Use hash table for global name lookups (bug 4496, r=fyren).

sourcepawn/compiler/AMBuilder

@@ -53,7 +53,8 @@ files = [
 	'zlib/inftrees.c',
 	'zlib/trees.c',
 	'zlib/uncompr.c',
-	'zlib/zutil.c'
+	'zlib/zutil.c',
+	'sp_symhash.c'
 	]
 if AMBuild.target['platform'] == 'linux':
 	files.append('binreloc.c')

sourcepawn/compiler/sc.h

@@ -565,7 +565,6 @@ SC_FUNC void delete_symbol(symbol *root,symbol *sym);
 SC_FUNC void delete_symbols(symbol *root,int level,int del_labels,int delete_functions);
 SC_FUNC int refer_symbol(symbol *entry,symbol *bywhom);
 SC_FUNC void markusage(symbol *sym,int usage);
-SC_FUNC uint32_t namehash(const char *name);
 SC_FUNC symbol *findglb(const char *name,int filter);
 SC_FUNC symbol *findloc(const char *name);
 SC_FUNC symbol *findconst(const char *name,int *matchtag);
@@ -771,6 +770,8 @@ SC_FUNC int state_conflict_id(int listid1,int listid2);
 
 /* external variables (defined in scvars.c) */
 #if !defined SC_SKIP_VDECL
+typedef struct HashTable HashTable;
+SC_VDECL struct HashTable *sp_Globals;
 SC_VDECL symbol loctab;       /* local symbol table */
 SC_VDECL symbol glbtab;       /* global symbol table */
 SC_VDECL cell *litq;          /* the literal queue */

sourcepawn/compiler/sc1.c

@@ -68,6 +68,7 @@
 #include "sc.h"
 #include <sourcemod_version.h>
 #include "sctracker.h"
+#include "sp_symhash.h"
 #define VERSION_STR "3.2.3636"
 #define VERSION_INT 0x0302
 
@@ -208,6 +209,10 @@ int pc_compile(int argc, char *argv[])
   if ((jmpcode=setjmp(errbuf))!=0)
     goto cleanup;
 
+  sp_Globals = NewHashTable();
+  if (!sp_Globals)
+    error(123);
+
   /* allocate memory for fixed tables */
   inpfname=(char*)malloc(_MAX_PATH);
   if (inpfname==NULL)
@@ -511,6 +516,7 @@ cleanup:
   delete_symbols(&loctab,0,TRUE,TRUE);    /* delete local variables if not yet
                                            * done (i.e. on a fatal error) */
   delete_symbols(&glbtab,0,TRUE,TRUE);
+  DestroyHashTable(sp_Globals);
   delete_consttable(&tagname_tab);
   delete_consttable(&libname_tab);
   delete_consttable(&sc_automaton_tab);
@@ -3837,8 +3843,10 @@ static int operatoradjust(int opertok,symbol *sym,char *opername,int resulttag)
         refer_symbol(sym,oldsym->refer[i]);
     delete_symbol(&glbtab,oldsym);
   } /* if */
+  RemoveFromHashTable(sp_Globals, sym);
   strcpy(sym->name,tmpname);
-  sym->hash=namehash(sym->name);/* calculate new hash */
+  sym->hash=NameHash(sym->name);/* calculate new hash */
+  AddToHashTable(sp_Globals, sym);
 
   /* operators should return a value, except the '~' operator */
   if (opertok!='~')

sourcepawn/compiler/sc2.c

@@ -31,6 +31,7 @@
 #if defined LINUX || defined __FreeBSD__ || defined __OpenBSD__
   #include <sclinux.h>
 #endif
+#include "sp_symhash.h"
 
 #if defined FORTIFY
   #include <alloc/fortify.h>
@@ -2392,6 +2393,7 @@ SC_FUNC int ishex(char c)
 static symbol *add_symbol(symbol *root,symbol *entry,int sort)
 {
   symbol *newsym;
+  int global = root==&glbtab;
 
   if (sort)
     while (root->next!=NULL && strcmp(entry->name,root->next->name)>0)
@@ -2404,6 +2406,8 @@ static symbol *add_symbol(symbol *root,symbol *entry,int sort)
   memcpy(newsym,entry,sizeof(symbol));
   newsym->next=root->next;
   root->next=newsym;
+  if (global)
+    AddToHashTable(sp_Globals, newsym);
   return newsym;
 }
 
@@ -2453,6 +2457,7 @@ static void free_symbol(symbol *sym)
 
