[PATCH] hex: add and use strbuf_add_oid_hex()

[PATCH] hex: add and use strbuf_add_oid_hex()

[René Scharfe]

Add a function for adding the full hexadecimal hash value of an object
ID to a strbuf.  It's thread-safe and slightly more efficient than using
strbuf_addstr() with oid_to_hex() because it doesn't have to determine
the length of the string or copy it from the intermediate static buffer.

Add and apply a semantic patch to use it throughout the code base.

I get a tiny speedup for git log showing a single hash per commit:

Benchmark 1: ./git_main log --format=%H
  Time (mean ± σ):      91.2 ms ±   0.7 ms    [User: 51.9 ms, System: 38.6 ms]
  Range (min … max):    89.8 ms …  92.6 ms    31 runs

Benchmark 2: ./git log --format=%H
  Time (mean ± σ):      90.5 ms ±   0.7 ms    [User: 51.0 ms, System: 38.8 ms]
  Range (min … max):    89.2 ms …  92.3 ms    32 runs

Summary
  ./git log --format=%H ran
    1.01 ± 0.01 times faster than ./git_main log --format=%H

Signed-off-by: René Scharfe <l.s.r@web.de>
---
 bisect.c                      |  2 +-
 builtin/bisect.c              |  2 +-
 builtin/cat-file.c            |  5 ++---
 builtin/replace.c             |  2 +-
 convert.c                     |  2 +-
 fsck.c                        |  2 +-
 hex.c                         | 10 ++++++++++
 hex.h                         |  5 +++++
 pretty.c                      |  8 ++++----
 refs.c                        |  2 +-
 sequencer.c                   |  4 ++--
 shallow.c                     |  2 +-
 tools/coccinelle/strbuf.cocci |  6 ++++++
 transport-helper.c            |  2 +-
 14 files changed, 37 insertions(+), 17 deletions(-)

diff --git a/bisect.c b/bisect.c
index ef17a442e5..e67226a6dc 100644
--- a/bisect.c
+++ b/bisect.c
@@ -512,7 +512,7 @@ static char *join_oid_array_hex(struct oid_array *array, char delim)
 	int i;
 
 	for (i = 0; i < array->nr; i++) {
-		strbuf_addstr(&joined_hexs, oid_to_hex(array->oid + i));
+		strbuf_add_oid_hex(&joined_hexs, array->oid + i);
 		if (i + 1 < array->nr)
 			strbuf_addch(&joined_hexs, delim);
 	}
diff --git a/builtin/bisect.c b/builtin/bisect.c
index 4520e585d0..0f679e7af9 100644
--- a/builtin/bisect.c
+++ b/builtin/bisect.c
@@ -833,7 +833,7 @@ static enum bisect_error bisect_start(struct bisect_terms *terms, int argc,
 		if (!repo_get_oid(the_repository, head, &head_oid) &&
 		    !starts_with(head, "refs/heads/")) {
 			strbuf_reset(&start_head);
-			strbuf_addstr(&start_head, oid_to_hex(&head_oid));
+			strbuf_add_oid_hex(&start_head, &head_oid);
 		} else if (!repo_get_oid(the_repository, head, &head_oid) &&
 			   skip_prefix(head, "refs/heads/", &head)) {
 			strbuf_addstr(&start_head, head);
diff --git a/builtin/cat-file.c b/builtin/cat-file.c
index d9fbad5358..f015e5f415 100644
--- a/builtin/cat-file.c
+++ b/builtin/cat-file.c
@@ -320,7 +320,7 @@ static int expand_atom(struct strbuf *sb, const char *atom, int len,
 {
 	if (is_atom("objectname", atom, len)) {
 		if (!data->mark_query)
-			strbuf_addstr(sb, oid_to_hex(&data->oid));
+			strbuf_add_oid_hex(sb, &data->oid);
 	} else if (is_atom("objecttype", atom, len)) {
 		if (data->mark_query)
 			data->info.typep = &data->type;
@@ -345,8 +345,7 @@ static int expand_atom(struct strbuf *sb, const char *atom, int len,
 		if (data->mark_query)
 			data->info.delta_base_oid = &data->delta_base_oid;
 		else
-			strbuf_addstr(sb,
-				      oid_to_hex(&data->delta_base_oid));
+			strbuf_add_oid_hex(sb, &data->delta_base_oid);
 	} else if (is_atom("objectmode", atom, len)) {
 		if (!data->mark_query && !(S_IFINVALID == data->mode))
 			strbuf_addf(sb, "%06o", data->mode);
diff --git a/builtin/replace.c b/builtin/replace.c
index 4c62c5ab58..aed6b2c8de 100644
--- a/builtin/replace.c
+++ b/builtin/replace.c
@@ -127,7 +127,7 @@ static int for_each_replace_name(const char **argv, each_replace_name_fn fn)
 		}
 
