[PATCH v2 09/10] t/unit-tests: convert reftable stack test to use clar

[Seyi Kuforiji <kuforiji98@gmail.com>]

Adapt reftable stack test file to use clar by using clar assertions
where necessary.

This marks the end of all unit tests migrated away from the
`unit-tests/t-*.c` pattern, there are no longer any files matching that
glob. Remove the sanity check for `t-*.c` files to prevent Meson
configuration errors during CI and local builds.

Signed-off-by: Seyi Kuforiji <kuforiji98@gmail.com>
---
 Makefile                        |    2 +-
 t/meson.build                   |    3 +-
 t/unit-tests/t-reftable-stack.c | 1451 -------------------------------
 t/unit-tests/u-reftable-stack.c | 1247 ++++++++++++++++++++++++++
 4 files changed, 1249 insertions(+), 1454 deletions(-)
 delete mode 100644 t/unit-tests/t-reftable-stack.c
 create mode 100644 t/unit-tests/u-reftable-stack.c

diff --git a/Makefile b/Makefile
index cd9db9adf1..0b42893611 100644
--- a/Makefile
+++ b/Makefile
@@ -1369,6 +1369,7 @@ CLAR_TEST_SUITES += u-reftable-pq
 CLAR_TEST_SUITES += u-reftable-reader
 CLAR_TEST_SUITES += u-reftable-readwrite
 CLAR_TEST_SUITES += u-reftable-record
+CLAR_TEST_SUITES += u-reftable-stack
 CLAR_TEST_SUITES += u-reftable-tree
 CLAR_TEST_SUITES += u-strbuf
 CLAR_TEST_SUITES += u-strcmp-offset
@@ -1381,7 +1382,6 @@ CLAR_TEST_OBJS += $(UNIT_TEST_DIR)/clar/clar.o
 CLAR_TEST_OBJS += $(UNIT_TEST_DIR)/unit-test.o
 CLAR_TEST_OBJS += $(UNIT_TEST_DIR)/lib-oid.o
 
-UNIT_TEST_PROGRAMS += t-reftable-stack
 UNIT_TEST_PROGS = $(patsubst %,$(UNIT_TEST_BIN)/%$X,$(UNIT_TEST_PROGRAMS))
 UNIT_TEST_OBJS += $(UNIT_TEST_DIR)/test-lib.o
 UNIT_TEST_OBJS += $(UNIT_TEST_DIR)/lib-reftable.o
diff --git a/t/meson.build b/t/meson.build
index 756cb2a2dd..8fa00fc9ef 100644
--- a/t/meson.build
+++ b/t/meson.build
@@ -15,6 +15,7 @@ clar_test_suites = [
   'unit-tests/u-reftable-reader.c',
   'unit-tests/u-reftable-readwrite.c',
   'unit-tests/u-reftable-record.c',
+  'unit-tests/u-reftable-stack.c',
   'unit-tests/u-reftable-tree.c',
   'unit-tests/u-strbuf.c',
   'unit-tests/u-strcmp-offset.c',
@@ -61,7 +62,6 @@ clar_unit_tests = executable('unit-tests',
 test('unit-tests', clar_unit_tests)
 
 unit_test_programs = [
-  'unit-tests/t-reftable-stack.c',
 ]
 
 foreach unit_test_program : unit_test_programs
@@ -1102,7 +1102,6 @@ integration_tests = [
 # sufficient to catch missing test suites in our CI though.
 foreach glob, tests : {
   't[0-9][0-9][0-9][0-9]-*.sh': integration_tests,
-  'unit-tests/t-*.c': unit_test_programs,
   'unit-tests/u-*.c': clar_test_suites,
 }
   actual_tests = run_command(shell, '-c', 'ls ' + glob,
diff --git a/t/unit-tests/t-reftable-stack.c b/t/unit-tests/t-reftable-stack.c
deleted file mode 100644
index c3f0059c34..0000000000
--- a/t/unit-tests/t-reftable-stack.c
+++ /dev/null
@@ -1,1451 +0,0 @@
-/*
-Copyright 2020 Google LLC
-
-Use of this source code is governed by a BSD-style
-license that can be found in the LICENSE file or at
-https://developers.google.com/open-source/licenses/bsd
-*/
-
-#define DISABLE_SIGN_COMPARE_WARNINGS
-
-#include "test-lib.h"
-#include "lib-reftable.h"
-#include "dir.h"
-#include "reftable/merged.h"
-#include "reftable/reader.h"
-#include "reftable/reftable-error.h"
-#include "reftable/stack.h"
-#include "strbuf.h"
-#include "tempfile.h"
-#include <dirent.h>
-
-static void clear_dir(const char *dirname)
-{
-	struct strbuf path = REFTABLE_BUF_INIT;
-	strbuf_addstr(&path, dirname);
-	remove_dir_recursively(&path, 0);
-	strbuf_release(&path);
-}
-
-static int count_dir_entries(const char *dirname)
-{
-	DIR *dir = opendir(dirname);
-	int len = 0;
-	struct dirent *d;
-	if (!dir)
-		return 0;
-
-	while ((d = readdir(dir))) {
-		/*
-		 * Besides skipping over "." and "..", we also need to
-		 * skip over other files that have a leading ".". This
-		 * is due to behaviour of NFS, which will rename files
-		 * to ".nfs*" to emulate delete-on-last-close.
-		 *
-		 * In any case this should be fine as the reftable
-		 * library will never write files with leading dots
-		 * anyway.
-		 */
-		if (starts_with(d->d_name, "."))
-			continue;
-		len++;
-	}
-	closedir(dir);
-	return len;
-}
-
-/*
- * Work linenumber into the tempdir, so we can see which tests forget to
- * cleanup.
