+	(utf8_to_ucn): New.
+	(cpp_spell_token): Add FORSTRING parameter.  Use it.
+	(cpp_token_as_text): Update call to cpp_spell_token.
+	(cpp_output_token): Write UCNs back out.
+	(stringify_arg): Update call to cpp_spell_token.
+	(paste_tokens): Likewise.
+	(cpp_macro_definition): Likewise.
+	* macro.c (stringify_arg): Likewise.
+	(paste_tokens): Likewise.
+	(cpp_macro_definition): Likewise.
+	* include/cpplib.h: Add parameter to cpp_spell_token.
+
 2005-03-04  Jakub Jelinek  <jakub@redhat.com>
 
 	PR bootstrap/20282
diff --git a/libcpp/charset.c b/libcpp/charset.c
index 6b6c360f73d..cd25f10a2e6 100644
--- a/libcpp/charset.c
+++ b/libcpp/charset.c
@@ -907,6 +907,15 @@ _cpp_valid_ucn (cpp_reader *pfile, const uchar **pstr,
 		 (int) (str - base), base);
       result = 1;
     }
+  else if (identifier_pos && result == 0x24 
+	   && CPP_OPTION (pfile, dollars_in_ident))
+    {
+      if (CPP_OPTION (pfile, warn_dollars) && !pfile->state.skipping)
+	{
+	  CPP_OPTION (pfile, warn_dollars) = 0;
+	  cpp_error (pfile, CPP_DL_PEDWARN, "'$' in identifier or number");
+	}
+    }
   else if (identifier_pos)
     {
       int validity = ucn_valid_in_identifier (pfile, result);
@@ -1414,7 +1423,60 @@ cpp_interpret_charconst (cpp_reader *pfile, const cpp_token *token,
 
   return result;
 }
+
+/* Convert an identifier denoted by ID and LEN, which might contain
+   UCN escapes, to the source character set, either UTF-8 or
+   UTF-EBCDIC.  Assumes that the identifier is actually a valid identifier.  */
+cpp_hashnode *
+_cpp_interpret_identifier (cpp_reader *pfile, const uchar *id, size_t len)
+{
+  /* It turns out that a UCN escape always turns into fewer characters
+     than the escape itself, so we can allocate a temporary in advance.  */
+  uchar * buf = alloca (len + 1);
+  uchar * bufp = buf;
+  size_t idp;
+  
+  for (idp = 0; idp < len; idp++)
+    if (id[idp] != '\\')
+      *bufp++ = id[idp];
+    else
+      {
+	unsigned length = id[idp+1] == 'u' ? 4 : 8;
+	cppchar_t value = 0;
+	size_t bufleft = len - (bufp - buf);
+	int rval;
+
+	idp += 2;
+	while (length && idp < len && ISXDIGIT (id[idp]))
+	  {
+	    value = (value << 4) + hex_value (id[idp]);
+	    idp++;
+	    length--;
+	  }
+	idp--;
+
+	/* Special case for EBCDIC: if the identifier contains
+	   a '$' specified using a UCN, translate it to EBCDIC.  */
+	if (value == 0x24)
+	  {
+	    *bufp++ = '$';
+	    continue;
+	  }
+
+	rval = one_cppchar_to_utf8 (value, &bufp, &bufleft);
+	if (rval)
+	  {
+	    errno = rval;
+	    cpp_errno (pfile, CPP_DL_ERROR,
+		       "converting UCN to source character set");
+	    break;
+	  }
+      }
 
+  return CPP_HASHNODE (ht_lookup (pfile->hash_table, 
+				  buf, bufp - buf, HT_ALLOC));
+}
+
 /* Convert an input buffer (containing the complete contents of one
    source file) from INPUT_CHARSET to the source character set.  INPUT
    points to the input buffer, SIZE is its allocated size, and LEN is
diff --git a/libcpp/directives.c b/libcpp/directives.c
index 84065052630..957e879caec 100644
--- a/libcpp/directives.c
+++ b/libcpp/directives.c
@@ -608,7 +608,8 @@ glue_header_name (cpp_reader *pfile)
       if (token->flags & PREV_WHITE)
 	buffer[total_len++] = ' ';
 