 		strbuf_setlen(&ref, base_len);
-		strbuf_addstr(&ref, oid_to_hex(&oid));
+		strbuf_add_oid_hex(&ref, &oid);
 		full_hex = ref.buf + base_len;
 
 		if (refs_read_ref(get_main_ref_store(the_repository), ref.buf, &oid)) {
diff --git a/convert.c b/convert.c
index eae36c8a59..036506842c 100644
--- a/convert.c
+++ b/convert.c
@@ -1239,7 +1239,7 @@ static int ident_to_worktree(const char *src, size_t len,
 
 		/* step 4: substitute */
 		strbuf_addstr(buf, "Id: ");
-		strbuf_addstr(buf, oid_to_hex(&oid));
+		strbuf_add_oid_hex(buf, &oid);
 		strbuf_addstr(buf, " $");
 	}
 	strbuf_add(buf, src, len);
diff --git a/fsck.c b/fsck.c
index b72200c352..b4ffee6a04 100644
--- a/fsck.c
+++ b/fsck.c
@@ -344,7 +344,7 @@ const char *fsck_describe_object(struct fsck_options *options,
 	buf = bufs + b;
 	b = (b + 1) % ARRAY_SIZE(bufs);
 	strbuf_reset(buf);
-	strbuf_addstr(buf, oid_to_hex(oid));
+	strbuf_add_oid_hex(buf, oid);
 	if (name)
 		strbuf_addf(buf, " (%s)", name);
 
diff --git a/hex.c b/hex.c
index bc756722ca..f02832140d 100644
--- a/hex.c
+++ b/hex.c
@@ -3,6 +3,7 @@
 #include "git-compat-util.h"
 #include "hash.h"
 #include "hex.h"
+#include "strbuf.h"
 
 static int get_hash_hex_algop(const char *hex, unsigned char *hash,
 			      const struct git_hash_algo *algop)
@@ -122,3 +123,12 @@ char *oid_to_hex(const struct object_id *oid)
 {
 	return hash_to_hex_algop(oid->hash, &hash_algos[oid->algo]);
 }
+
+void strbuf_add_oid_hex(struct strbuf *sb, const struct object_id *oid)
+{
+	const struct git_hash_algo *algop = oid->algo ?
+		&hash_algos[oid->algo] : the_hash_algo;
+	strbuf_grow(sb, algop->hexsz);
+	hash_to_hex_algop_r(sb->buf + sb->len, oid->hash, algop);
+	strbuf_setlen(sb, sb->len + algop->hexsz);
+}
diff --git a/hex.h b/hex.h
index 1e9a65d83a..f15c7e2220 100644
--- a/hex.h
+++ b/hex.h
@@ -33,6 +33,11 @@ char *oid_to_hex_r(char *out, const struct object_id *oid);
 char *hash_to_hex_algop(const unsigned char *hash, const struct git_hash_algo *);	/* static buffer result! */
 char *oid_to_hex(const struct object_id *oid);						/* same static buffer */
 