- */
-static char *get_tmp_template(int linenumber)
-{
-	const char *tmp = getenv("TMPDIR");
-	static char template[1024];
-	snprintf(template, sizeof(template) - 1, "%s/stack_test-%d.XXXXXX",
-		 tmp ? tmp : "/tmp", linenumber);
-	return template;
-}
-
-static char *get_tmp_dir(int linenumber)
-{
-	char *dir = get_tmp_template(linenumber);
-	check(mkdtemp(dir) != NULL);
-	return dir;
-}
-
-static void t_read_file(void)
-{
-	char *fn = get_tmp_template(__LINE__);
-	struct tempfile *tmp = mks_tempfile(fn);
-	int fd = get_tempfile_fd(tmp);
-	char out[1024] = "line1\n\nline2\nline3";
-	int n, err;
-	char **names = NULL;
-	const char *want[] = { "line1", "line2", "line3" };
-
-	check_int(fd, >, 0);
-	n = write_in_full(fd, out, strlen(out));
-	check_int(n, ==, strlen(out));
-	err = close(fd);
-	check_int(err, >=, 0);
-
-	err = read_lines(fn, &names);
-	check(!err);
-
-	for (size_t i = 0; names[i]; i++)
-		check_str(want[i], names[i]);
-	free_names(names);
-	(void) remove(fn);
-	delete_tempfile(&tmp);
-}
-
-static int write_test_ref(struct reftable_writer *wr, void *arg)
-{
-	struct reftable_ref_record *ref = arg;
-	check(!reftable_writer_set_limits(wr, ref->update_index,
-					  ref->update_index));
-	return reftable_writer_add_ref(wr, ref);
-}
-
-static void write_n_ref_tables(struct reftable_stack *st,
-			       size_t n)
-{
-	int disable_auto_compact;
-	int err;
-
-	disable_auto_compact = st->opts.disable_auto_compact;
-	st->opts.disable_auto_compact = 1;
-
-	for (size_t i = 0; i < n; i++) {
-		struct reftable_ref_record ref = {
-			.update_index = reftable_stack_next_update_index(st),
-			.value_type = REFTABLE_REF_VAL1,
-		};
-		char buf[128];
-
-		snprintf(buf, sizeof(buf), "refs/heads/branch-%04"PRIuMAX, (uintmax_t)i);
-		ref.refname = buf;
-		t_reftable_set_hash(ref.value.val1, i, REFTABLE_HASH_SHA1);
-
-		err = reftable_stack_add(st, &write_test_ref, &ref);
-		check(!err);
-	}
-
-	st->opts.disable_auto_compact = disable_auto_compact;
-}
-
-struct write_log_arg {
-	struct reftable_log_record *log;
-	uint64_t update_index;
-};
-
-static int write_test_log(struct reftable_writer *wr, void *arg)
-{
-	struct write_log_arg *wla = arg;
-
-	check(!reftable_writer_set_limits(wr, wla->update_index,
-					  wla->update_index));
-	return reftable_writer_add_log(wr, wla->log);
-}
-
-static void t_reftable_stack_add_one(void)
-{
-	char *dir = get_tmp_dir(__LINE__);
-	struct reftable_buf scratch = REFTABLE_BUF_INIT;
-	int mask = umask(002);
-	struct reftable_write_options opts = {
-		.default_permissions = 0660,
-	};
-	struct reftable_stack *st = NULL;
-	int err;
-	struct reftable_ref_record ref = {
-		.refname = (char *) "HEAD",
-		.update_index = 1,
-		.value_type = REFTABLE_REF_SYMREF,
-		.value.symref = (char *) "master",
-	};
-	struct reftable_ref_record dest = { 0 };
-	struct stat stat_result = { 0 };
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	err = reftable_stack_add(st, write_test_ref, &ref);
-	check(!err);
-
-	err = reftable_stack_read_ref(st, ref.refname, &dest);
-	check(!err);
-	check(reftable_ref_record_equal(&ref, &dest, REFTABLE_HASH_SIZE_SHA1));
-	check_int(st->readers_len, >, 0);
-
-#ifndef GIT_WINDOWS_NATIVE
-	check(!reftable_buf_addstr(&scratch, dir));
-	check(!reftable_buf_addstr(&scratch, "/tables.list"));
-	err = stat(scratch.buf, &stat_result);
-	check(!err);
-	check_int((stat_result.st_mode & 0777), ==, opts.default_permissions);
-
-	reftable_buf_reset(&scratch);
-	check(!reftable_buf_addstr(&scratch, dir));
-	check(!reftable_buf_addstr(&scratch, "/"));
-	/* do not try at home; not an external API for reftable. */
-	check(!reftable_buf_addstr(&scratch, st->readers[0]->name));
-	err = stat(scratch.buf, &stat_result);
-	check(!err);
-	check_int((stat_result.st_mode & 0777), ==, opts.default_permissions);
-#else
-	(void) stat_result;
-#endif
-
-	reftable_ref_record_release(&dest);
-	reftable_stack_destroy(st);
-	reftable_buf_release(&scratch);
-	clear_dir(dir);
-	umask(mask);
-}
-
-static void t_reftable_stack_uptodate(void)
-{
-	struct reftable_write_options opts = { 0 };
-	struct reftable_stack *st1 = NULL;
-	struct reftable_stack *st2 = NULL;
-	char *dir = get_tmp_dir(__LINE__);
-
-	int err;
-	struct reftable_ref_record ref1 = {
-		.refname = (char *) "HEAD",
-		.update_index = 1,
-		.value_type = REFTABLE_REF_SYMREF,
-		.value.symref = (char *) "master",
-	};
-	struct reftable_ref_record ref2 = {
-		.refname = (char *) "branch2",
-		.update_index = 2,
-		.value_type = REFTABLE_REF_SYMREF,
-		.value.symref = (char *) "master",
-	};
-
-
-	/* simulate multi-process access to the same stack
-	   by creating two stacks for the same directory.