 SC_FUNC void delete_symbol(symbol *root,symbol *sym)
 {
+  symbol *origRoot=root;
   /* find the symbol and its predecessor
    * (this function assumes that you will never delete a symbol that is not
    * in the table pointed at by "root")
@@ -2463,6 +2468,9 @@ SC_FUNC void delete_symbol(symbol *root,symbol *sym)
     assert(root!=NULL);
   } /* while */
 
+  if (origRoot==&glbtab)
+    RemoveFromHashTable(sp_Globals, sym);
+
   /* unlink it, then free it */
   root->next=sym->next;
   free_symbol(sym);
@@ -2470,6 +2478,7 @@ SC_FUNC void delete_symbol(symbol *root,symbol *sym)
 
 SC_FUNC void delete_symbols(symbol *root,int level,int delete_labels,int delete_functions)
 {
+  symbol *origRoot=root;
   symbol *sym,*parent_sym;
   constvalue *stateptr;
   int mustdelete;
@@ -2525,6 +2534,8 @@ SC_FUNC void delete_symbols(symbol *root,int level,int delete_labels,int delete_
       break;
     } /* switch */
     if (mustdelete) {
+      if (origRoot == &glbtab)
+        RemoveFromHashTable(sp_Globals, sym);
       root->next=sym->next;
       free_symbol(sym);
     } else {
@@ -2549,26 +2560,12 @@ SC_FUNC void delete_symbols(symbol *root,int level,int delete_labels,int delete_
   } /* if */
 }
 
-/* The purpose of the hash is to reduce the frequency of a "name"
- * comparison (which is costly). There is little interest in avoiding
- * clusters in similar names, which is why this function is plain simple.
- */
-SC_FUNC uint32_t namehash(const char *name)
-{
-  const unsigned char *ptr=(const unsigned char *)name;
-  int len=strlen(name);
-  if (len==0)
-    return 0L;
-  assert(len<256);
-  return (len<<24Lu) + (ptr[0]<<16Lu) + (ptr[len-1]<<8Lu) + (ptr[len>>1Lu]);
-}
-
 static symbol *find_symbol(const symbol *root,const char *name,int fnumber,int automaton,int *cmptag)
 {
   symbol *firstmatch=NULL;
   symbol *sym=root->next;
   int count=0;
-  unsigned long hash=namehash(name);
+  unsigned long hash=NameHash(name);
   while (sym!=NULL) {
     if (hash==sym->hash && strcmp(name,sym->name)==0        /* check name */
         && (sym->parent==NULL || sym->ident==iCONSTEXPR)    /* sub-types (hierarchical types) are skipped, except for enum fields */
@@ -2705,7 +2702,7 @@ SC_FUNC symbol *findglb(const char *name,int filter)
    * that has no state(s) attached to it
    */
   if (sym==NULL)
-    sym=find_symbol(&glbtab,name,fcurrent,-1,NULL);
+    sym=FindInHashTable(sp_Globals,name,fcurrent);
   return sym;
 }
 
@@ -2724,8 +2721,12 @@ SC_FUNC symbol *findconst(const char *name,int *cmptag)
   symbol *sym;
 
   sym=find_symbol(&loctab,name,-1,-1,cmptag);  /* try local symbols first */
-  if (sym==NULL || sym->ident!=iCONSTEXPR)     /* not found, or not a constant */
+  if (sym==NULL || sym->ident!=iCONSTEXPR) {   /* not found, or not a constant */
-    sym=find_symbol(&glbtab,name,fcurrent,-1,cmptag);
+    if (cmptag)
+      sym=FindTaggedInHashTable(sp_Globals,name,fcurrent,cmptag);
+    else
+      sym=FindInHashTable(sp_Globals,name,fcurrent);
+  }
   if (sym==NULL || sym->ident!=iCONSTEXPR)
     return NULL;
   assert(sym->parent==NULL || (sym->usage & uENUMFIELD)!=0);
@@ -2765,7 +2766,7 @@ SC_FUNC symbol *addsym(const char *name,cell addr,int ident,int vclass,int tag,i
   /* first fill in the entry */
   memset(&entry,0,sizeof entry);
   strcpy(entry.name,name);
-  entry.hash=namehash(name);
+  entry.hash=NameHash(name);
   entry.addr=addr;
   entry.codeaddr=code_idx;
   entry.vclass=(char)vclass;
@@ -2780,8 +2781,7 @@ SC_FUNC symbol *addsym(const char *name,cell addr,int ident,int vclass,int tag,i
   /* then insert it in the list */
   if (vclass==sGLOBAL)
     return add_symbol(&glbtab,&entry,TRUE);
-  else
+  return add_symbol(&loctab,&entry,FALSE);
-    return add_symbol(&loctab,&entry,FALSE);
 }
 