+struct strbuf;
+
+/* Apply oid_to_hex_r() to a strbuf to append the hexadecimal hash. */
+void strbuf_add_oid_hex(struct strbuf *sb, const struct object_id *oid);
+
 /*
  * Parse a 40-character hexadecimal object ID starting from hex, updating the
  * pointer specified by end when parsing stops.  The resulting object ID is
diff --git a/pretty.c b/pretty.c
index 814803980b..2684223946 100644
--- a/pretty.c
+++ b/pretty.c
@@ -662,7 +662,7 @@ static void add_merge_info(const struct pretty_print_context *pp,
 		if (pp->abbrev)
 			strbuf_add_unique_abbrev(sb, oidp, pp->abbrev);
 		else
-			strbuf_addstr(sb, oid_to_hex(oidp));
+			strbuf_add_oid_hex(sb, oidp);
 		parent = parent->next;
 	}
 	strbuf_addch(sb, '\n');
@@ -1567,7 +1567,7 @@ static size_t format_commit_one(struct strbuf *sb, /* in UTF-8 */
 	switch (placeholder[0]) {
 	case 'H':		/* commit hash */
 		strbuf_addstr(sb, diff_get_color(c->auto_color, DIFF_COMMIT));
-		strbuf_addstr(sb, oid_to_hex(&commit->object.oid));
+		strbuf_add_oid_hex(sb, &commit->object.oid);
 		strbuf_addstr(sb, diff_get_color(c->auto_color, DIFF_RESET));
 		return 1;
 	case 'h':		/* abbreviated commit hash */
@@ -1577,7 +1577,7 @@ static size_t format_commit_one(struct strbuf *sb, /* in UTF-8 */
 		strbuf_addstr(sb, diff_get_color(c->auto_color, DIFF_RESET));
 		return 1;
 	case 'T':		/* tree hash */
-		strbuf_addstr(sb, oid_to_hex(get_commit_tree_oid(commit)));
+		strbuf_add_oid_hex(sb, get_commit_tree_oid(commit));
 		return 1;
 	case 't':		/* abbreviated tree hash */
 		strbuf_add_unique_abbrev(sb,
@@ -1588,7 +1588,7 @@ static size_t format_commit_one(struct strbuf *sb, /* in UTF-8 */
 		for (p = commit->parents; p; p = p->next) {
 			if (p != commit->parents)
 				strbuf_addch(sb, ' ');
-			strbuf_addstr(sb, oid_to_hex(&p->item->object.oid));
+			strbuf_add_oid_hex(sb, &p->item->object.oid);
 		}
 		return 1;
 	case 'p':		/* abbreviated parent hashes */
diff --git a/refs.c b/refs.c
index 844785219d..ee92f18d41 100644
--- a/refs.c
+++ b/refs.c
@@ -2498,7 +2498,7 @@ int refs_update_symref_extended(struct ref_store *refs, const char *ref,
 	if (referent && refs_read_symbolic_ref(refs, ref, referent) == NOT_A_SYMREF) {
 		struct object_id oid;
 		if (!refs_read_ref(refs, ref, &oid)) {
-			strbuf_addstr(referent, oid_to_hex(&oid));
+			strbuf_add_oid_hex(referent, &oid);
 			ret = NOT_A_SYMREF;
 		}
 	}
diff --git a/sequencer.c b/sequencer.c
index b7d8dca47f..b4df04b672 100644
--- a/sequencer.c
+++ b/sequencer.c
@@ -2223,7 +2223,7 @@ static void refer_to_commit(struct repository *r, struct strbuf *msgbuf,
 		repo_format_commit_message(r, commit,
 					   "%h (%s, %ad)", msgbuf, &ctx);
 	} else {
-		strbuf_addstr(msgbuf, oid_to_hex(&commit->object.oid));
+		strbuf_add_oid_hex(msgbuf, &commit->object.oid);
 	}
 }
 
@@ -2395,7 +2395,7 @@ static int do_pick_commit(struct repository *r,
 			if (!has_conforming_footer(&ctx->message, NULL, 0))
 				strbuf_addch(&ctx->message, '\n');
 			strbuf_addstr(&ctx->message, cherry_picked_prefix);
-			strbuf_addstr(&ctx->message, oid_to_hex(&commit->object.oid));
+			strbuf_add_oid_hex(&ctx->message, &commit->object.oid);
 			strbuf_addstr(&ctx->message, ")\n");
 		}
 		if (!is_fixup(command))
diff --git a/shallow.c b/shallow.c
index a8ad92e303..b4b4e2e32a 100644
--- a/shallow.c
+++ b/shallow.c
@@ -395,7 +395,7 @@ static int write_shallow_commits_1(struct strbuf *out, int use_pack_protocol,
 	if (!extra)
 		return data.count;
 	for (size_t i = 0; i < extra->nr; i++) {
-		strbuf_addstr(out, oid_to_hex(extra->oid + i));
+		strbuf_add_oid_hex(out, extra->oid + i);
 		strbuf_addch(out, '\n');
 		data.count++;
 	}
diff --git a/tools/coccinelle/strbuf.cocci b/tools/coccinelle/strbuf.cocci
index f586128329..667903d1d4 100644
--- a/tools/coccinelle/strbuf.cocci
+++ b/tools/coccinelle/strbuf.cocci
@@ -78,3 +78,9 @@ struct strbuf SB;
 @@
 - SB.buf ? SB.buf : ""
 + SB.buf
+
+@@
+expression SB, OID;
+@@
+- strbuf_addstr(SB, oid_to_hex(OID))
++ strbuf_add_oid_hex(SB, OID)
diff --git a/transport-helper.c b/transport-helper.c
index 4614036c99..4a54769789 100644
--- a/transport-helper.c
+++ b/transport-helper.c
@@ -1051,7 +1051,7 @@ static int push_refs_with_push(struct transport *transport,
 			if (ref->peer_ref)
 				strbuf_addstr(&buf, ref->peer_ref->name);
 			else
-				strbuf_addstr(&buf, oid_to_hex(&ref->new_oid));
+				strbuf_add_oid_hex(&buf, &ref->new_oid);
 		}
 		strbuf_addch(&buf, ':');
 		strbuf_addstr(&buf, ref->name);
-- 
2.54.0

