From 65b04e233e09e3cc2e0fda3c28e155b95c5a4baf Mon Sep 17 00:00:00 2001
From: Sverker Eriksson <sverker@erlang.org>
Date: Tue, 7 Mar 2017 17:04:40 +0100
Subject: erts: Do not generate atoms on old latin1 external format

term_to_binary will always generate utf8 atoms
ATOM_UTF8_EXT and SMALL_ATOM_UTF8_EXT.

Old latin1 atoms, ATOM_EXT and SMALL_ATOM_EXT, are still decoded.
---
 erts/emulator/beam/dist.h     |  1 +
 erts/emulator/beam/external.c | 39 +++------------------------------------
 2 files changed, 4 insertions(+), 36 deletions(-)

diff --git a/erts/emulator/beam/dist.h b/erts/emulator/beam/dist.h
index 8f6be1061a..6ed36a478e 100644
--- a/erts/emulator/beam/dist.h
+++ b/erts/emulator/beam/dist.h
@@ -53,6 +53,7 @@
 			       | DFLAG_EXPORT_PTR_TAG		\
 			       | DFLAG_BIT_BINARIES             \
 			       | DFLAG_MAP_TAG                  \
+                               | DFLAG_UTF8_ATOMS               \
                                | DFLAG_BIG_CREATION)
 
 /* opcodes used in distribution messages */
diff --git a/erts/emulator/beam/external.c b/erts/emulator/beam/external.c
index 186b250c7e..285ae4ac78 100644
--- a/erts/emulator/beam/external.c
+++ b/erts/emulator/beam/external.c
@@ -2093,7 +2093,6 @@ enc_atom(ErtsAtomCacheMap *acmp, Eterm atom, byte *ep, Uint32 dflags)
 {
     int iix;
     int len;
-    int utf8_atoms = (int) (dflags & DFLAG_UTF8_ATOMS);
 
     ASSERT(is_atom(atom));
 
@@ -2122,8 +2121,8 @@ enc_atom(ErtsAtomCacheMap *acmp, Eterm atom, byte *ep, Uint32 dflags)
     if (iix < 0) {
 	Atom *a = atom_tab(atom_val(atom));
 	len = a->len;
-	if (utf8_atoms || a->latin1_chars < 0) {
-	    if (len > 255) {
+        {
+            if (len > 255) {
 		*ep++ = ATOM_UTF8_EXT;
 		put_int16(len, ep);
 		ep += 2;
@@ -2135,32 +2134,6 @@ enc_atom(ErtsAtomCacheMap *acmp, Eterm atom, byte *ep, Uint32 dflags)
 	    }
 	    sys_memcpy((char *) ep, (char *) a->name, len);
 	}
-	else {
-	    if (a->latin1_chars <= 255 && (dflags & DFLAG_SMALL_ATOM_TAGS)) {
-		*ep++ = SMALL_ATOM_EXT;
-		if (len == a->latin1_chars) {
-		    sys_memcpy(ep+1, a->name, len);
-		}
-		else {
-		    len = erts_utf8_to_latin1(ep+1, a->name, len);
-		    ASSERT(len == a->latin1_chars);
-		}
-		put_int8(len, ep);
-		ep++;
-	    }
-	    else {
-		*ep++ = ATOM_EXT;
-		if (len == a->latin1_chars) {
-		    sys_memcpy(ep+2, a->name, len);
-		}
-		else {
-		    len = erts_utf8_to_latin1(ep+2, a->name, len);
-		    ASSERT(len == a->latin1_chars);
-		}
-		put_int16(len, ep);
-		ep += 2;
-	    }	    
-	}
 	ep += len;
 	return ep;
     }
@@ -4085,19 +4058,13 @@ encode_size_struct_int(TTBSizeContext* ctx, ErtsAtomCacheMap *acmp, Eterm obj,
 	    else {
 		Atom *a = atom_tab(atom_val(obj));
 		int alen;
-		if ((dflags & DFLAG_UTF8_ATOMS) || a->latin1_chars < 0) {
+                {
 		    alen = a->len;
 		    result += 1 + 1 + alen;
 		    if (alen > 255) {
 			result++; /* ATOM_UTF8_EXT (not small) */
 		    }
 		}
-		else {
-		    alen = a->latin1_chars;
-		    result += 1 + 1 + alen;
-		    if (alen > 255 || !(dflags & DFLAG_SMALL_ATOM_TAGS))
-			result++; /* ATOM_EXT (not small) */
-		}
 		insert_acache_map(acmp, obj, dflags);
 	    }
 	    break;
-- 
cgit v1.2.3