Back out 21533 because it broke Encode's build in really weird ways

p4raw-id: //depot/perl@21534

diff --git a/hv.c b/hv.c
index 485e220..7a1d25b 100644 (file)
--- a/hv.c
+++ b/hv.c
@@ -20,8 +20,6 @@
 #define PERL_IN_HV_C
 #include "perl.h"
 
-#define HV_MAX_LENGTH_BEFORE_SPLIT 4
-
 STATIC HE*
 S_new_he(pTHX)
 {
@@ -315,7 +313,7 @@ S_hv_fetch_flags(pTHX_ HV *hv, const char *key, I32 klen, I32 lval, int flags)
             }
             else
                 HeKFLAGS(entry) = flags;
-            if (flags & HVhek_ENABLEHVKFLAGS)
+            if (flags)
                 HvHASKFLAGS_on(hv);
         }
         if (flags & HVhek_FREEKEY)
@@ -489,7 +487,7 @@ Perl_hv_fetch_ent(pTHX_ HV *hv, SV *keysv, I32 lval, register U32 hash)
             }
             else
                 HeKFLAGS(entry) = flags;
-            if (flags & HVhek_ENABLEHVKFLAGS)
+            if (flags)
                 HvHASKFLAGS_on(hv);
         }
        if (key != keysave)
@@ -605,7 +603,7 @@ Perl_hv_store_flags(pTHX_ HV *hv, const char *key, I32 klen, SV *val,
                  register U32 hash, int flags)
 {
     register XPVHV* xhv;
-    register U32 n_links;
+    register I32 i;
     register HE *entry;
     register HE **oentry;
 
@@ -652,10 +650,9 @@ Perl_hv_store_flags(pTHX_ HV *hv, const char *key, I32 klen, SV *val,
 
     /* oentry = &(HvARRAY(hv))[hash & (I32) HvMAX(hv)]; */
     oentry = &((HE**)xhv->xhv_array)[hash & (I32) xhv->xhv_max];
+    i = 1;
 
-    n_links = 0;
-
-    for (entry = *oentry; entry; ++n_links, entry = HeNEXT(entry)) {
+    for (entry = *oentry; entry; i=0, entry = HeNEXT(entry)) {
        if (HeHASH(entry) != hash)              /* strings can't be equal */
            continue;
        if (HeKLEN(entry) != (I32)klen)
@@ -722,17 +719,9 @@ Perl_hv_store_flags(pTHX_ HV *hv, const char *key, I32 klen, SV *val,
     *oentry = entry;
 
     xhv->xhv_keys++; /* HvKEYS(hv)++ */
-    if (!n_links) {                            /* initial entry? */
+    if (i) {                           /* initial entry? */
        xhv->xhv_fill++; /* HvFILL(hv)++ */
-    } else if ((n_links > HV_MAX_LENGTH_BEFORE_SPLIT)
-              && (!HvREHASH(hv)
-                  || (xhv->xhv_keys > (IV)xhv->xhv_max))) {
-       /* Use the old HvKEYS(hv) > HvMAX(hv) condition to limit bucket
-          splits on a rehashed hash, as we're not going to split it again,
-          and if someone is lucky (evil) enough to get all the keys in one
-          list they could exhaust our memory as we repeatedly double the
-          number of buckets on every entry. Linear search feels a less worse
-          thing to do.  */
+    } else if (xhv->xhv_keys > (IV)xhv->xhv_max /* HvKEYS(hv) > HvMAX(hv) */) {
         hsplit(hv);
     }
 
@@ -774,7 +763,7 @@ Perl_hv_store_ent(pTHX_ HV *hv, SV *keysv, SV *val, U32 hash)
     XPVHV* xhv;
     char *key;
     STRLEN klen;
-    U32 n_links;
+    I32 i;
     HE *entry;
     HE **oentry;
     bool is_utf8;
@@ -841,9 +830,9 @@ Perl_hv_store_ent(pTHX_ HV *hv, SV *keysv, SV *val, U32 hash)
 