-      total_len = (cpp_spell_token (pfile, token, (uchar *) &buffer[total_len])
+      total_len = (cpp_spell_token (pfile, token, (uchar *) &buffer[total_len],
+				    true)
 		   - (uchar *) buffer);
     }
 
diff --git a/libcpp/include/cpplib.h b/libcpp/include/cpplib.h
index 70f8d895afd..ccf8bff47e4 100644
--- a/libcpp/include/cpplib.h
+++ b/libcpp/include/cpplib.h
@@ -637,7 +637,7 @@ extern unsigned int cpp_errors (cpp_reader *);
 extern unsigned int cpp_token_len (const cpp_token *);
 extern unsigned char *cpp_token_as_text (cpp_reader *, const cpp_token *);
 extern unsigned char *cpp_spell_token (cpp_reader *, const cpp_token *,
-				       unsigned char *);
+				       unsigned char *, bool);
 extern void cpp_register_pragma (cpp_reader *, const char *, const char *,
 				 void (*) (cpp_reader *), bool);
 extern void cpp_handle_deferred_pragma (cpp_reader *, const cpp_string *);
diff --git a/libcpp/internal.h b/libcpp/internal.h
index 0ae13d58cb6..af823d766b3 100644
--- a/libcpp/internal.h
+++ b/libcpp/internal.h
@@ -571,6 +571,9 @@ extern unsigned char *_cpp_convert_input (cpp_reader *, const char *,
 					  unsigned char *, size_t, size_t,
 					  off_t *);
 extern const char *_cpp_default_encoding (void);
+extern cpp_hashnode * _cpp_interpret_identifier (cpp_reader *pfile,
+						 const unsigned char *id,
+						 size_t len);
 
 /* Utility routines and macros.  */
 #define DSC(str) (const unsigned char *)str, sizeof str - 1
diff --git a/libcpp/lex.c b/libcpp/lex.c
index 62a28f81b87..8398c7ca061 100644
--- a/libcpp/lex.c
+++ b/libcpp/lex.c
@@ -53,7 +53,7 @@ static const struct token_spelling token_spellings[N_TTYPES] = { TTYPE_TABLE };
 static void add_line_note (cpp_buffer *, const uchar *, unsigned int);
 static int skip_line_comment (cpp_reader *);
 static void skip_whitespace (cpp_reader *, cppchar_t);
-static cpp_hashnode *lex_identifier (cpp_reader *, const uchar *);
+static cpp_hashnode *lex_identifier (cpp_reader *, const uchar *, bool);
 static void lex_number (cpp_reader *, cpp_string *);
 static bool forms_identifier_p (cpp_reader *, int);
 static void lex_string (cpp_reader *, cpp_token *, const uchar *);
@@ -453,7 +453,7 @@ forms_identifier_p (cpp_reader *pfile, int first)
     }
 
   /* Is this a syntactically valid UCN?  */
-  if (0 && *buffer->cur == '\\'
+  if (*buffer->cur == '\\'
       && (buffer->cur[1] == 'u' || buffer->cur[1] == 'U'))
     {
       buffer->cur += 2;
@@ -467,39 +467,39 @@ forms_identifier_p (cpp_reader *pfile, int first)
 
 /* Lex an identifier starting at BUFFER->CUR - 1.  */
 static cpp_hashnode *
-lex_identifier (cpp_reader *pfile, const uchar *base)
+lex_identifier (cpp_reader *pfile, const uchar *base, bool starts_ucn)
 {
   cpp_hashnode *result;
-  const uchar *cur, *limit;
+  const uchar *cur;
   unsigned int len;
   unsigned int hash = HT_HASHSTEP (0, *base);
 
   cur = pfile->buffer->cur;
-  for (;;)
+  if (! starts_ucn)
+    while (ISIDNUM (*cur))
+      {
+	hash = HT_HASHSTEP (hash, *cur);
+	cur++;
+      }
+  pfile->buffer->cur = cur;
+  if (starts_ucn || forms_identifier_p (pfile, false))
     {
-      /* N.B. ISIDNUM does not include $.  */
-      while (ISIDNUM (*cur))
-	{
-	  hash = HT_HASHSTEP (hash, *cur);
-	  cur++;
-	}
-
-      pfile->buffer->cur = cur;
-      if (!forms_identifier_p (pfile, false))
-	break;
-
-      limit = pfile->buffer->cur;
-      while (cur < limit)
-	{
-	  hash = HT_HASHSTEP (hash, *cur);
-	  cur++;
-	}
+      /* Slower version for identifiers containing UCNs (or $).  */
+      do {
+	while (ISIDNUM (*pfile->buffer->cur))
+	  pfile->buffer->cur++;
+      } while (forms_identifier_p (pfile, false));
+      result = _cpp_interpret_identifier (pfile, base,
+					  pfile->buffer->cur - base);
     }
-  len = cur - base;
-  hash = HT_HASHFINISH (hash, len);
+  else
+    {
+      len = cur - base;
+      hash = HT_HASHFINISH (hash, len);
 
