Merge branch 'egil/erts/remove-broken-hash/OTP-13827'

* egil/erts/remove-broken-hash/OTP-13827: Update test cases for erlang:hash/2 removal mnesia: Remove mnesia_frag_old_hash hash module stdlib: Produce correct warning for erlang:hash/2 erts: Remove erlang:hash/2 from documentation erts: Remove broken hash from Erlang
kmizu · Jan 26, 2017 · 9153f7a · 9153f7a
2 parents 7ac0749 + 52097fa
commit 9153f7a
Show file tree

Hide file tree

Showing 25 changed files with 143 additions and 745 deletions.
diff --git a/erts/doc/src/erlang.xml b/erts/doc/src/erlang.xml
@@ -1954,26 +1954,6 @@ os_prompt%</pre>
       </desc>
     </func>
 
-    <func>
-      <name name="hash" arity="2"/>
-      <fsummary>Hash function (deprecated).</fsummary>
-      <desc>
-        <p>Returns a hash value for <c><anno>Term</anno></c> within the range
-          <c>1..<anno>Range</anno></c>. The maximum range is 1..2^27-1.</p>
-        <warning>
-          <p><em>This BIF is deprecated, as the hash value can differ on
-            different architectures.</em> The hash values for integer
-            terms &gt; 2^27 and large binaries are
-            poor. The BIF is retained for backward compatibility
-            reasons (it can have been used to hash records into a file),
-            but all new code is to use one of the BIFs
-            <seealso marker="#phash/2"><c>erlang:phash/2</c></seealso> or
-            <seealso marker="#phash2/1"><c>erlang:phash2/1,2</c></seealso>
-            instead.</p>
-        </warning>
-      </desc>
-    </func>
-
     <func>
       <name name="hd" arity="1"/>
       <fsummary>Head of a list.</fsummary>
@@ -3818,9 +3798,6 @@ RealSystem = system + MissedSystem</code>
           <c><anno>Term</anno></c> within the range
           <c>1..<anno>Range</anno></c>. The maximum value for
           <c><anno>Range</anno></c> is 2^32.</p>
-        <p>This BIF can be used instead of the old deprecated BIF
-          <c>erlang:hash/2</c>, as it calculates better hashes for
-          all data types, but consider using <c>phash2/1,2</c> instead.</p>
       </desc>
     </func>
 

diff --git a/erts/emulator/beam/bif.c b/erts/emulator/beam/bif.c
@@ -4723,25 +4723,6 @@ BIF_RETTYPE system_flag_2(BIF_ALIST_2)
 
 /**********************************************************************/
 
-BIF_RETTYPE hash_2(BIF_ALIST_2)
-{
-    Uint32 hash;
-    Sint range;
-
-    if (is_not_small(BIF_ARG_2)) {
-	BIF_ERROR(BIF_P, BADARG);
-    }
-    if ((range = signed_val(BIF_ARG_2)) <= 0) {  /* [1..MAX_SMALL] */
-	BIF_ERROR(BIF_P, BADARG);
-    }
-#if defined(ARCH_64)
-    if (range > ((1L << 27) - 1))
-	BIF_ERROR(BIF_P, BADARG);
-#endif
-    hash = make_broken_hash(BIF_ARG_1);
-    BIF_RET(make_small(1 + (hash % range)));   /* [1..range] */
-}
-
 BIF_RETTYPE phash_2(BIF_ALIST_2)
 {
     Uint32 hash;

diff --git a/erts/emulator/beam/bif.tab b/erts/emulator/beam/bif.tab
@@ -670,9 +670,3 @@ gcbif erlang:ceil/1
 bif math:floor/1
 bif math:ceil/1
 bif math:fmod/2
-
-#
-# Obsolete
-#
-
-bif erlang:hash/2
diff --git a/erts/emulator/beam/erl_utils.h b/erts/emulator/beam/erl_utils.h
@@ -117,7 +117,6 @@ int erts_fit_in_bits_int32(Sint32);
 int erts_fit_in_bits_uint(Uint);
 Sint erts_list_length(Eterm);
 int erts_is_builtin(Eterm, Eterm, int);
-Uint32 make_broken_hash(Eterm);
 Uint32 block_hash(byte *, unsigned, Uint32);
 Uint32 make_hash2(Eterm);
 Uint32 make_hash(Eterm);

diff --git a/erts/emulator/beam/utils.c b/erts/emulator/beam/utils.c
@@ -700,12 +700,7 @@ erts_bld_atom_2uint_3tup_list(Uint **hpp, Uint *szp, Sint length,
 /* make a hash index from an erlang term */
 
