Unicode support for erlang:display_string/1

3 files changed, 56 insertions, 6 deletions

diff --git a/erts/emulator/beam/bif.c b/erts/emulator/beam/bif.c
index 4c92b9ba2b..22805b5c02 100644
--- a/erts/emulator/beam/bif.c
+++ b/erts/emulator/beam/bif.c
@@ -3947,15 +3947,18 @@ BIF_RETTYPE display_string_1(BIF_ALIST_1)
 {
     Process* p = BIF_P;
     Eterm string = BIF_ARG_1;
-    Sint len = is_string(string);
-    char *str;
+    Sint len = erts_unicode_list_to_buf_len(string);
+    Sint written;
+    byte *str;
+    int res;
 
-    if (len <= 0) {
+    if (len < 0) {
 	BIF_ERROR(p, BADARG);
     }
-    str = (char *) erts_alloc(ERTS_ALC_T_TMP, sizeof(char)*(len + 1));
-    if (intlist_to_buf(string, str, len) != len)
-	erts_exit(ERTS_ERROR_EXIT, "%s:%d: Internal error\n", __FILE__, __LINE__);
+    str = (byte *) erts_alloc(ERTS_ALC_T_TMP, sizeof(char)*(len + 1));
+    res = erts_unicode_list_to_buf(string, str, len, &written);
+    if (res != 0 || written != len)
+	erts_exit(ERTS_ERROR_EXIT, "%s:%d: Internal error (%d)\n", __FILE__, __LINE__, res);
     str[len] = '\0';
     erts_fprintf(stderr, "%s", str);
     erts_free(ERTS_ALC_T_TMP, (void *) str);
diff --git a/erts/emulator/beam/global.h b/erts/emulator/beam/global.h
index fcb88712e9..64e543586a 100644
--- a/erts/emulator/beam/global.h
+++ b/erts/emulator/beam/global.h
@@ -1288,6 +1288,7 @@ int erts_utf8_to_latin1(byte* dest, const byte* source, int slen);
 void bin_write(fmtfn_t, void*, byte*, size_t);
 Sint intlist_to_buf(Eterm, char*, Sint); /* most callers pass plain char*'s */
 int erts_unicode_list_to_buf(Eterm list, byte *buf, Sint len, Sint* written);
+Sint erts_unicode_list_to_buf_len(Eterm list);
 
 struct Sint_buf {
 #if defined(ARCH_64)
diff --git a/erts/emulator/beam/utils.c b/erts/emulator/beam/utils.c
index ba7596cde4..16523659d9 100644
--- a/erts/emulator/beam/utils.c
+++ b/erts/emulator/beam/utils.c
@@ -3704,6 +3704,52 @@ erts_unicode_list_to_buf(Eterm list, byte *buf, Sint len, Sint* written)
     return res;
 }
 
+Sint
+erts_unicode_list_to_buf_len(Eterm list)
+{
+    Eterm* listptr;
+    Sint sz = 0;
+
+    if (is_nil(list)) {
+	return 0;
+    }
+    if (is_not_list(list)) {
+	return -1;
+    }
+    listptr = list_val(list);
+
+    while (1) {
+	Sint val;
+
+	if (is_not_small(CAR(listptr))) {
+	    return -1;
+	}
+	val = signed_val(CAR(listptr));
+	if (0 <= val && val < 0x80) {
+	    sz++;
+	} else if (val < 0x800) {
+	    sz += 2;
+	} else if (val < 0x10000UL) {
+	    if (0xD800 <= val && val <= 0xDFFF) {
+		return -1;
+	    }
+	    sz += 3;
+	} else if (val < 0x110000) {
+	    sz += 4;
+	} else {
+	    return -1;
+	}
+	list = CDR(listptr);
+	if (is_nil(list)) {
+	    return sz;
+	}
+	if (is_not_list(list)) {
+	    return -1;
+	}
+	listptr = list_val(list);
+    }
+}
+
 /*
 ** Convert an integer to a byte list
 ** return pointer to converted stuff (need not to be at start of buf!)