-  result = (cpp_hashnode *)
-    ht_lookup_with_hash (pfile->hash_table, base, len, hash, HT_ALLOC);
+      result = (cpp_hashnode *)
+	ht_lookup_with_hash (pfile->hash_table, base, len, hash, HT_ALLOC);
+    }
 
   /* Rarely, identifiers require diagnostics when lexed.  */
   if (__builtin_expect ((result->flags & NODE_DIAGNOSTIC)
@@ -922,7 +922,7 @@ _cpp_lex_direct (cpp_reader *pfile)
     case 'S': case 'T': case 'U': case 'V': case 'W': case 'X':
     case 'Y': case 'Z':
       result->type = CPP_NAME;
-      result->val.node = lex_identifier (pfile, buffer->cur - 1);
+      result->val.node = lex_identifier (pfile, buffer->cur - 1, false);
 
       /* Convert named operators to their proper types.  */
       if (result->val.node->flags & NODE_OPERATOR)
@@ -1155,7 +1155,7 @@ _cpp_lex_direct (cpp_reader *pfile)
 	if (forms_identifier_p (pfile, true))
 	  {
 	    result->type = CPP_NAME;
-	    result->val.node = lex_identifier (pfile, base);
+	    result->val.node = lex_identifier (pfile, base, true);
 	    break;
 	  }
 	buffer->cur++;
@@ -1180,19 +1180,56 @@ cpp_token_len (const cpp_token *token)
     {
     default:		len = 4;				break;
     case SPELL_LITERAL:	len = token->val.str.len;		break;
-    case SPELL_IDENT:	len = NODE_LEN (token->val.node);	break;
+    case SPELL_IDENT:	len = NODE_LEN (token->val.node) * 10;	break;
     }
 
   return len;
 }
 
+/* Parse UTF-8 out of NAMEP and place a \U escape in BUFFER.
+   Return the number of bytes read out of NAME.  (There are always
+   10 bytes written to BUFFER.)  */
+
+static size_t
+utf8_to_ucn (unsigned char *buffer, const unsigned char *name)
+{
+  int j;
+  int ucn_len = 0;
+  int ucn_len_c;
+  unsigned t;
+  unsigned long utf32;
+  
+  /* Compute the length of the UTF-8 sequence.  */
+  for (t = *name; t & 0x80; t <<= 1)
+    ucn_len++;
+  
+  utf32 = *name & (0x7F >> ucn_len);
+  for (ucn_len_c = 1; ucn_len_c < ucn_len; ucn_len_c++)
+    {
+      utf32 = (utf32 << 6) | (*++name & 0x3F);
+      
+      /* Ill-formed UTF-8.  */
+      if ((*name & ~0x3F) != 0x80)
+	abort ();
+    }
+  
+  *buffer++ = '\\';
+  *buffer++ = 'U';
+  for (j = 7; j >= 0; j--)
+    *buffer++ = "0123456789abcdef"[(utf32 >> (4 * j)) & 0xF];
+  return ucn_len;
+}
+
+
 /* Write the spelling of a token TOKEN to BUFFER.  The buffer must
    already contain the enough space to hold the token's spelling.
    Returns a pointer to the character after the last character written.
+   FORSTRING is true if this is to be the spelling after translation
+   phase 1 (this is different for UCNs).
    FIXME: Would be nice if we didn't need the PFILE argument.  */
 unsigned char *
 cpp_spell_token (cpp_reader *pfile, const cpp_token *token,
-		 unsigned char *buffer)
+		 unsigned char *buffer, bool forstring)
 {
   switch (TOKEN_SPELL (token))
     {
@@ -1216,8 +1253,26 @@ cpp_spell_token (cpp_reader *pfile, const cpp_token *token,
 