-	 */
-	err = reftable_new_stack(&st1, dir, &opts);
-	check(!err);
-
-	err = reftable_new_stack(&st2, dir, &opts);
-	check(!err);
-
-	err = reftable_stack_add(st1, write_test_ref, &ref1);
-	check(!err);
-
-	err = reftable_stack_add(st2, write_test_ref, &ref2);
-	check_int(err, ==, REFTABLE_OUTDATED_ERROR);
-
-	err = reftable_stack_reload(st2);
-	check(!err);
-
-	err = reftable_stack_add(st2, write_test_ref, &ref2);
-	check(!err);
-	reftable_stack_destroy(st1);
-	reftable_stack_destroy(st2);
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_transaction_api(void)
-{
-	char *dir = get_tmp_dir(__LINE__);
-	struct reftable_write_options opts = { 0 };
-	struct reftable_stack *st = NULL;
-	int err;
-	struct reftable_addition *add = NULL;
-
-	struct reftable_ref_record ref = {
-		.refname = (char *) "HEAD",
-		.update_index = 1,
-		.value_type = REFTABLE_REF_SYMREF,
-		.value.symref = (char *) "master",
-	};
-	struct reftable_ref_record dest = { 0 };
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	reftable_addition_destroy(add);
-
-	err = reftable_stack_new_addition(&add, st, 0);
-	check(!err);
-
-	err = reftable_addition_add(add, write_test_ref, &ref);
-	check(!err);
-
-	err = reftable_addition_commit(add);
-	check(!err);
-
-	reftable_addition_destroy(add);
-
-	err = reftable_stack_read_ref(st, ref.refname, &dest);
-	check(!err);
-	check_int(REFTABLE_REF_SYMREF, ==, dest.value_type);
-	check(reftable_ref_record_equal(&ref, &dest, REFTABLE_HASH_SIZE_SHA1));
-
-	reftable_ref_record_release(&dest);
-	reftable_stack_destroy(st);
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_transaction_with_reload(void)
-{
-	char *dir = get_tmp_dir(__LINE__);
-	struct reftable_stack *st1 = NULL, *st2 = NULL;
-	int err;
-	struct reftable_addition *add = NULL;
-	struct reftable_ref_record refs[2] = {
-		{
-			.refname = (char *) "refs/heads/a",
-			.update_index = 1,
-			.value_type = REFTABLE_REF_VAL1,
-			.value.val1 = { '1' },
-		},
-		{
-			.refname = (char *) "refs/heads/b",
-			.update_index = 2,
-			.value_type = REFTABLE_REF_VAL1,
-			.value.val1 = { '1' },
-		},
-	};
-	struct reftable_ref_record ref = { 0 };
-
-	err = reftable_new_stack(&st1, dir, NULL);
-	check(!err);
-	err = reftable_new_stack(&st2, dir, NULL);
-	check(!err);
-
-	err = reftable_stack_new_addition(&add, st1, 0);
-	check(!err);
-	err = reftable_addition_add(add, write_test_ref, &refs[0]);
-	check(!err);
-	err = reftable_addition_commit(add);
-	check(!err);
-	reftable_addition_destroy(add);
-
-	/*
-	 * The second stack is now outdated, which we should notice. We do not
-	 * create the addition and lock the stack by default, but allow the
-	 * reload to happen when REFTABLE_STACK_NEW_ADDITION_RELOAD is set.
-	 */
-	err = reftable_stack_new_addition(&add, st2, 0);
-	check_int(err, ==, REFTABLE_OUTDATED_ERROR);
-	err = reftable_stack_new_addition(&add, st2, REFTABLE_STACK_NEW_ADDITION_RELOAD);
-	check(!err);
-	err = reftable_addition_add(add, write_test_ref, &refs[1]);
-	check(!err);
-	err = reftable_addition_commit(add);
-	check(!err);
-	reftable_addition_destroy(add);
-
-	for (size_t i = 0; i < ARRAY_SIZE(refs); i++) {
-		err = reftable_stack_read_ref(st2, refs[i].refname, &ref);
-		check(!err);
-		check(reftable_ref_record_equal(&refs[i], &ref, REFTABLE_HASH_SIZE_SHA1));
-	}
-
-	reftable_ref_record_release(&ref);
-	reftable_stack_destroy(st1);
-	reftable_stack_destroy(st2);
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_transaction_api_performs_auto_compaction(void)
-{
-	char *dir = get_tmp_dir(__LINE__);
-	struct reftable_write_options opts = {0};
-	struct reftable_addition *add = NULL;
-	struct reftable_stack *st = NULL;
-	size_t n = 20;
-	int err;
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	for (size_t i = 0; i <= n; i++) {
-		struct reftable_ref_record ref = {
-			.update_index = reftable_stack_next_update_index(st),
-			.value_type = REFTABLE_REF_SYMREF,
-			.value.symref = (char *) "master",
-		};
-		char name[100];
-
-		snprintf(name, sizeof(name), "branch%04"PRIuMAX, (uintmax_t)i);
-		ref.refname = name;
-
-		/*
-		 * Disable auto-compaction for all but the last runs. Like this
-		 * we can ensure that we indeed honor this setting and have
-		 * better control over when exactly auto compaction runs.
-		 */
-		st->opts.disable_auto_compact = i != n;
-
-		err = reftable_stack_new_addition(&add, st, 0);
-		check(!err);
-
-		err = reftable_addition_add(add, write_test_ref, &ref);
-		check(!err);
-
-		err = reftable_addition_commit(add);
-		check(!err);
-
-		reftable_addition_destroy(add);
-
-		/*
-		 * The stack length should grow continuously for all runs where
-		 * auto compaction is disabled. When enabled, we should merge
-		 * all tables in the stack.