 /*
-** There are three hash functions.
-** make_broken_hash: the one used for backward compatibility
-** is called from the bif erlang:hash/2. Should never be used
-** as it a) hashes only a part of binaries, b) hashes bignums really poorly,
-** c) hashes bignums differently on different endian processors and d) hashes 
-** small integers with different weights on different bytes.
+** There are two hash functions.
 **
 ** make_hash: A hash function that will give the same values for the same
 ** terms regardless of the internal representation. Small integers are 
@@ -1045,6 +1040,10 @@ Uint32 make_hash(Eterm term_arg)
     DESTROY_WSTACK(stack);
     return hash;
 
+#undef MAKE_HASH_TUPLE_OP
+#undef MAKE_HASH_TERM_ARRAY_OP
+#undef MAKE_HASH_CDR_PRE_OP
+#undef MAKE_HASH_CDR_POST_OP
 #undef UINT32_HASH_STEP
 #undef UINT32_HASH_RET
 }
@@ -1954,259 +1953,6 @@ make_internal_hash(Eterm term)
 #undef HCONST
 #undef MIX
 
-
-Uint32 make_broken_hash(Eterm term)
-{
-    Uint32 hash = 0;
-    DECLARE_WSTACK(stack);
-    unsigned op;
-tail_recur:
-    op = tag_val_def(term);
-    for (;;) {
-    switch (op) {
-    case NIL_DEF:
-	hash = hash*FUNNY_NUMBER3 + 1;
-	break;
-    case ATOM_DEF:
-	hash = hash*FUNNY_NUMBER1 +
-	    (atom_tab(atom_val(term))->slot.bucket.hvalue);
-	break;
-    case SMALL_DEF:
-#if defined(ARCH_64)
-    {
-	Sint y1 = signed_val(term);
-	Uint y2 = y1 < 0 ? -(Uint)y1 : y1;
-	Uint32 y3 = (Uint32) (y2 >> 32);
-	int arity = 1;
-
-#if defined(WORDS_BIGENDIAN)
-	if (!IS_SSMALL28(y1))
-	{   /* like a bignum */
-	    Uint32 y4 = (Uint32) y2;
-	    hash = hash*FUNNY_NUMBER2 + ((y4 << 16) | (y4 >> 16));
-	    if (y3) {
-		hash = hash*FUNNY_NUMBER2 + ((y3 << 16) | (y3 >> 16));
-		arity++;
-	    }
-	    hash = hash * (y1 < 0 ? FUNNY_NUMBER3 : FUNNY_NUMBER2) + arity;
-	} else {
-	    hash = hash*FUNNY_NUMBER2 + (((Uint) y1) & 0xfffffff);
-	}
-#else
-	if  (!IS_SSMALL28(y1))
-	{   /* like a bignum */
-	    hash = hash*FUNNY_NUMBER2 + ((Uint32) y2);
-	    if (y3)
-	    {
-		hash = hash*FUNNY_NUMBER2 + y3;
-		arity++;
-	    }
-	    hash = hash * (y1 < 0 ? FUNNY_NUMBER3 : FUNNY_NUMBER2) + arity;
-	} else {
-	    hash = hash*FUNNY_NUMBER2 + (((Uint) y1) & 0xfffffff);
-	}
-#endif
-    }
-#else
-	hash = hash*FUNNY_NUMBER2 + unsigned_val(term);
-#endif
-	break;
-
-    case BINARY_DEF:
-	{
-	    size_t sz = binary_size(term);
-	    size_t i = (sz < 15) ? sz : 15;
-
-	    hash = hash_binary_bytes(term, i, hash);
-	    hash = hash*FUNNY_NUMBER4 + sz;
-	    break;
-	}
-
-    case EXPORT_DEF:
-	{
-	    Export* ep = *((Export **) (export_val(term) + 1));
-
-	    hash = hash * FUNNY_NUMBER11 + ep->info.mfa.arity;