 SC_FUNC symbol *addvariable(const char *name,cell addr,int ident,int vclass,int tag,

sourcepawn/compiler/scvars.c

@@ -25,6 +25,7 @@
 #include <stdio.h>
 #include <stdlib.h>     /* for _MAX_PATH */
 #include "sc.h"
+#include "sp_symhash.h"
 
 /*  global variables
  *
@@ -100,6 +101,8 @@ SC_VDEFINE FILE *outf    = NULL;   /* (intermediate) text file written to */
 
 SC_VDEFINE jmp_buf errbuf;
 
+SC_VDEFINE HashTable *sp_Globals = NULL;
+
 #if !defined SC_LIGHT
   SC_VDEFINE int sc_makereport=FALSE; /* generate a cross-reference report */
 #endif

sourcepawn/compiler/sp_symhash.c

@@ -0,0 +1,229 @@
+/* vim: set ts=4 sw=4 tw=99 et: */
+#include <stdlib.h>
+#include <string.h>
+#include <assert.h>
+#include "sc.h"
+#include "sp_symhash.h"
+
+SC_FUNC uint32_t
+NameHash(const char *str)
+{
+  size_t len = strlen(str);
+  const uint8_t *data = (uint8_t *)str;
+  #undef get16bits
+  #if (defined(__GNUC__) && defined(__i386__)) || defined(__WATCOMC__) \
+      || defined(_MSC_VER) || defined (__BORLANDC__) || defined (__TURBOC__)
+    #define get16bits(d) (*((const uint16_t *) (d)))
+  #endif
+  #if !defined (get16bits)
+    #define get16bits(d) ((((uint32_t)(((const uint8_t *)(d))[1])) << 8)\
+                     +(uint32_t)(((const uint8_t *)(d))[0]) )
+  #endif
+  uint32_t hash = len, tmp;
+  int rem;
+
+  if (len <= 0 || data == NULL) return 0;
+
+  rem = len & 3;
+  len >>= 2;
+
+  /* Main loop */
+  for (;len > 0; len--) {
+      hash  += get16bits (data);
+      tmp    = (get16bits (data+2) << 11) ^ hash;
+      hash   = (hash << 16) ^ tmp;
+      data  += 2*sizeof (uint16_t);
+      hash  += hash >> 11;
+  }
+
+  /* Handle end cases */
+  switch (rem) {
+    case 3: hash += get16bits (data);
+      hash ^= hash << 16;
+      hash ^= data[sizeof (uint16_t)] << 18;
+      hash += hash >> 11;
+      break;
+    case 2: hash += get16bits (data);
+      hash ^= hash << 11;
+      hash += hash >> 17;
+      break;
+    case 1: hash += *data;
+      hash ^= hash << 10;
+      hash += hash >> 1;
+  }
+
+  /* Force "avalanching" of final 127 bits */
+  hash ^= hash << 3;
+  hash += hash >> 5;
+  hash ^= hash << 4;
+  hash += hash >> 17;
+  hash ^= hash << 25;
+  hash += hash >> 6;
+
+  return hash;
+
+  #undef get16bits
+}
+
+SC_FUNC HashTable *NewHashTable()
+{
+    HashTable *ht = (HashTable*)malloc(sizeof(HashTable));
+    if (!ht)
+        return ht;
+    ht->buckets = (HashEntry **)calloc(32, sizeof(HashEntry *));
+    if (!ht->buckets) {
+        free(ht);
+        return NULL;
+    }
+    ht->nbuckets = 32;
+    ht->nused = 0;
+    ht->bucketmask = 32 - 1;
+    return ht;
+}
+
+SC_FUNC void
+DestroyHashTable(HashTable *ht)
+{
+    uint32_t i;
+    if (!ht)
+        return;
+    for (i = 0; i < ht->nbuckets; i++) {
+        HashEntry *he = ht->buckets[i];
+        while (he != NULL) {
+            HashEntry *next = he->next;
+            free(he);
+            he = next;
+        }
+    }
+    free(ht->buckets);
+    free(ht);
+}
+
+SC_FUNC symbol *
+FindTaggedInHashTable(HashTable *ht, const char *name, int fnumber,
+                      int *cmptag)
+{
+    int count=0;
+    symbol *firstmatch=NULL;
+    uint32_t hash = NameHash(name);
+    uint32_t bucket = hash & ht->bucketmask;
+    HashEntry *he = ht->buckets[bucket];
+
+    assert(cmptag!=NULL);
+
+    while (he != NULL) {
+        symbol *sym = he->sym;
+        if ((sym->parent==NULL || sym->ident==iCONSTEXPR) &&
+            (sym->fnumber<0 || sym->fnumber==fnumber) &&
+            (strcmp(sym->name, name) == 0))
+        {