-		 */
-		if (i != n)
-			check_int(st->merged->readers_len, ==, i + 1);
-		else
-			check_int(st->merged->readers_len, ==, 1);
-	}
-
-	reftable_stack_destroy(st);
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_auto_compaction_fails_gracefully(void)
-{
-	struct reftable_ref_record ref = {
-		.refname = (char *) "refs/heads/master",
-		.update_index = 1,
-		.value_type = REFTABLE_REF_VAL1,
-		.value.val1 = {0x01},
-	};
-	struct reftable_write_options opts = { 0 };
-	struct reftable_stack *st;
-	struct reftable_buf table_path = REFTABLE_BUF_INIT;
-	char *dir = get_tmp_dir(__LINE__);
-	int err;
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	err = reftable_stack_add(st, write_test_ref, &ref);
-	check(!err);
-	check_int(st->merged->readers_len, ==, 1);
-	check_int(st->stats.attempts, ==, 0);
-	check_int(st->stats.failures, ==, 0);
-
-	/*
-	 * Lock the newly written table such that it cannot be compacted.
-	 * Adding a new table to the stack should not be impacted by this, even
-	 * though auto-compaction will now fail.
-	 */
-	check(!reftable_buf_addstr(&table_path, dir));
-	check(!reftable_buf_addstr(&table_path, "/"));
-	check(!reftable_buf_addstr(&table_path, st->readers[0]->name));
-	check(!reftable_buf_addstr(&table_path, ".lock"));
-	write_file_buf(table_path.buf, "", 0);
-
-	ref.update_index = 2;
-	err = reftable_stack_add(st, write_test_ref, &ref);
-	check(!err);
-	check_int(st->merged->readers_len, ==, 2);
-	check_int(st->stats.attempts, ==, 1);
-	check_int(st->stats.failures, ==, 1);
-
-	reftable_stack_destroy(st);
-	reftable_buf_release(&table_path);
-	clear_dir(dir);
-}
-
-static int write_error(struct reftable_writer *wr UNUSED, void *arg)
-{
-	return *((int *)arg);
-}
-
-static void t_reftable_stack_update_index_check(void)
-{
-	char *dir = get_tmp_dir(__LINE__);
-	struct reftable_write_options opts = { 0 };
-	struct reftable_stack *st = NULL;
-	int err;
-	struct reftable_ref_record ref1 = {
-		.refname = (char *) "name1",
-		.update_index = 1,
-		.value_type = REFTABLE_REF_SYMREF,
-		.value.symref = (char *) "master",
-	};
-	struct reftable_ref_record ref2 = {
-		.refname = (char *) "name2",
-		.update_index = 1,
-		.value_type = REFTABLE_REF_SYMREF,
-		.value.symref = (char *) "master",
-	};
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	err = reftable_stack_add(st, write_test_ref, &ref1);
-	check(!err);
-
-	err = reftable_stack_add(st, write_test_ref, &ref2);
-	check_int(err, ==, REFTABLE_API_ERROR);
-	reftable_stack_destroy(st);
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_lock_failure(void)
-{
-	char *dir = get_tmp_dir(__LINE__);
-	struct reftable_write_options opts = { 0 };
-	struct reftable_stack *st = NULL;
-	int err, i;
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-	for (i = -1; i != REFTABLE_EMPTY_TABLE_ERROR; i--) {
-		err = reftable_stack_add(st, write_error, &i);
-		check_int(err, ==, i);
-	}
-
-	reftable_stack_destroy(st);
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_add(void)
-{
-	int err = 0;
-	struct reftable_write_options opts = {
-		.exact_log_message = 1,
-		.default_permissions = 0660,
-		.disable_auto_compact = 1,
-	};
-	struct reftable_stack *st = NULL;
-	char *dir = get_tmp_dir(__LINE__);
-	struct reftable_ref_record refs[2] = { 0 };
-	struct reftable_log_record logs[2] = { 0 };
-	struct reftable_buf path = REFTABLE_BUF_INIT;
-	struct stat stat_result;
-	size_t i, N = ARRAY_SIZE(refs);
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	for (i = 0; i < N; i++) {
-		char buf[256];
-		snprintf(buf, sizeof(buf), "branch%02"PRIuMAX, (uintmax_t)i);
-		refs[i].refname = xstrdup(buf);
-		refs[i].update_index = i + 1;
-		refs[i].value_type = REFTABLE_REF_VAL1;
-		t_reftable_set_hash(refs[i].value.val1, i, REFTABLE_HASH_SHA1);
-
-		logs[i].refname = xstrdup(buf);
-		logs[i].update_index = N + i + 1;
-		logs[i].value_type = REFTABLE_LOG_UPDATE;
-		logs[i].value.update.email = xstrdup("identity@invalid");
-		t_reftable_set_hash(logs[i].value.update.new_hash, i, REFTABLE_HASH_SHA1);
-	}
-
-	for (i = 0; i < N; i++) {
-		int err = reftable_stack_add(st, write_test_ref, &refs[i]);
-		check(!err);
-	}
-
-	for (i = 0; i < N; i++) {
-		struct write_log_arg arg = {
-			.log = &logs[i],
-			.update_index = reftable_stack_next_update_index(st),
-		};
-		int err = reftable_stack_add(st, write_test_log, &arg);
-		check(!err);
-	}
-
-	err = reftable_stack_compact_all(st, NULL);
-	check(!err);
-
-	for (i = 0; i < N; i++) {
-		struct reftable_ref_record dest = { 0 };
-
-		int err = reftable_stack_read_ref(st, refs[i].refname, &dest);
-		check(!err);
-		check(reftable_ref_record_equal(&dest, refs + i,
-						 REFTABLE_HASH_SIZE_SHA1));
-		reftable_ref_record_release(&dest);
-	}
-
-	for (i = 0; i < N; i++) {
-		struct reftable_log_record dest = { 0 };
-		int err = reftable_stack_read_log(st, refs[i].refname, &dest);
-		check(!err);
-		check(reftable_log_record_equal(&dest, logs + i,
-						 REFTABLE_HASH_SIZE_SHA1));
-		reftable_log_record_release(&dest);
-	}
-
-#ifndef GIT_WINDOWS_NATIVE
-	check(!reftable_buf_addstr(&path, dir));
-	check(!reftable_buf_addstr(&path, "/tables.list"));
-	err = stat(path.buf, &stat_result);
-	check(!err);
-	check_int((stat_result.st_mode & 0777), ==, opts.default_permissions);
-
