xdiff: split xrecord_t.ha into line_hash and minimal_perfect_hash
The ha field is serving two different purposes, which makes the code harder to read. At first glance, it looks like many places assume there could never be hash collisions between lines of the two input files. In reality, line_hash is used together with xdl_recmatch() to ensure correct comparisons of lines, even when collisions occur. To make this clearer, the old ha field has been split: * line_hash: a straightforward hash of a line, independent of any external context. Its type is uint64_t, as it comes from a fixed width hash function. * minimal_perfect_hash: Not a new concept, but now a separate field. It comes from the classifier's general-purpose hash table, which assigns each line a unique and minimal hash across the two files. A size_t is used here because it's meant to be used to index an array. This also avoids ` as usize` casts on the Rust side when using it to index a slice. Signed-off-by: Ezekiel Newren <ezekielnewren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Ezekiel Newren committed
Nov 18, 2025 at 22:34 UTC
6a26019c81faa07ba811541b4cf35be9e8ee1ead
5 files changed
+21
-20
xdiff/xdiffi.c
+3
-3
@@ -22,9 +22,9 @@
22
23
#include "xinclude.h"
24
25
-static unsigned long get_hash(xdfile_t *xdf, long index)
25
+static size_t get_hash(xdfile_t *xdf, long index)
26
{
27
- return xdf->recs[xdf->rindex[index]].ha;
27
+ return xdf->recs[xdf->rindex[index]].minimal_perfect_hash;
28
}
29
30
#define XDL_MAX_COST_MIN 256
@@ -385,7 +385,7 @@ static xdchange_t *xdl_add_change(xdchange_t *xscr, long i1, long i2, long chg1,
385
386
static int recs_match(xrecord_t *rec1, xrecord_t *rec2)
387
{
388
- return (rec1->ha == rec2->ha);
388
+ return rec1->minimal_perfect_hash == rec2->minimal_perfect_hash;
389
}
390
391
/*
xdiff/xhistogram.c
+2
-2
@@ -90,7 +90,7 @@ struct region {
90
91
static int cmp_recs(xrecord_t *r1, xrecord_t *r2)
92
{
93
- return r1->ha == r2->ha;
93
+ return r1->minimal_perfect_hash == r2->minimal_perfect_hash;
94
95
}
96
@@ -98,7 +98,7 @@ static int cmp_recs(xrecord_t *r1, xrecord_t *r2)
98
(cmp_recs(REC(i->env, s1, l1), REC(i->env, s2, l2)))
99
100
#define TABLE_HASH(index, side, line) \
101
- XDL_HASHLONG((REC(index->env, side, line))->ha, index->table_bits)
101
+ XDL_HASHLONG((REC(index->env, side, line))->minimal_perfect_hash, index->table_bits)
102
103
static int scanA(struct histindex *index, int line1, int count1)
104
{
xdiff/xpatience.c
+5
-5
@@ -48,7 +48,7 @@
48
struct hashmap {
49
int nr, alloc;
50
struct entry {
51
- unsigned long hash;
51
+ size_t minimal_perfect_hash;
52
/*
53
* 0 = unused entry, 1 = first line, 2 = second, etc.
54
* line2 is NON_UNIQUE if the line is not unique
@@ -101,10 +101,10 @@ static void insert_record(xpparam_t const *xpp, int line, struct hashmap *map,
101
* So we multiply ha by 2 in the hope that the hashing was
102
* "unique enough".
103
*/
104
- int index = (int)((record->ha << 1) % map->alloc);
104
+ int index = (int)((record->minimal_perfect_hash << 1) % map->alloc);
105
106
while (map->entries[index].line1) {
107
- if (map->entries[index].hash != record->ha) {
107
+ if (map->entries[index].minimal_perfect_hash != record->minimal_perfect_hash) {
108
if (++index >= map->alloc)
109
index = 0;
110
continue;
@@ -120,7 +120,7 @@ static void insert_record(xpparam_t const *xpp, int line, struct hashmap *map,
120
if (pass == 2)
121
return;
122
map->entries[index].line1 = line;
123
- map->entries[index].hash = record->ha;
123
+ map->entries[index].minimal_perfect_hash = record->minimal_perfect_hash;
124
map->entries[index].anchor = is_anchor(xpp, (const char *)map->env->xdf1.recs[line - 1].ptr);
