@matheusbsbs / matheus-bagliano / commits / 2d83cc3f84

packfile: widen unpack_entry()'s size out-parameter to size_t

The topic `js/objects-larger-than-4gb-on-windows` widened the streaming, index-pack and unpack-objects paths to `size_t` but deliberately stopped at the in-memory `unpack_entry()` cascade, which still hands back the unpacked size through `unsigned long *`. On Windows that boundary truncates above 4 GiB because that data type is only 32 bits wide on that platform. Widen the code path. Except `packed_object_info_with_index_pos()`: It cannot yet pass `oi->sizep` directly because the field is still `unsigned long *`; bridge it with a `size_t` temporary that narrows back, and let a later commit drop the bridge once the field is wide too. `gfi_unpack_entry()` keeps its narrow signature because fast-import tracks sizes through `unsigned long` everywhere it crosses subsystem boundaries, keeping its signature allows the scope of this commit to be somewhat reasonable, still. Assisted-by: Opus 4.7 Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Johannes Schindelin committed Jun 15, 2026 at 11:52 UTC 2d83cc3f84594dc1fb79626c3eae4af3e942e882

4 files changed +27 -16

builtin/fast-import.c

+6 -1

@@ -1239,6 +1239,8 @@ static void *gfi_unpack_entry(

1239 unsigned long *sizep)

1240 {

1241 enum object_type type;

1242 + size_t size_st = 0;

1243 + void *data;

1244 struct packed_git *p = all_packs[oe->pack_id];

1245 if (p == pack_data && p->pack_size < (pack_size + the_hash_algo->rawsz)) {

1246 /* The object is stored in the packfile we are writing to

@@ -1260,7 +1262,10 @@ static void *gfi_unpack_entry(

1262 */

1263 p->pack_size = pack_size + the_hash_algo->rawsz;

1264 }

1263 - return unpack_entry(the_repository, p, oe->idx.offset, &type, sizep);

1265 + data = unpack_entry(the_repository, p, oe->idx.offset, &type, &size_st);

1266 + if (sizep)

1267 + *sizep = cast_size_t_to_ulong(size_st);

1268 + return data;

1269 }

1270

1271 static void load_tree(struct tree_entry *root)

pack-check.c

+2 -3

@@ -143,9 +143,8 @@ static int verify_packfile(struct repository *r,

143 data = NULL;

144 data_valid = 0;

145 } else {

146 - unsigned long sz;

147 - data = unpack_entry(r, p, entries[i].offset, &type, &sz);

148 - size = sz;

146 + data = unpack_entry(r, p, entries[i].offset, &type,

147 + &size);

148 data_valid = 1;

149 }

150

packfile.c

+17 -11

@@ -1454,7 +1454,7 @@ struct delta_base_cache_entry {

1454 struct delta_base_cache_key key;

1455 struct list_head lru;

1456 void *data;

1457 - unsigned long size;

1457 + size_t size;

1458 enum object_type type;

1459 };

1460

@@ -1525,7 +1525,7 @@ static void detach_delta_base_cache_entry(struct delta_base_cache_entry *ent)

1525 }

1526

1527 static void *cache_or_unpack_entry(struct repository *r, struct packed_git *p,

1528 - off_t base_offset, unsigned long *base_size,

1528 + off_t base_offset, size_t *base_size,

1529 enum object_type *type)

1530 {

1531 struct delta_base_cache_entry *ent;

@@ -1558,8 +1558,8 @@ void clear_delta_base_cache(void)

1558 }

1559

1560 static void add_delta_base_cache(struct packed_git *p, off_t base_offset,

1561 - void *base, unsigned long base_size,

1562 - unsigned long delta_base_cache_limit,

1561 + void *base, size_t base_size,

1562 + size_t delta_base_cache_limit,

1563 enum object_type type)

1564 {

1565 struct delta_base_cache_entry *ent;

@@ -1614,10 +1614,13 @@ static int packed_object_info_with_index_pos(struct packed_git *p, off_t obj_off

1614 * a "real" type later if the caller is interested.