-	reftable_buf_reset(&path);
-	check(!reftable_buf_addstr(&path, dir));
-	check(!reftable_buf_addstr(&path, "/"));
-	/* do not try at home; not an external API for reftable. */
-	check(!reftable_buf_addstr(&path, st->readers[0]->name));
-	err = stat(path.buf, &stat_result);
-	check(!err);
-	check_int((stat_result.st_mode & 0777), ==, opts.default_permissions);
-#else
-	(void) stat_result;
-#endif
-
-	/* cleanup */
-	reftable_stack_destroy(st);
-	for (i = 0; i < N; i++) {
-		reftable_ref_record_release(&refs[i]);
-		reftable_log_record_release(&logs[i]);
-	}
-	reftable_buf_release(&path);
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_iterator(void)
-{
-	struct reftable_write_options opts = { 0 };
-	struct reftable_stack *st = NULL;
-	char *dir = get_tmp_dir(__LINE__);
-	struct reftable_ref_record refs[10] = { 0 };
-	struct reftable_log_record logs[10] = { 0 };
-	struct reftable_iterator it = { 0 };
-	size_t N = ARRAY_SIZE(refs), i;
-	int err;
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	for (i = 0; i < N; i++) {
-		refs[i].refname = xstrfmt("branch%02"PRIuMAX, (uintmax_t)i);
-		refs[i].update_index = i + 1;
-		refs[i].value_type = REFTABLE_REF_VAL1;
-		t_reftable_set_hash(refs[i].value.val1, i, REFTABLE_HASH_SHA1);
-
-		logs[i].refname = xstrfmt("branch%02"PRIuMAX, (uintmax_t)i);
-		logs[i].update_index = i + 1;
-		logs[i].value_type = REFTABLE_LOG_UPDATE;
-		logs[i].value.update.email = xstrdup("johndoe@invalid");
-		logs[i].value.update.message = xstrdup("commit\n");
-		t_reftable_set_hash(logs[i].value.update.new_hash, i, REFTABLE_HASH_SHA1);
-	}
-
-	for (i = 0; i < N; i++) {
-		err = reftable_stack_add(st, write_test_ref, &refs[i]);
-		check(!err);
-	}
-
-	for (i = 0; i < N; i++) {
-		struct write_log_arg arg = {
-			.log = &logs[i],
-			.update_index = reftable_stack_next_update_index(st),
-		};
-
-		err = reftable_stack_add(st, write_test_log, &arg);
-		check(!err);
-	}
-
-	reftable_stack_init_ref_iterator(st, &it);
-	reftable_iterator_seek_ref(&it, refs[0].refname);
-	for (i = 0; ; i++) {
-		struct reftable_ref_record ref = { 0 };
-
-		err = reftable_iterator_next_ref(&it, &ref);
-		if (err > 0)
-			break;
-		check(!err);
-		check(reftable_ref_record_equal(&ref, &refs[i], REFTABLE_HASH_SIZE_SHA1));
-		reftable_ref_record_release(&ref);
-	}
-	check_int(i, ==, N);
-
-	reftable_iterator_destroy(&it);
-
-	err = reftable_stack_init_log_iterator(st, &it);
-	check(!err);
-
-	reftable_iterator_seek_log(&it, logs[0].refname);
-	for (i = 0; ; i++) {
-		struct reftable_log_record log = { 0 };
-
-		err = reftable_iterator_next_log(&it, &log);
-		if (err > 0)
-			break;
-		check(!err);
-		check(reftable_log_record_equal(&log, &logs[i], REFTABLE_HASH_SIZE_SHA1));
-		reftable_log_record_release(&log);
-	}
-	check_int(i, ==, N);
-
-	reftable_stack_destroy(st);
-	reftable_iterator_destroy(&it);
-	for (i = 0; i < N; i++) {
-		reftable_ref_record_release(&refs[i]);
-		reftable_log_record_release(&logs[i]);
-	}
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_log_normalize(void)
-{
-	int err = 0;
-	struct reftable_write_options opts = {
-		0,
-	};
-	struct reftable_stack *st = NULL;
-	char *dir = get_tmp_dir(__LINE__);
-	struct reftable_log_record input = {
-		.refname = (char *) "branch",
-		.update_index = 1,
-		.value_type = REFTABLE_LOG_UPDATE,
-		.value = {
-			.update = {
-				.new_hash = { 1 },
-				.old_hash = { 2 },
-			},
-		},
-	};
-	struct reftable_log_record dest = {
-		.update_index = 0,
-	};
-	struct write_log_arg arg = {
-		.log = &input,
-		.update_index = 1,
-	};
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	input.value.update.message = (char *) "one\ntwo";
-	err = reftable_stack_add(st, write_test_log, &arg);
-	check_int(err, ==, REFTABLE_API_ERROR);
-
-	input.value.update.message = (char *) "one";
-	err = reftable_stack_add(st, write_test_log, &arg);
-	check(!err);
-
-	err = reftable_stack_read_log(st, input.refname, &dest);
-	check(!err);
-	check_str(dest.value.update.message, "one\n");
-
-	input.value.update.message = (char *) "two\n";
-	arg.update_index = 2;
-	err = reftable_stack_add(st, write_test_log, &arg);
-	check(!err);
-	err = reftable_stack_read_log(st, input.refname, &dest);
-	check(!err);
-	check_str(dest.value.update.message, "two\n");
-
-	/* cleanup */
-	reftable_stack_destroy(st);
-	reftable_log_record_release(&dest);
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_tombstone(void)
-{
-	char *dir = get_tmp_dir(__LINE__);
-	struct reftable_write_options opts = { 0 };
-	struct reftable_stack *st = NULL;
-	int err;
-	struct reftable_ref_record refs[2] = { 0 };
-	struct reftable_log_record logs[2] = { 0 };
-	size_t i, N = ARRAY_SIZE(refs);
-	struct reftable_ref_record dest = { 0 };
-	struct reftable_log_record log_dest = { 0 };
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	/* even entries add the refs, odd entries delete them. */
-	for (i = 0; i < N; i++) {
-		const char *buf = "branch";
-		refs[i].refname = xstrdup(buf);
-		refs[i].update_index = i + 1;
-		if (i % 2 == 0) {
-			refs[i].value_type = REFTABLE_REF_VAL1;
-			t_reftable_set_hash(refs[i].value.val1, i,
-					    REFTABLE_HASH_SHA1);
-		}
-
-		logs[i].refname = xstrdup(buf);
-		/*
-		 * update_index is part of the key so should be constant.