125
if (!map->first)
126
map->first = map->entries + index;
@@ -248,7 +248,7 @@ static int match(struct hashmap *map, int line1, int line2)
248
{
249
xrecord_t *record1 = &map->env->xdf1.recs[line1 - 1];
250
xrecord_t *record2 = &map->env->xdf2.recs[line2 - 1];
251
- return record1->ha == record2->ha;
251
+ return record1->minimal_perfect_hash == record2->minimal_perfect_hash;
252
}
253
254
static int patience_diff(xpparam_t const *xpp, xdfenv_t *env,
xdiff/xprepare.c
+9
-9
@@ -93,12 +93,12 @@ static void xdl_free_classifier(xdlclassifier_t *cf) {
93
94
95
static int xdl_classify_record(unsigned int pass, xdlclassifier_t *cf, xrecord_t *rec) {
96
- long hi;
96
+ size_t hi;
97
xdlclass_t *rcrec;
98
99
- hi = (long) XDL_HASHLONG(rec->ha, cf->hbits);
99
+ hi = XDL_HASHLONG(rec->line_hash, cf->hbits);
100
for (rcrec = cf->rchash[hi]; rcrec; rcrec = rcrec->next)
101
- if (rcrec->rec.ha == rec->ha &&
101
+ if (rcrec->rec.line_hash == rec->line_hash &&
102
xdl_recmatch((const char *)rcrec->rec.ptr, (long)rcrec->rec.size,
103
(const char *)rec->ptr, (long)rec->size, cf->flags))
104
break;
@@ -120,7 +120,7 @@ static int xdl_classify_record(unsigned int pass, xdlclassifier_t *cf, xrecord_t
120
121
(pass == 1) ? rcrec->len1++ : rcrec->len2++;
122
123
- rec->ha = (unsigned long) rcrec->idx;
123
+ rec->minimal_perfect_hash = (size_t)rcrec->idx;
124
125
return 0;
126
}
@@ -158,7 +158,7 @@ static int xdl_prepare_ctx(unsigned int pass, mmfile_t *mf, long narec, xpparam_
158
crec = &xdf->recs[xdf->nrec++];
159
crec->ptr = prev;
160
crec->size = cur - prev;
161
- crec->ha = hav;
161
+ crec->line_hash = hav;
162
if (xdl_classify_record(pass, cf, crec) < 0)
163
goto abort;
164
}
@@ -290,7 +290,7 @@ static int xdl_cleanup_records(xdlclassifier_t *cf, xdfile_t *xdf1, xdfile_t *xd
290
if ((mlim = xdl_bogosqrt(xdf1->nrec)) > XDL_MAX_EQLIMIT)
291
mlim = XDL_MAX_EQLIMIT;
292
for (i = xdf1->dstart, recs = &xdf1->recs[xdf1->dstart]; i <= xdf1->dend; i++, recs++) {
293
- rcrec = cf->rcrecs[recs->ha];
293
+ rcrec = cf->rcrecs[recs->minimal_perfect_hash];
294
nm = rcrec ? rcrec->len2 : 0;
295
action1[i] = (nm == 0) ? DISCARD: (nm >= mlim && !need_min) ? INVESTIGATE: KEEP;
296
}
@@ -298,7 +298,7 @@ static int xdl_cleanup_records(xdlclassifier_t *cf, xdfile_t *xdf1, xdfile_t *xd
298
if ((mlim = xdl_bogosqrt(xdf2->nrec)) > XDL_MAX_EQLIMIT)
299
mlim = XDL_MAX_EQLIMIT;
300
for (i = xdf2->dstart, recs = &xdf2->recs[xdf2->dstart]; i <= xdf2->dend; i++, recs++) {
301
- rcrec = cf->rcrecs[recs->ha];
301
+ rcrec = cf->rcrecs[recs->minimal_perfect_hash];
302
nm = rcrec ? rcrec->len1 : 0;
303
action2[i] = (nm == 0) ? DISCARD: (nm >= mlim && !need_min) ? INVESTIGATE: KEEP;
304
}
@@ -350,7 +350,7 @@ static int xdl_trim_ends(xdfile_t *xdf1, xdfile_t *xdf2) {
350
recs2 = xdf2->recs;
351
for (i = 0, lim = XDL_MIN(xdf1->nrec, xdf2->nrec); i < lim;
352
i++, recs1++, recs2++)
353
- if (recs1->ha != recs2->ha)
353
+ if (recs1->minimal_perfect_hash != recs2->minimal_perfect_hash)
354
break;
355
356
xdf1->dstart = xdf2->dstart = i;
@@ -358,7 +358,7 @@ static int xdl_trim_ends(xdfile_t *xdf1, xdfile_t *xdf2) {
358
recs1 = xdf1->recs + xdf1->nrec - 1;
359
recs2 = xdf2->recs + xdf2->nrec - 1;
360
for (lim -= i, i = 0; i < lim; i++, recs1--, recs2--)
361
- if (recs1->ha != recs2->ha)
361
+ if (recs1->minimal_perfect_hash != recs2->minimal_perfect_hash)
362
break;
363
364
xdf1->dend = xdf1->nrec - i - 1;
xdiff/xtypes.h
+2
-1
@@ -41,7 +41,8 @@ typedef struct s_chastore {
41
typedef struct s_xrecord {
42
uint8_t const *ptr;
43
size_t size;
44
- unsigned long ha;
44
+ uint64_t line_hash;
45
+ size_t minimal_perfect_hash;
46
} xrecord_t;
47
48
typedef struct s_xdfile {