-	    hash = hash*FUNNY_NUMBER1 +
-		(atom_tab(atom_val(ep->info.mfa.module))->slot.bucket.hvalue);
-	    hash = hash*FUNNY_NUMBER1 +
-		(atom_tab(atom_val(ep->info.mfa.function))->slot.bucket.hvalue);
-	    break;
-	}
-
-    case FUN_DEF:
-	{
-	    ErlFunThing* funp = (ErlFunThing *) fun_val(term);
-	    Uint num_free = funp->num_free;
-
-	    hash = hash * FUNNY_NUMBER10 + num_free;
-	    hash = hash*FUNNY_NUMBER1 +
-		(atom_tab(atom_val(funp->fe->module))->slot.bucket.hvalue);
-	    hash = hash*FUNNY_NUMBER2 + funp->fe->old_index;
-	    hash = hash*FUNNY_NUMBER2 + funp->fe->old_uniq;
-	    if (num_free > 0) {
-		if (num_free > 1) {
-		    WSTACK_PUSH3(stack, (UWord) &funp->env[1], (num_free-1), MAKE_HASH_TERM_ARRAY_OP);
-		}
-		term = funp->env[0];
-		goto tail_recur;
-	    }
-	    break;
-	}
-
-    case PID_DEF:
-	hash = hash*FUNNY_NUMBER5 + internal_pid_number(term);
-	break;
-    case EXTERNAL_PID_DEF:
-	hash = hash*FUNNY_NUMBER5 + external_pid_number(term);
-        break;
-    case PORT_DEF:
-	hash = hash*FUNNY_NUMBER9 + internal_port_number(term);
-	break;
-    case EXTERNAL_PORT_DEF:
-	hash = hash*FUNNY_NUMBER9 + external_port_number(term);
-	break;
-    case REF_DEF:
-	hash = hash*FUNNY_NUMBER9 + internal_ref_numbers(term)[0];
-	break;
-    case EXTERNAL_REF_DEF:
-	hash = hash*FUNNY_NUMBER9 + external_ref_numbers(term)[0];
-	break;
-    case FLOAT_DEF:
-	{
-            FloatDef ff;
-            GET_DOUBLE(term, ff);
-            if (ff.fd == 0.0f) {
-                /* ensure positive 0.0 */
-                ff.fd = erts_get_positive_zero_float();
-            }
-            hash = hash*FUNNY_NUMBER6 + (ff.fw[0] ^ ff.fw[1]);
-	}
-	break;
-    case MAKE_HASH_CDR_PRE_OP:
-	term = (Eterm) WSTACK_POP(stack);
-	if (is_not_list(term)) {
-	    WSTACK_PUSH(stack, (UWord) MAKE_HASH_CDR_POST_OP);
-	    goto tail_recur;
-	}
-	/*fall through*/
-    case LIST_DEF:
-	{
-	    Eterm* list = list_val(term);
-	    WSTACK_PUSH2(stack, (UWord) CDR(list),
-			 (UWord) MAKE_HASH_CDR_PRE_OP);
-	    term = CAR(list);
-	    goto tail_recur;
-	}
-
-    case MAKE_HASH_CDR_POST_OP:
-	hash *= FUNNY_NUMBER8;
-	break;
-
-     case BIG_DEF:
-	{
-	    Eterm* ptr  = big_val(term);
-	    int is_neg = BIG_SIGN(ptr);
-	    Uint arity = BIG_ARITY(ptr);
-	    Uint i = arity;
-	    ptr++;
-#if D_EXP == 16
-	    /* hash over 32 bit LE */
-
-	    while(i--) {
-		hash = hash*FUNNY_NUMBER2 + *ptr++;
-	    }
-#elif D_EXP == 32
-
-#if defined(WORDS_BIGENDIAN)
-	    while(i--) {
-		Uint d = *ptr++;
-		hash = hash*FUNNY_NUMBER2 + ((d << 16) | (d >> 16));
-	    }
-#else
-	    while(i--) {
-		hash = hash*FUNNY_NUMBER2 + *ptr++;
