simon/post-scarcity
1
0
Fork 0
Code Issues 4 Pull requests Projects Releases Packages Wiki Activity Actions

Greatly improved performance of equal for string like things.

Browse source
This commit is contained in:
Simon Brooke 2026-02-25 22:10:37 +00:00
parent 90e862cc59
commit ffceda5edc
3 changed files with 39 additions and 16 deletions
Download patch file Download diff file Expand all files Collapse all files

47
src/ops/equal.c
View file

@ -9,6 +9,7 @@
#include <math.h> #include <math.h>
#include <stdbool.h> #include <stdbool.h>
#include <string.h>
#include "arith/integer.h" #include "arith/integer.h"
#include "arith/peano.h" #include "arith/peano.h"
@ -363,22 +364,38 @@ bool equal( struct cons_pointer a, struct cons_pointer b ) {
/* TODO: it is not OK to do this on the stack since list-like /* TODO: it is not OK to do this on the stack since list-like
* structures can be of indefinite extent. It *must* be done by * structures can be of indefinite extent. It *must* be done by
* iteration (and even that is problematic) */ * iteration (and even that is problematic) */
if (cell_a->payload.string.hash == cell_b->payload.string.hash) {
wchar_t a_buff[ STRING_SHIPYARD_SIZE], b_buff[ STRING_SHIPYARD_SIZE];
uint32_t tag = cell_a->tag.value;
int i = 0;
memset(a_buff,0,sizeof(a_buff));
memset(b_buff,0,sizeof(b_buff));
for (; (i < (STRING_SHIPYARD_SIZE - 1)) && !nilp( a) && !nilp( b); i++) {
a_buff[i] = cell_a->payload.string.character;
a = c_cdr(a);
cell_a = &pointer2cell( a );
b_buff[i] = cell_b->payload.string.character;
b = c_cdr( b);
cell_b = &pointer2cell( b);
}
#ifdef DEBUG #ifdef DEBUG
debug_print( L"Comparing '", DEBUG_ARITH); debug_print( L"Comparing '", DEBUG_LAMBDA);
debug_print_object( a, DEBUG_ARITH); debug_print( a_buff, DEBUG_LAMBDA);
debug_print( L"' to '", DEBUG_ARITH); debug_print( L"' to '", DEBUG_LAMBDA);
debug_print_object( b, DEBUG_ARITH); debug_print( b_buff, DEBUG_LAMBDA);
debug_print( L"'\n", DEBUG_LAMBDA);
#endif #endif
result =
cell_a->payload.string.hash == cell_b->payload.string.hash /* OK, now we have wchar string buffers loaded from the objects. We
&& cell_a->payload.string.character == * may not have exhausted either string, so the buffers being equal
cell_b->payload.string.character * isn't sufficient. So we recurse at least once. */
&&
( equal result = (wcsncmp( a_buff, b_buff, i) == 0) && equal( c_cdr(a), c_cdr(b));
( cell_a->payload.string.cdr, }
cell_b->payload.string.cdr )
|| ( end_of_string( cell_a->payload.string.cdr )
&& end_of_string( cell_b->payload.string.cdr ) ) );
break; break;
case VECTORPOINTTV: case VECTORPOINTTV:
if ( cell_b->tag.value == VECTORPOINTTV) { if ( cell_b->tag.value == VECTORPOINTTV) {
@ -403,7 +420,7 @@ bool equal( struct cons_pointer a, struct cons_pointer b ) {
* I'll ignore them, too, for now. * I'll ignore them, too, for now.
*/ */
debug_printf( DEBUG_ARITH, L"\nequal returning %d\n", result ); debug_printf( DEBUG_LAMBDA, L"\nequal returning %d\n", result );
return result; return result;
} }

6
src/ops/equal.h
View file

@ -15,6 +15,12 @@
#ifndef __equal_h #ifndef __equal_h
#define __equal_h #define __equal_h
/**
* size of buffer for assembling strings. Likely to be useful to
* read, too.
*/
#define STRING_SHIPYARD_SIZE 1024
/** /**
* Shallow, and thus cheap, equality: true if these two objects are * Shallow, and thus cheap, equality: true if these two objects are
* the same object, else false. * the same object, else false.