     spell_ident:
     case SPELL_IDENT:
-      memcpy (buffer, NODE_NAME (token->val.node), NODE_LEN (token->val.node));
-      buffer += NODE_LEN (token->val.node);
+      if (forstring)
+	{
+	  memcpy (buffer, NODE_NAME (token->val.node),
+		  NODE_LEN (token->val.node));
+	  buffer += NODE_LEN (token->val.node);
+	}
+      else
+	{
+	  size_t i;
+	  const unsigned char * name = NODE_NAME (token->val.node);
+	  
+	  for (i = 0; i < NODE_LEN (token->val.node); i++)
+	    if (name[i] & ~0x7F)
+	      {
+		i += utf8_to_ucn (buffer, name + i) - 1;
+		buffer += 10;
+	      }
+	    else
+	      *buffer++ = NODE_NAME (token->val.node)[i];
+	}
       break;
 
     case SPELL_LITERAL:
@@ -1242,7 +1297,7 @@ cpp_token_as_text (cpp_reader *pfile, const cpp_token *token)
   unsigned int len = cpp_token_len (token) + 1;
   unsigned char *start = _cpp_unaligned_alloc (pfile, len), *end;
 
-  end = cpp_spell_token (pfile, token, start);
+  end = cpp_spell_token (pfile, token, start, false);
   end[0] = '\0';
 
   return start;
@@ -1286,8 +1341,21 @@ cpp_output_token (const cpp_token *token, FILE *fp)
 
     spell_ident:
     case SPELL_IDENT:
-      fwrite (NODE_NAME (token->val.node), 1, NODE_LEN (token->val.node), fp);
-    break;
+      {
+	size_t i;
+	const unsigned char * name = NODE_NAME (token->val.node);
+	
+	for (i = 0; i < NODE_LEN (token->val.node); i++)
+	  if (name[i] & ~0x7F)
+	    {
+	      unsigned char buffer[10];
+	      i += utf8_to_ucn (buffer, name + i) - 1;
+	      fwrite (buffer, 1, 10, fp);
+	    }
+	  else
+	    fputc (NODE_NAME (token->val.node)[i], fp);
+      }
+      break;
 
     case SPELL_LITERAL:
       fwrite (token->val.str.text, 1, token->val.str.len, fp);
diff --git a/libcpp/macro.c b/libcpp/macro.c
index 7d65886a390..441b3b32ed3 100644
--- a/libcpp/macro.c
+++ b/libcpp/macro.c
@@ -380,12 +380,12 @@ stringify_arg (cpp_reader *pfile, macro_arg *arg)
 	{
 	  _cpp_buff *buff = _cpp_get_buff (pfile, len);
 	  unsigned char *buf = BUFF_FRONT (buff);
-	  len = cpp_spell_token (pfile, token, buf) - buf;
+	  len = cpp_spell_token (pfile, token, buf, true) - buf;
 	  dest = cpp_quote_string (dest, buf, len);
 	  _cpp_release_buff (pfile, buff);
 	}
       else
-	dest = cpp_spell_token (pfile, token, dest);
+	dest = cpp_spell_token (pfile, token, dest, true);
 
       if (token->type == CPP_OTHER && token->val.str.text[0] == '\\')
 	backslash_count++;
@@ -422,7 +422,7 @@ paste_tokens (cpp_reader *pfile, const cpp_token **plhs, const cpp_token *rhs)
   lhs = *plhs;
   len = cpp_token_len (lhs) + cpp_token_len (rhs) + 1;
   buf = alloca (len);
-  end = cpp_spell_token (pfile, lhs, buf);
+  end = cpp_spell_token (pfile, lhs, buf, false);
 
   /* Avoid comment headers, since they are still processed in stage 3.
      It is simpler to insert a space here, rather than modifying the
@@ -430,7 +430,7 @@ paste_tokens (cpp_reader *pfile, const cpp_token **plhs, const cpp_token *rhs)
      false doesn't work, since we want to clear the PASTE_LEFT flag.  */
   if (lhs->type == CPP_DIV && rhs->type != CPP_EQ)
     *end++ = ' ';
-  end = cpp_spell_token (pfile, rhs, end);
+  end = cpp_spell_token (pfile, rhs, end, false);
   *end = '\n';
 
   cpp_push_buffer (pfile, buf, end - buf, /* from_stage3 */ true);
@@ -1751,7 +1751,7 @@ cpp_macro_definition (cpp_reader *pfile, const cpp_hashnode *node)
 	      buffer += NODE_LEN (macro->params[token->val.arg_no - 1]);
 	    }
 	  else
-	    buffer = cpp_spell_token (pfile, token, buffer);
+	    buffer = cpp_spell_token (pfile, token, buffer, false);
 
 	  if (token->flags & PASTE_LEFT)
 	    {