The very aggressive -Ooverlap-strings

code.c

@@ -59,7 +59,8 @@ void code_init() {
     int                    i               = 0;
 
     code_entfields = 0;
-    code_string_cache = util_htnew(1024);
+
+    code_string_cache = util_htnew(OPTS_OPTIMIZATION(OPTIM_OVERLAP_STRINGS) ? 0x100 : 1024);
 
     /*
      * The way progs.dat is suppose to work is odd, there needs to be
@@ -75,6 +76,7 @@ void code_init() {
     vec_push(code_fields,     empty_def);
 }
 
+void *code_util_str_htgeth(hash_table_t *ht, const char *key, size_t bin);
 uint32_t code_genstring(const char *str)
 {
     uint32_t off;
@@ -92,8 +94,14 @@ uint32_t code_genstring(const char *str)
         return code_string_cached_empty;
     }
 
-    hash = util_hthash(code_string_cache, str);
-    existing = util_htgeth(code_string_cache, str, hash);
+    if (OPTS_OPTIMIZATION(OPTIM_OVERLAP_STRINGS)) {
+        hash = ((unsigned char*)str)[strlen(str)-1];
+        existing = code_util_str_htgeth(code_string_cache, str, hash);
+    } else {
+        hash = util_hthash(code_string_cache, str);
+        existing = util_htgeth(code_string_cache, str, hash);
+    }
+
     if (existing)
         return HASH_ENTRY_TO_QCINT(existing);
 
@@ -101,7 +109,6 @@ uint32_t code_genstring(const char *str)
     vec_upload(code_chars, str, strlen(str)+1);
 
     util_htseth(code_string_cache, str, hash, QCINT_TO_HASH_ENTRY(off));
-    existing = util_htgeth(code_string_cache, str, hash);
     return off;
 }
 

doc/gmqcc.1

@@ -389,6 +389,24 @@ possible.
 Don't generate defs for immediate values or even declared constants.
 Meaning variables which are implicitly constant or qualified as such
 using the 'const' keyword.
+.TP
+.B -Ooverlap-strings
+Aggressively reuse strings in the string section. When a string should
+be added which is the trailing substring of an already existing
+string, the existing string's tail will be returned instead of the new
+string being added.
+
+For example the following code will only generate 1 string:
+
+.in +4
+.nf
+print("Hell you!\\n");
+print("you!\\n"); // trailing substring of "Hello you!\\n"
+.fi
+.in
+There's however one limitation. Strings are still processed in order,
+so if the above print statements were reversed, this optimization
+would not happen.
 .SH CONFIG
 The configuration file is similar to regular .ini files. Comments
 start with hashtags or semicolons, sections are written in square

opts.def

@@ -85,6 +85,7 @@
     GMQCC_DEFINE_FLAG(TAIL_CALLS,           2)
     GMQCC_DEFINE_FLAG(OVERLAP_LOCALS,       3)
     GMQCC_DEFINE_FLAG(STRIP_CONSTANT_NAMES, 1)
+    GMQCC_DEFINE_FLAG(OVERLAP_STRINGS,      2)
 #endif
 
 /* some cleanup so we don't have to */

util.c

@@ -514,6 +514,40 @@ void *util_htget(hash_table_t *ht, const char *key) {
     return util_htgeth(ht, key, util_hthash(ht, key));
 }
 
+void *code_util_str_htgeth(hash_table_t *ht, const char *key, size_t bin) {
+    hash_node_t *pair;
+    size_t len, keylen;
+    int cmp;
+
+    keylen = strlen(key);
+
+    pair = ht->table[bin];
+    while (pair && pair->key) {
+        len = strlen(pair->key);
+        if (len < keylen) {
+            pair = pair->next;
+            continue;
+        }
+        if (keylen == len) {
+            cmp = strcmp(key, pair->key);
+            if (cmp == 0)
+                return pair->value;
+            if (cmp < 0)
+                return NULL;
+            pair = pair->next;
+            continue;
+        }
+        cmp = strcmp(key, pair->key + len - keylen);
+        if (cmp == 0) {
+            uintptr_t up = (uintptr_t)pair->value;
+            up += len - keylen;
+            return (void*)up;
+        }
+        pair = pair->next;
+    }
+    return NULL;
+}
+
 /*
  * Free all allocated data in a hashtable, this is quite the amount
  * of work.