     /* oentry = &(HvARRAY(hv))[hash & (I32) HvMAX(hv)]; */
     oentry = &((HE**)xhv->xhv_array)[hash & (I32) xhv->xhv_max];
-    n_links = 0;
+    i = 1;
     entry = *oentry;
-    for (; entry; ++n_links, entry = HeNEXT(entry)) {
+    for (; entry; i=0, entry = HeNEXT(entry)) {
        if (HeHASH(entry) != hash)              /* strings can't be equal */
            continue;
        if (HeKLEN(entry) != (I32)klen)
@@ -897,17 +886,10 @@ Perl_hv_store_ent(pTHX_ HV *hv, SV *keysv, SV *val, U32 hash)
     *oentry = entry;
 
     xhv->xhv_keys++; /* HvKEYS(hv)++ */
-    if (!n_links) {                            /* initial entry? */
+    if (i) {                           /* initial entry? */
        xhv->xhv_fill++; /* HvFILL(hv)++ */
-    } else if ((xhv->xhv_keys > (IV)xhv->xhv_max)
-              || ((n_links > HV_MAX_LENGTH_BEFORE_SPLIT) && !HvREHASH(hv))) {
-       /* Use only the old HvKEYS(hv) > HvMAX(hv) condition to limit bucket
-          splits on a rehashed hash, as we're not going to split it again,
-          and if someone is lucky (evil) enough to get all the keys in one
-          list they could exhaust our memory as we repeatedly double the
-          number of buckets on every entry. Linear search feels a less worse
-          thing to do.  */
-        hsplit(hv);
+    } else if (xhv->xhv_keys > (IV)xhv->xhv_max /* HvKEYS(hv) > HvMAX(hv) */) {
+       hsplit(hv);
     }
 
     return entry;
@@ -1529,7 +1511,7 @@ S_hsplit(pTHX_ HV *hv)
 
 
     /* Pick your policy for "hashing isn't working" here:  */
-    if (longest_chain <= HV_MAX_LENGTH_BEFORE_SPLIT /* split worked?  */
+    if (longest_chain < 8 || longest_chain * 2 < HvTOTALKEYS(hv)
        || HvREHASH(hv)) {
        return;
     }
@@ -1551,6 +1533,7 @@ S_hsplit(pTHX_ HV *hv)
     xhv->xhv_fill = 0;
     HvSHAREKEYS_off(hv);
     HvREHASH_on(hv);
+    HvHASKFLAGS_on(hv);
 
     aep = (HE **) xhv->xhv_array;
 
@@ -2402,9 +2385,6 @@ S_share_hek_flags(pTHX_ const char *str, I32 len, register U32 hash, int flags)
 
     if (!(Svp = hv_fetch(PL_strtab, str, len, FALSE)))
        hv_store(PL_strtab, str, len, Nullsv, hash);
-
-       Can't rehash the shared string table, so not sure if it's worth
-       counting the number of entries in the linked list
     */
     xhv = (XPVHV*)SvANY(PL_strtab);
     /* assert(xhv_array != 0) */

diff --git a/hv.h b/hv.h
index 5f83440..1c7fcda 100644 (file)
--- a/hv.h
+++ b/hv.h
@@ -212,7 +212,6 @@ C<SV*>.
  * is utf8 (including 8 bit keys that were entered as utf8, and need upgrading
  * when retrieved during iteration. It may still be set when there are no longer
  * any utf8 keys.
- * See HVhek_ENABLEHVKFLAGS for the trigger.
  */
 #define HvHASKFLAGS(hv)                (SvFLAGS(hv) & SVphv_HASKFLAGS)
 #define HvHASKFLAGS_on(hv)     (SvFLAGS(hv) |= SVphv_HASKFLAGS)
@@ -284,16 +283,6 @@ C<SV*>.
                                * (may change, but Storable is a core module) */
 #define HVhek_MASK     0xFF
 
-/* Which flags enable HvHASKFLAGS? Somewhat a hack on a hack, as
-   HVhek_REHASH is only needed because the rehash flag has to be duplicated
-   into all keys as hv_iternext has no access to the hash flags. At this
-   point Storable's tests get upset, because sometimes hashes are "keyed"
-   and sometimes not, depending on the order of data insertion, and whether
-   it triggered rehashing. So currently HVhek_REHAS is exempt.
-*/
-   
-#define HVhek_ENABLEHVKFLAGS   (HVhek_MASK - HVhek_REHASH)
-
 #define HEK_UTF8(hek)          (HEK_FLAGS(hek) & HVhek_UTF8)
 #define HEK_UTF8_on(hek)       (HEK_FLAGS(hek) |= HVhek_UTF8)
 #define HEK_UTF8_off(hek)      (HEK_FLAGS(hek) &= ~HVhek_UTF8)