Re: [PATCH] hex: add and use strbuf_add_oid_hex()

[Jeff King]

On Wed, May 13, 2026 at 05:49:11PM +0200, René Scharfe wrote:

> Add a function for adding the full hexadecimal hash value of an object
> ID to a strbuf.  It's thread-safe and slightly more efficient than using
> strbuf_addstr() with oid_to_hex() because it doesn't have to determine
> the length of the string or copy it from the intermediate static buffer.
> 
> Add and apply a semantic patch to use it throughout the code base.
> 
> I get a tiny speedup for git log showing a single hash per commit:
> 
> Benchmark 1: ./git_main log --format=%H
>   Time (mean ± σ):      91.2 ms ±   0.7 ms    [User: 51.9 ms, System: 38.6 ms]
>   Range (min … max):    89.8 ms …  92.6 ms    31 runs
> 
> Benchmark 2: ./git log --format=%H
>   Time (mean ± σ):      90.5 ms ±   0.7 ms    [User: 51.0 ms, System: 38.8 ms]
>   Range (min … max):    89.2 ms …  92.3 ms    32 runs

Probably the most extreme benchmark would be:

  git cat-file --batch-all-objects --batch-check='%(objectname)'

which is really just dumping the oids from packfiles. I got ~3% speedup,
though like yours it's within the run-to-run noise.

I think this is worth doing solely for removing more instances of global
buffers, though.

-Peff

Re: [PATCH] hex: add and use strbuf_add_oid_hex()

[René Scharfe]

On 5/13/26 6:01 PM, Jeff King wrote:
> On Wed, May 13, 2026 at 05:49:11PM +0200, René Scharfe wrote:
> 
>> Add a function for adding the full hexadecimal hash value of an object
>> ID to a strbuf.  It's thread-safe and slightly more efficient than using
>> strbuf_addstr() with oid_to_hex() because it doesn't have to determine
>> the length of the string or copy it from the intermediate static buffer.
>>
>> Add and apply a semantic patch to use it throughout the code base.
>>
>> I get a tiny speedup for git log showing a single hash per commit:
>>
>> Benchmark 1: ./git_main log --format=%H
>>   Time (mean ± σ):      91.2 ms ±   0.7 ms    [User: 51.9 ms, System: 38.6 ms]
>>   Range (min … max):    89.8 ms …  92.6 ms    31 runs
>>
>> Benchmark 2: ./git log --format=%H
>>   Time (mean ± σ):      90.5 ms ±   0.7 ms    [User: 51.0 ms, System: 38.8 ms]
>>   Range (min … max):    89.2 ms …  92.3 ms    32 runs
> 
> Probably the most extreme benchmark would be:
> 
>   git cat-file --batch-all-objects --batch-check='%(objectname)'
> 
> which is really just dumping the oids from packfiles. I got ~3% speedup,
> though like yours it's within the run-to-run noise.

Hmm, should've lead with that; on an Apple M1:

Benchmark 1: ./git_main cat-file --batch-all-objects --batch-check='%(objectname)'
  Time (mean ± σ):     117.9 ms ±   0.2 ms    [User: 111.3 ms, System: 5.6 ms]
  Range (min … max):   117.5 ms … 118.5 ms    24 runs

Benchmark 2: ./git cat-file --batch-all-objects --batch-check='%(objectname)'
  Time (mean ± σ):     109.9 ms ±   0.2 ms    [User: 103.2 ms, System: 5.6 ms]
  Range (min … max):   109.5 ms … 110.8 ms    26 runs

Summary
  ./git cat-file --batch-all-objects --batch-check='%(objectname)' ran
    1.07 ± 0.00 times faster than ./git_main cat-file --batch-all-objects --batch-check='%(objectname)'

... and on an M5:

Benchmark 1: ./git_main cat-file --batch-all-objects --batch-check='%(objectname)'
  Time (mean ± σ):      76.8 ms ±   1.9 ms    [User: 73.0 ms, System: 3.2 ms]
  Range (min … max):    73.7 ms …  80.7 ms    38 runs

Benchmark 2: ./git cat-file --batch-all-objects --batch-check='%(objectname)'
  Time (mean ± σ):      70.8 ms ±   1.0 ms    [User: 66.9 ms, System: 3.3 ms]
  Range (min … max):    69.4 ms …  73.7 ms    41 runs

Summary
  ./git cat-file --batch-all-objects --batch-check='%(objectname)' ran
    1.08 ± 0.03 times faster than ./git_main cat-file --batch-all-objects --batch-check='%(objectname)'‚

René