+            /* return closest match or first match; count number of matches */
+            if (firstmatch==NULL)
+                firstmatch=sym;
+            if (*cmptag==0)
+                count++;
+            if (*cmptag==sym->tag) {
+                *cmptag=1;    /* good match found, set number of matches to 1 */
+                return sym;
+            }
+        }
+        he = he->next;
+    }
+
+    if (firstmatch!=NULL)
+        *cmptag=count;
+    return firstmatch;
+}
+
+SC_FUNC symbol *
+FindInHashTable(HashTable *ht, const char *name, int fnumber)
+{
+    uint32_t hash = NameHash(name);
+    uint32_t bucket = hash & ht->bucketmask;
+    HashEntry *he = ht->buckets[bucket];
+
+    while (he != NULL) {
+        symbol *sym = he->sym;
+        if ((sym->parent==NULL || sym->ident==iCONSTEXPR) &&
+            (sym->fnumber<0 || sym->fnumber==fnumber) &&
+            (strcmp(sym->name, name) == 0))
+        {
+            return sym;
+        }
+        he = he->next;
+    }
+
+    return NULL;
+}
+
+static void
+ResizeHashTable(HashTable *ht)
+{
+    uint32_t i;
+    uint32_t xnbuckets = ht->nbuckets * 2;
+    uint32_t xbucketmask = xnbuckets - 1;
+    HashEntry **xbuckets = (HashEntry **)calloc(xnbuckets, sizeof(HashEntry*));
+    if (!xbuckets)
+        return;
+
+    for (i = 0; i < ht->nbuckets; i++) {
+        HashEntry *he = ht->buckets[i];
+        while (he != NULL) {
+            HashEntry *next = he->next;
+            uint32_t bucket = he->sym->hash & xbucketmask;
+            he->next = xbuckets[bucket];
+            xbuckets[bucket] = he;
+            he = next;
+        }
+    }
+    free(ht->buckets);
+    ht->buckets = xbuckets;
+    ht->nbuckets = xnbuckets;
+    ht->bucketmask = xbucketmask;
+}
+
+SC_FUNC void
+AddToHashTable(HashTable *ht, symbol *sym)
+{
+    uint32_t bucket = sym->hash & ht->bucketmask;
+    HashEntry **hep;
+
+    hep = &ht->buckets[bucket];
+    while (*hep) {
+        assert((*hep)->sym != sym);
+        hep = &(*hep)->next;
+    }
+
+    HashEntry *he = (HashEntry *)malloc(sizeof(HashEntry));
+    if (!he)
+      error(123);
+    he->sym = sym;
+    he->next = NULL;
+    *hep = he;
+    ht->nused++;
+
+    if (ht->nused > ht->nbuckets && ht->nbuckets <= INT_MAX / 2)
+        ResizeHashTable(ht);
+}
+
+SC_FUNC void
+RemoveFromHashTable(HashTable *ht, symbol *sym)
+{
+    uint32_t bucket = sym->hash & ht->bucketmask;
+    HashEntry **hep = &ht->buckets[bucket];
+    HashEntry *he = *hep;
+
+    while (he != NULL) {
+        if (he->sym == sym) {
+            *hep = he->next;
+            free(he);
+            ht->nused--;
+            return;
+        }
+        hep = &he->next;
+        he = *hep;
+    }
+
+    assert(0);
+}
+

sourcepawn/compiler/sp_symhash.h

@@ -0,0 +1,28 @@
+/* vim: set ts=4 sw=4 tw=99 et: */
+#ifndef _INCLUDE_SPCOMP_SYMHASH_H_
+#define _INCLUDE_SPCOMP_SYMHASH_H_
+
+SC_FUNC uint32_t NameHash(const char *str);
+
+typedef struct HashEntry {
+    symbol *sym;
+    struct HashEntry *next;
+} HashEntry;
+
+struct HashTable {
+    uint32_t nbuckets;
+    uint32_t nused;
+    uint32_t bucketmask;
+    HashEntry **buckets;
+};
+
+SC_FUNC HashTable *NewHashTable();
+SC_FUNC void DestroyHashTable(HashTable *ht);
+SC_FUNC void AddToHashTable(HashTable *ht, symbol *sym);
+SC_FUNC void RemoveFromHashTable(HashTable *ht, symbol *sym);
+SC_FUNC symbol *FindInHashTable(HashTable *ht, const char *name, int fnumber);
+SC_FUNC symbol *FindTaggedInHashTable(HashTable *ht, const char *name, int fnumber,
+                                      int *cmptag);
+
+#endif /* _INCLUDE_SPCOMP_SYMHASH_H_ */
+