erl_interface: utf8 atoms continued

1 files changed, 132 insertions, 53 deletions

diff --git a/lib/erl_interface/src/decode/decode_atom.c b/lib/erl_interface/src/decode/decode_atom.c
index 84edf1766a..2ada418243 100644
--- a/lib/erl_interface/src/decode/decode_atom.c
+++ b/lib/erl_interface/src/decode/decode_atom.c
@@ -21,76 +21,155 @@
 #include "eiext.h"
 #include "putget.h"
 
-static int utf8_to_latin1(char* dest, const char* source, unsigned len);
 
 int ei_decode_atom(const char *buf, int *index, char *p)
 {
-  const char *s = buf + *index;
-  const char *s0 = s;
-  int len;
-
-  switch (get8(s)) {
-  case ERL_ATOM_EXT:
-      len = get16be(s);
-      if (len > MAXATOMLEN) return -1;
-      if (p) {
-	  memmove(p,s,len); 
-	  p[len] = (char)0;
-      }
-      break;
-
-  case ERL_SMALL_ATOM_EXT:
-      len = get8(s);
-      if (p) {
-	  memmove(p,s,len); 
-	  p[len] = (char)0;
-      }
-      break;
-
-  case ERL_UNICODE_ATOM_EXT:
-      len = get16be(s);
-
-      if (len > 2*MAXATOMLEN) return -1;
-
-      if (p && utf8_to_latin1(p, s, len) < 0) return -1;
-      break;
-
-  default:
-      return -1;
-  }
-
-  s += len;
-  *index += s-s0;
-  return 0;
+    return ei_decode_atom_as(buf, index, p, MAXATOMLEN, ERLANG_LATIN1, NULL, NULL);
 }
 
-int ei_internal_get_atom(const char** bufp, char* p)
+int ei_decode_atom_as(const char *buf, int *index, char* p, int destlen,
+		      enum erlang_char_encoding want_enc,
+		      enum erlang_char_encoding* was_encp,
+		      enum erlang_char_encoding* res_encp)
 {
-    int ix = 0;
-    if (ei_decode_atom(*bufp, &ix, p) < 0) return -1;
-    *bufp += ix;
+    const char *s = buf + *index;
+    const char *s0 = s;
+    int len;
+    enum erlang_char_encoding got_enc;
+    
+    switch (get8(s)) {
+    case ERL_ATOM_EXT:
+	len = get16be(s);
+	got_enc = ERLANG_LATIN1;
+	break;
+    case ERL_SMALL_ATOM_EXT:
+	len = get8(s);
+	got_enc = ERLANG_LATIN1;
+	break;
+    case ERL_ATOM_UTF8_EXT:
+	len = get16be(s);
+	got_enc = ERLANG_UTF8;
+	break;
+    case ERL_SMALL_ATOM_UTF8_EXT:
+	len = get8(s);
+	got_enc = ERLANG_UTF8;
+	break;
+    default:
+	return -1;
+    }
+    
+    if (want_enc == got_enc || want_enc == ERLANG_WHATEVER || want_enc == ERLANG_ASCII) {
+	int i, found_non_ascii = 0;
+	if (len >= destlen)
+	    return -1;
+	for (i=0; i<len; i++) {
+	    if (s[i] & 0x80) found_non_ascii = 1;
+	    if (p) p[i] = s[i];
+	}
+	if (p) p[len] = 0;
+	if (want_enc == ERLANG_ASCII && found_non_ascii) {
+	    return -1;
+	}
+	if (res_encp) {
+	    *res_encp = found_non_ascii ? got_enc : ERLANG_ASCII;
+	}
+    }
+    else {
+	int plen = (got_enc == ERLANG_LATIN1) ?
+	    utf8_to_latin1(p, s, len, destlen-1, res_encp) :
+	    latin1_to_utf8(p, s, len, destlen-1, res_encp);
+	if (plen < 0) return -1;
+	if (p) p[plen] = 0;
+    }
+    if (was_encp) {
+	*was_encp = got_enc;
+    }
+    
+    s += len;
+    *index += s-s0;
     return 0;
-}
+}	
+
 
-static int utf8_to_latin1(char* dest, const char* source, unsigned slen)
+int utf8_to_latin1(char* dst, const char* src, int slen, int destlen,
+		   enum erlang_char_encoding* res_encp)
 {
-    const char* dest_end = dest + MAXATOMLEN - 1;
+    const char* const dst_start = dst;
+    const char* const dst_end = dst + destlen;
+    int found_non_ascii = 0;
 
-    while (slen > 0 && dest < dest_end) {
-	if ((source[0] & 0x80) == 0) {
-	    *dest++ = *source++;
+    while (slen > 0) {
+	if (dst >= dst_end) return -1;
+	if ((src[0] & 0x80) == 0) {
+	    if (dst_start) {
+		*dst = *src;
+	    }
+	    ++dst;
+	    ++src;
 	    --slen;
 	}
 	else if (slen > 1 &&
-		 (source[0] & 0xFE) == 0xC2 &&
-		 (source[1] & 0xC0) == 0x80) {
-	    *dest++ = (char) ((source[0] << 6) | (source[1] & 0x3F));
-	    source += 2;
+		 (src[0] & 0xFE) == 0xC2 &&
+		 (src[1] & 0xC0) == 0x80) {
+	    if (dst_start) {
+		*dst = (char) ((src[0] << 6) | (src[1] & 0x3F));
+	    }
+	    ++dst;
+	    src += 2;
 	    slen -= 2;
+	    found_non_ascii = 1;
 	}
 	else return -1;
     }
-    *dest = 0;
+    if (res_encp) {
+	*res_encp = found_non_ascii ? ERLANG_LATIN1 : ERLANG_ASCII;
+    }
+    return dst - dst_start;
+}
+
+int latin1_to_utf8(char* dst, const char* src, int slen, int destlen,
+		   enum erlang_char_encoding* res_encp)
+{
+    const char* const src_end = src + slen;
+    const char* const dst_start = dst;
+    const char* const dst_end = dst + destlen;
+    int found_non_ascii = 0;
+
+    while (src < src_end) {
+	if (dst >= dst_end) return -1;
+	if ((src[0] & 0x80) == 0) {
+	    if (dst_start) {
+		*dst = *src;
+	    }
+	    ++dst;
+	}
+	else {
+	    if (dst_start) {
+		unsigned char ch = *src;
+		dst[0] = 0xC0 | (ch >> 6);
+		dst[1] = 0x80 | (ch & 0x3F);
+	    }
+	    dst += 2;
+	    found_non_ascii = 1;
+	}
+	++src;
+    }
+    if (res_encp) {
+	*res_encp = found_non_ascii ? ERLANG_UTF8 : ERLANG_ASCII;
+    }
+    return dst - dst_start;
+}
+
+
+
+int ei_internal_get_atom(const char** bufp, char* p,
+			 enum erlang_char_encoding* was_encp)
+{
+    int ix = 0;
+    if (ei_decode_atom_as(*bufp, &ix, p, MAXATOMLEN_UTF8, ERLANG_UTF8, was_encp, NULL) < 0)
+	return -1;
+    *bufp += ix;
     return 0;
 }
 
+