Merge branch 'rickard/t2b-latin1-atom'

OTP-14337

* rickard/t2b-latin1-atom:
  Update primary bootstrap
  Update preloaded modules
  Fix erl_interface tests
  Introduce minor vsn 2 in term_to_binary/2
  Revert "kernel: Try mend disk_log whitebox tests"
  Revert "erts: Do not generate atoms on old latin1 external format"

2 files changed, 39 insertions, 5 deletions

diff --git a/erts/emulator/beam/dist.h b/erts/emulator/beam/dist.h
index 93a651b24d..3e17645997 100644
--- a/erts/emulator/beam/dist.h
+++ b/erts/emulator/beam/dist.h
@@ -53,7 +53,6 @@
 			       | 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 1190d90b8e..95275847f8 100644
--- a/erts/emulator/beam/external.c
+++ b/erts/emulator/beam/external.c
@@ -1129,8 +1129,10 @@ BIF_RETTYPE term_to_binary_2(BIF_ALIST_2)
 		case 0:
 		    flags = TERM_TO_BINARY_DFLAGS & ~DFLAG_NEW_FLOATS;
 		    break;
-		case 1:
+		case 1: /* Current default... */
 		    flags = TERM_TO_BINARY_DFLAGS;
+                case 2:
+                    flags = TERM_TO_BINARY_DFLAGS | DFLAG_UTF8_ATOMS;
 		    break;
 		default:
 		    goto error;
@@ -2090,6 +2092,7 @@ 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));
 
@@ -2118,8 +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 (len > 255) {
+	if (utf8_atoms || a->latin1_chars < 0) {
+	    if (len > 255) {
 		*ep++ = ATOM_UTF8_EXT;
 		put_int16(len, ep);
 		ep += 2;
@@ -2131,6 +2134,32 @@ 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;
     }
@@ -4053,13 +4082,19 @@ 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;