-		 * The value itself should be less than the writer's upper
-		 * limit.
-		 */
-		logs[i].update_index = 1;
-		if (i % 2 == 0) {
-			logs[i].value_type = REFTABLE_LOG_UPDATE;
-			t_reftable_set_hash(logs[i].value.update.new_hash, i,
-					    REFTABLE_HASH_SHA1);
-			logs[i].value.update.email =
-				xstrdup("identity@invalid");
-		}
-	}
-	for (i = 0; i < N; i++) {
-		int err = reftable_stack_add(st, write_test_ref, &refs[i]);
-		check(!err);
-	}
-
-	for (i = 0; i < N; i++) {
-		struct write_log_arg arg = {
-			.log = &logs[i],
-			.update_index = reftable_stack_next_update_index(st),
-		};
-		int err = reftable_stack_add(st, write_test_log, &arg);
-		check(!err);
-	}
-
-	err = reftable_stack_read_ref(st, "branch", &dest);
-	check_int(err, ==, 1);
-	reftable_ref_record_release(&dest);
-
-	err = reftable_stack_read_log(st, "branch", &log_dest);
-	check_int(err, ==, 1);
-	reftable_log_record_release(&log_dest);
-
-	err = reftable_stack_compact_all(st, NULL);
-	check(!err);
-
-	err = reftable_stack_read_ref(st, "branch", &dest);
-	check_int(err, ==, 1);
-
-	err = reftable_stack_read_log(st, "branch", &log_dest);
-	check_int(err, ==, 1);
-	reftable_ref_record_release(&dest);
-	reftable_log_record_release(&log_dest);
-
-	/* cleanup */
-	reftable_stack_destroy(st);
-	for (i = 0; i < N; i++) {
-		reftable_ref_record_release(&refs[i]);
-		reftable_log_record_release(&logs[i]);
-	}
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_hash_id(void)
-{
-	char *dir = get_tmp_dir(__LINE__);
-	struct reftable_write_options opts = { 0 };
-	struct reftable_stack *st = NULL;
-	int err;
-
-	struct reftable_ref_record ref = {
-		.refname = (char *) "master",
-		.value_type = REFTABLE_REF_SYMREF,
-		.value.symref = (char *) "target",
-		.update_index = 1,
-	};
-	struct reftable_write_options opts32 = { .hash_id = REFTABLE_HASH_SHA256 };
-	struct reftable_stack *st32 = NULL;
-	struct reftable_write_options opts_default = { 0 };
-	struct reftable_stack *st_default = NULL;
-	struct reftable_ref_record dest = { 0 };
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	err = reftable_stack_add(st, write_test_ref, &ref);
-	check(!err);
-
-	/* can't read it with the wrong hash ID. */
-	err = reftable_new_stack(&st32, dir, &opts32);
-	check_int(err, ==, REFTABLE_FORMAT_ERROR);
-
-	/* check that we can read it back with default opts too. */
-	err = reftable_new_stack(&st_default, dir, &opts_default);
-	check(!err);
-
-	err = reftable_stack_read_ref(st_default, "master", &dest);
-	check(!err);
-
-	check(reftable_ref_record_equal(&ref, &dest, REFTABLE_HASH_SIZE_SHA1));
-	reftable_ref_record_release(&dest);
-	reftable_stack_destroy(st);
-	reftable_stack_destroy(st_default);
-	clear_dir(dir);
-}
-
-static void t_suggest_compaction_segment(void)
-{
-	uint64_t sizes[] = { 512, 64, 17, 16, 9, 9, 9, 16, 2, 16 };
-	struct segment min =
-		suggest_compaction_segment(sizes, ARRAY_SIZE(sizes), 2);
-	check_int(min.start, ==, 1);
-	check_int(min.end, ==, 10);
-}
-
-static void t_suggest_compaction_segment_nothing(void)
-{
-	uint64_t sizes[] = { 64, 32, 16, 8, 4, 2 };
-	struct segment result =
-		suggest_compaction_segment(sizes, ARRAY_SIZE(sizes), 2);
-	check_int(result.start, ==, result.end);
-}
-
-static void t_reflog_expire(void)
-{
-	char *dir = get_tmp_dir(__LINE__);
-	struct reftable_write_options opts = { 0 };
-	struct reftable_stack *st = NULL;
-	struct reftable_log_record logs[20] = { 0 };
-	size_t i, N = ARRAY_SIZE(logs) - 1;
-	int err;
-	struct reftable_log_expiry_config expiry = {
-		.time = 10,
-	};
-	struct reftable_log_record log = { 0 };
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	for (i = 1; i <= N; i++) {
-		char buf[256];
-		snprintf(buf, sizeof(buf), "branch%02"PRIuMAX, (uintmax_t)i);
-
-		logs[i].refname = xstrdup(buf);
-		logs[i].update_index = i;
-		logs[i].value_type = REFTABLE_LOG_UPDATE;
-		logs[i].value.update.time = i;
-		logs[i].value.update.email = xstrdup("identity@invalid");
-		t_reftable_set_hash(logs[i].value.update.new_hash, i,
-				    REFTABLE_HASH_SHA1);
-	}
-
-	for (i = 1; i <= N; i++) {
-		struct write_log_arg arg = {
-			.log = &logs[i],
-			.update_index = reftable_stack_next_update_index(st),
-		};
-		int err = reftable_stack_add(st, write_test_log, &arg);
-		check(!err);
-	}
-
-	err = reftable_stack_compact_all(st, NULL);