-	    }
-#endif
-
-#elif D_EXP == 64
-	    {
-	      Uint32 h = 0, l;
-#if defined(WORDS_BIGENDIAN)
-	      while(i--) {
-		  Uint d = *ptr++;
-		  l = d & 0xffffffff;
-		  h = d >> 32;
-		  hash = hash*FUNNY_NUMBER2 + ((l << 16) | (l >> 16));
-		  if (h || i)
-		      hash = hash*FUNNY_NUMBER2 + ((h << 16) | (h >> 16));
-	      }
-#else
-	      while(i--) {
-		  Uint d = *ptr++;
-		  l = d & 0xffffffff;
-		  h = d >> 32;
-		  hash = hash*FUNNY_NUMBER2 + l;
-		  if (h || i)
-		      hash = hash*FUNNY_NUMBER2 + h;
-	      }
-#endif
-	      /* adjust arity to match 32 bit mode */
-	      arity = (arity << 1) - (h == 0);
-	    }
-
-#else
-#error "unsupported D_EXP size"
-#endif
-	    hash = hash * (is_neg ? FUNNY_NUMBER3 : FUNNY_NUMBER2) + arity;
-	}
-	break;
-
-    case MAP_DEF:
-        hash = hash*FUNNY_NUMBER13 + FUNNY_NUMBER14 + make_hash2(term);
-        break;
-    case TUPLE_DEF:
-	{
-	    Eterm* ptr = tuple_val(term);
-	    Uint arity = arityval(*ptr);
-
-	    WSTACK_PUSH3(stack, (UWord) arity, (UWord) (ptr+1), (UWord) arity);
-	    op = MAKE_HASH_TUPLE_OP;
-	}/*fall through*/
-    case MAKE_HASH_TUPLE_OP:
-    case MAKE_HASH_TERM_ARRAY_OP:
-	{
-	    Uint i = (Uint) WSTACK_POP(stack);
-	    Eterm* ptr = (Eterm*) WSTACK_POP(stack);
-	    if (i != 0) {
-		term = *ptr;
-		WSTACK_PUSH3(stack, (UWord)(ptr+1), (UWord) i-1, (UWord) op);
-		goto tail_recur;
-	    }
-	    if (op == MAKE_HASH_TUPLE_OP) {
-		Uint32 arity = (UWord) WSTACK_POP(stack);
-		hash = hash*FUNNY_NUMBER9 + arity;
-	    }
-	    break;
-	}
-
-    default:
-	erts_exit(ERTS_ERROR_EXIT, "Invalid tag in make_broken_hash\n");
-	return 0;
-      }
-      if (WSTACK_ISEMPTY(stack)) break;
-      op = (Uint) WSTACK_POP(stack);
-    }
-
-    DESTROY_WSTACK(stack);
-    return hash;
-
-#undef MAKE_HASH_TUPLE_OP
-#undef MAKE_HASH_TERM_ARRAY_OP
-#undef MAKE_HASH_CDR_PRE_OP
-#undef MAKE_HASH_CDR_POST_OP
-}
-
 static Eterm
 do_allocate_logger_message(Eterm gleader, Eterm **hp, ErlOffHeap **ohp,
 			   ErlHeapFragment **bp, Process **p, Uint sz)

diff --git a/erts/emulator/test/binary_SUITE.erl b/erts/emulator/test/binary_SUITE.erl
@@ -19,7 +19,6 @@
 %%
 
 -module(binary_SUITE).
--compile({nowarn_deprecated_function, {erlang,hash,2}}).
 
 %% Tests binaries and the BIFs:
 %%	list_to_binary/1
@@ -392,7 +391,6 @@ test_hash(List) ->
     Bin = list_to_binary(List),
     Sbin = make_sub_binary(List),
     Unaligned = make_unaligned_sub_binary(Sbin),
-    test_hash_1(Bin, Sbin, Unaligned, fun erlang:hash/2),
     test_hash_1(Bin, Sbin, Unaligned, fun erlang:phash/2),
     test_hash_1(Bin, Sbin, Unaligned, fun erlang:phash2/2).