1615 */

1616 if (oi->contentp) {

1617 - *oi->contentp = cache_or_unpack_entry(p->repo, p, obj_offset, oi->sizep,

1618 - &type);

1617 + size_t size_st = 0;

1618 + *oi->contentp = cache_or_unpack_entry(p->repo, p, obj_offset,

1619 + &size_st, &type);

1620 if (!*oi->contentp)

1621 type = OBJ_BAD;

1622 + else if (oi->sizep)

1623 + *oi->sizep = cast_size_t_to_ulong(size_st);

1624 } else if (oi->sizep || oi->typep || oi->delta_base_oid) {

1625 type = unpack_object_header(p, &w_curs, &curpos, &size);

1626 }

@@ -1735,7 +1738,7 @@ int packed_object_info(struct packed_git *p, off_t obj_offset,

1738 static void *unpack_compressed_entry(struct packed_git *p,

1739 struct pack_window **w_curs,

1740 off_t curpos,

1738 - unsigned long size)

1741 + size_t size)

1742 {

1743 int st;

1744 git_zstream stream;

@@ -1790,11 +1793,11 @@ int do_check_packed_object_crc;

1793 struct unpack_entry_stack_ent {

1794 off_t obj_offset;

1795 off_t curpos;

1793 - unsigned long size;

1796 + size_t size;

1797 };

1798

1799 void *unpack_entry(struct repository *r, struct packed_git *p, off_t obj_offset,

1797 - enum object_type *final_type, unsigned long *final_size)

1800 + enum object_type *final_type, size_t *final_size)

1801 {

1802 struct pack_window *w_curs = NULL;

1803 off_t curpos = obj_offset;

@@ -1911,7 +1914,7 @@ void *unpack_entry(struct repository *r, struct packed_git *p, off_t obj_offset,

1914 void *delta_data;

1915 void *base = data;

1916 void *external_base = NULL;

1914 - unsigned long delta_size, base_size = size;

1917 + size_t delta_size, base_size = size;

1918 int i;

1919 off_t base_obj_offset = obj_offset;

1920

@@ -1928,6 +1931,7 @@ void *unpack_entry(struct repository *r, struct packed_git *p, off_t obj_offset,

1931 struct object_id base_oid;

1932 if (!(offset_to_pack_pos(p, obj_offset, &pos))) {

1933 struct object_info oi = OBJECT_INFO_INIT;

1934 + unsigned long bsz_ul = 0;

1935

1936 nth_packed_object_id(&base_oid, p,

1937 pack_pos_to_index(p, pos));

@@ -1938,11 +1942,13 @@ void *unpack_entry(struct repository *r, struct packed_git *p, off_t obj_offset,

1942 mark_bad_packed_object(p, &base_oid);

1943

1944 oi.typep = &type;

1941 - oi.sizep = &base_size;

1945 + oi.sizep = &bsz_ul;

1946 oi.contentp = &base;

1947 if (odb_read_object_info_extended(r->objects, &base_oid,

1948 &oi, 0) < 0)

1949 base = NULL;

1950 + else

1951 + base_size = bsz_ul;

1952

1953 external_base = base;

1954 }

packfile.h

+2 -1

@@ -455,7 +455,8 @@ off_t nth_packed_object_offset(const struct packed_git *, uint32_t n);

455 off_t find_pack_entry_one(const struct object_id *oid, struct packed_git *);

456

457 int is_pack_valid(struct packed_git *);

458 -void *unpack_entry(struct repository *r, struct packed_git *, off_t, enum object_type *, unsigned long *);

458 +void *unpack_entry(struct repository *r, struct packed_git *, off_t,

459 + enum object_type *, size_t *);

460 unsigned long unpack_object_header_buffer(const unsigned char *buf, unsigned long len, enum object_type *type, size_t *sizep);

461 unsigned long get_size_from_delta(struct packed_git *, struct pack_window **, off_t);

462 int unpack_object_header(struct packed_git *, struct pack_window **, off_t *, size_t *);