-	check(!err);
-
-	err = reftable_stack_compact_all(st, &expiry);
-	check(!err);
-
-	err = reftable_stack_read_log(st, logs[9].refname, &log);
-	check_int(err, ==, 1);
-
-	err = reftable_stack_read_log(st, logs[11].refname, &log);
-	check(!err);
-
-	expiry.min_update_index = 15;
-	err = reftable_stack_compact_all(st, &expiry);
-	check(!err);
-
-	err = reftable_stack_read_log(st, logs[14].refname, &log);
-	check_int(err, ==, 1);
-
-	err = reftable_stack_read_log(st, logs[16].refname, &log);
-	check(!err);
-
-	/* cleanup */
-	reftable_stack_destroy(st);
-	for (i = 0; i <= N; i++)
-		reftable_log_record_release(&logs[i]);
-	clear_dir(dir);
-	reftable_log_record_release(&log);
-}
-
-static int write_nothing(struct reftable_writer *wr, void *arg UNUSED)
-{
-	check(!reftable_writer_set_limits(wr, 1, 1));
-	return 0;
-}
-
-static void t_empty_add(void)
-{
-	struct reftable_write_options opts = { 0 };
-	struct reftable_stack *st = NULL;
-	int err;
-	char *dir = get_tmp_dir(__LINE__);
-	struct reftable_stack *st2 = NULL;
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	err = reftable_stack_add(st, write_nothing, NULL);
-	check(!err);
-
-	err = reftable_new_stack(&st2, dir, &opts);
-	check(!err);
-	clear_dir(dir);
-	reftable_stack_destroy(st);
-	reftable_stack_destroy(st2);
-}
-
-static int fastlogN(uint64_t sz, uint64_t N)
-{
-	int l = 0;
-	if (sz == 0)
-		return 0;
-	for (; sz; sz /= N)
-		l++;
-	return l - 1;
-}
-
-static void t_reftable_stack_auto_compaction(void)
-{
-	struct reftable_write_options opts = {
-		.disable_auto_compact = 1,
-	};
-	struct reftable_stack *st = NULL;
-	char *dir = get_tmp_dir(__LINE__);
-	int err;
-	size_t i, N = 100;
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	for (i = 0; i < N; i++) {
-		char name[100];
-		struct reftable_ref_record ref = {
-			.refname = name,
-			.update_index = reftable_stack_next_update_index(st),
-			.value_type = REFTABLE_REF_SYMREF,
-			.value.symref = (char *) "master",
-		};
-		snprintf(name, sizeof(name), "branch%04"PRIuMAX, (uintmax_t)i);
-
-		err = reftable_stack_add(st, write_test_ref, &ref);
-		check(!err);
-
-		err = reftable_stack_auto_compact(st);
-		check(!err);
-		check(i < 2 || st->merged->readers_len < 2 * fastlogN(i, 2));
-	}
-
-	check_int(reftable_stack_compaction_stats(st)->entries_written, <,
-	       (uint64_t)(N * fastlogN(N, 2)));
-
-	reftable_stack_destroy(st);
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_auto_compaction_factor(void)
-{
-	struct reftable_write_options opts = {
-		.auto_compaction_factor = 5,
-	};
-	struct reftable_stack *st = NULL;
-	char *dir = get_tmp_dir(__LINE__);
-	int err;
-	size_t N = 100;
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	for (size_t i = 0; i < N; i++) {
-		char name[20];
-		struct reftable_ref_record ref = {
-			.refname = name,
-			.update_index = reftable_stack_next_update_index(st),
-			.value_type = REFTABLE_REF_VAL1,
-		};
-		xsnprintf(name, sizeof(name), "branch%04"PRIuMAX, (uintmax_t)i);
-
-		err = reftable_stack_add(st, &write_test_ref, &ref);
-		check(!err);
-
-		check(i < 5 || st->merged->readers_len < 5 * fastlogN(i, 5));
-	}
-
-	reftable_stack_destroy(st);
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_auto_compaction_with_locked_tables(void)
-{
-	struct reftable_write_options opts = {
-		.disable_auto_compact = 1,
-	};
-	struct reftable_stack *st = NULL;
-	struct reftable_buf buf = REFTABLE_BUF_INIT;
-	char *dir = get_tmp_dir(__LINE__);
-	int err;
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	write_n_ref_tables(st, 5);
-	check_int(st->merged->readers_len, ==, 5);
-
-	/*
-	 * Given that all tables we have written should be roughly the same
-	 * size, we expect that auto-compaction will want to compact all of the
-	 * tables. Locking any of the tables will keep it from doing so.
-	 */
-	check(!reftable_buf_addstr(&buf, dir));
-	check(!reftable_buf_addstr(&buf, "/"));
-	check(!reftable_buf_addstr(&buf, st->readers[2]->name));
-	check(!reftable_buf_addstr(&buf, ".lock"));
-	write_file_buf(buf.buf, "", 0);
-
-	/*
-	 * When parts of the stack are locked, then auto-compaction does a best
-	 * effort compaction of those tables which aren't locked. So while this
-	 * would in theory compact all tables, due to the preexisting lock we
-	 * only compact the newest two tables.
