From 32d13d59910a384ad09682cafc83c7300c96c694 Mon Sep 17 00:00:00 2001
From: Sverker Eriksson <sverker@erlang.org>
Date: Wed, 8 Mar 2017 14:46:24 +0100
Subject: erl_interface: Do not generate atoms on old latin1 ext format

Solved by letting ei_encode_atom_as ignore 'to_enc' argument
and always encode in UTF8 format.
---
 lib/erl_interface/src/legacy/erl_eterm.c   | 16 +++++++++++-----
 lib/erl_interface/src/legacy/erl_marshal.c | 13 ++++---------
 2 files changed, 15 insertions(+), 14 deletions(-)

(limited to 'lib/erl_interface/src/legacy')

diff --git a/lib/erl_interface/src/legacy/erl_eterm.c b/lib/erl_interface/src/legacy/erl_eterm.c
index e4b3b49c7d..5153d0f2e7 100644
--- a/lib/erl_interface/src/legacy/erl_eterm.c
+++ b/lib/erl_interface/src/legacy/erl_eterm.c
@@ -188,14 +188,20 @@ char* erl_atom_ptr_latin1(Erl_Atom_data* a)
 char* erl_atom_ptr_utf8(Erl_Atom_data* a)
 {
     if (a->utf8 == NULL) {
-	int dlen = a->lenL * 2; /* over estimation */
-	a->utf8 = malloc(dlen + 1); 
-	a->lenU = latin1_to_utf8(a->utf8, a->latin1, a->lenL, dlen, NULL);
-	a->utf8[a->lenU] = '\0';
+        erlang_char_encoding enc;
+        a->lenU = latin1_to_utf8(NULL, a->latin1, a->lenL, a->lenL*2, &enc);
+        if (enc == ERLANG_ASCII) {
+            a->utf8 = a->latin1;
+        }
+        else {
+            a->utf8 = malloc(a->lenU + 1);
+            latin1_to_utf8(a->utf8, a->latin1, a->lenL, a->lenU, NULL);
+            a->utf8[a->lenU] = '\0';
+        }
     }
     return a->utf8;
-
 }
+
 int erl_atom_size_latin1(Erl_Atom_data* a)
 {
     if (a->latin1 == NULL) {
diff --git a/lib/erl_interface/src/legacy/erl_marshal.c b/lib/erl_interface/src/legacy/erl_marshal.c
index 527ae0ef8f..b7a8455313 100644
--- a/lib/erl_interface/src/legacy/erl_marshal.c
+++ b/lib/erl_interface/src/legacy/erl_marshal.c
@@ -175,10 +175,9 @@ static void encode_atom(Erl_Atom_data* a, unsigned char **ext)
     int ix = 0;
     if (a->latin1) {
 	ei_encode_atom_len_as((char*)*ext, &ix, a->latin1, a->lenL,
-			      ERLANG_LATIN1, ERLANG_LATIN1);
+			      ERLANG_LATIN1, ERLANG_UTF8);
     }
-    else if (ei_encode_atom_len_as((char*)*ext, &ix, a->utf8, a->lenU,
-				   ERLANG_UTF8, ERLANG_LATIN1) < 0) {
+    else {
 	ei_encode_atom_len_as((char*)*ext, &ix, a->utf8, a->lenU,
 			      ERLANG_UTF8, ERLANG_UTF8);
     }
@@ -542,12 +541,8 @@ int erl_term_len(ETERM *ep)
 
 static int atom_len_helper(Erl_Atom_data* a)
 {
-    if (erl_atom_ptr_latin1(a)) {
-	return 1 + 2 + a->lenL; /* ERL_ATOM_EXT */
-    }
-    else {
-	return 1 + 1 + (a->lenU > 255) + a->lenU;
-    }
+    (void) erl_atom_ptr_utf8(a);
+    return 1 + 1 + (a->lenU > 255) + a->lenU;
 }
 
 static int erl_term_len_helper(ETERM *ep, int dist)
-- 
cgit v1.2.3