-	 */
-	err = reftable_stack_auto_compact(st);
-	check(!err);
-	check_int(st->stats.failures, ==, 0);
-	check_int(st->merged->readers_len, ==, 4);
-
-	reftable_stack_destroy(st);
-	reftable_buf_release(&buf);
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_add_performs_auto_compaction(void)
-{
-	struct reftable_write_options opts = { 0 };
-	struct reftable_stack *st = NULL;
-	char *dir = get_tmp_dir(__LINE__);
-	int err;
-	size_t i, n = 20;
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	for (i = 0; i <= n; i++) {
-		struct reftable_ref_record ref = {
-			.update_index = reftable_stack_next_update_index(st),
-			.value_type = REFTABLE_REF_SYMREF,
-			.value.symref = (char *) "master",
-		};
-		char buf[128];
-
-		/*
-		 * Disable auto-compaction for all but the last runs. Like this
-		 * we can ensure that we indeed honor this setting and have
-		 * better control over when exactly auto compaction runs.
-		 */
-		st->opts.disable_auto_compact = i != n;
-
-		snprintf(buf, sizeof(buf), "branch-%04"PRIuMAX, (uintmax_t)i);
-		ref.refname = buf;
-
-		err = reftable_stack_add(st, write_test_ref, &ref);
-		check(!err);
-
-		/*
-		 * The stack length should grow continuously for all runs where
-		 * auto compaction is disabled. When enabled, we should merge
-		 * all tables in the stack.
-		 */
-		if (i != n)
-			check_int(st->merged->readers_len, ==, i + 1);
-		else
-			check_int(st->merged->readers_len, ==, 1);
-	}
-
-	reftable_stack_destroy(st);
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_compaction_with_locked_tables(void)
-{
-	struct reftable_write_options opts = {
-		.disable_auto_compact = 1,
-	};
-	struct reftable_stack *st = NULL;
-	struct reftable_buf buf = REFTABLE_BUF_INIT;
-	char *dir = get_tmp_dir(__LINE__);
-	int err;
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	write_n_ref_tables(st, 3);
-	check_int(st->merged->readers_len, ==, 3);
-
-	/* Lock one of the tables that we're about to compact. */
-	check(!reftable_buf_addstr(&buf, dir));
-	check(!reftable_buf_addstr(&buf, "/"));
-	check(!reftable_buf_addstr(&buf, st->readers[1]->name));
-	check(!reftable_buf_addstr(&buf, ".lock"));
-	write_file_buf(buf.buf, "", 0);
-
-	/*
-	 * Compaction is expected to fail given that we were not able to
-	 * compact all tables.
-	 */
-	err = reftable_stack_compact_all(st, NULL);
-	check_int(err, ==, REFTABLE_LOCK_ERROR);
-	check_int(st->stats.failures, ==, 1);
-	check_int(st->merged->readers_len, ==, 3);
-
-	reftable_stack_destroy(st);
-	reftable_buf_release(&buf);
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_compaction_concurrent(void)
-{
-	struct reftable_write_options opts = { 0 };
-	struct reftable_stack *st1 = NULL, *st2 = NULL;
-	char *dir = get_tmp_dir(__LINE__);
-	int err;
-
-	err = reftable_new_stack(&st1, dir, &opts);
-	check(!err);
-	write_n_ref_tables(st1, 3);
-
-	err = reftable_new_stack(&st2, dir, &opts);
-	check(!err);
-
-	err = reftable_stack_compact_all(st1, NULL);
-	check(!err);
-
-	reftable_stack_destroy(st1);
-	reftable_stack_destroy(st2);
-
-	check_int(count_dir_entries(dir), ==, 2);
-	clear_dir(dir);
-}
-
-static void unclean_stack_close(struct reftable_stack *st)
-{
-	/* break abstraction boundary to simulate unclean shutdown. */
-	for (size_t i = 0; i < st->readers_len; i++)
-		reftable_reader_decref(st->readers[i]);
-	st->readers_len = 0;
-	REFTABLE_FREE_AND_NULL(st->readers);
-}
-
-static void t_reftable_stack_compaction_concurrent_clean(void)
-{
-	struct reftable_write_options opts = { 0 };
-	struct reftable_stack *st1 = NULL, *st2 = NULL, *st3 = NULL;
-	char *dir = get_tmp_dir(__LINE__);
-	int err;
-
-	err = reftable_new_stack(&st1, dir, &opts);
-	check(!err);
-	write_n_ref_tables(st1, 3);
-
-	err = reftable_new_stack(&st2, dir, &opts);
-	check(!err);
-
-	err = reftable_stack_compact_all(st1, NULL);
-	check(!err);
-
-	unclean_stack_close(st1);
-	unclean_stack_close(st2);
-
-	err = reftable_new_stack(&st3, dir, &opts);
-	check(!err);
-
-	err = reftable_stack_clean(st3);
-	check(!err);
-	check_int(count_dir_entries(dir), ==, 2);
-
-	reftable_stack_destroy(st1);
-	reftable_stack_destroy(st2);
-	reftable_stack_destroy(st3);
-
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_read_across_reload(void)
-{
-	struct reftable_write_options opts = { 0 };
-	struct reftable_stack *st1 = NULL, *st2 = NULL;
-	struct reftable_ref_record rec = { 0 };
-	struct reftable_iterator it = { 0 };
-	char *dir = get_tmp_dir(__LINE__);
-	int err;
-
-	/* Create a first stack and set up an iterator for it. */
-	err = reftable_new_stack(&st1, dir, &opts);
-	check(!err);
-	write_n_ref_tables(st1, 2);
-	check_int(st1->merged->readers_len, ==, 2);
-	reftable_stack_init_ref_iterator(st1, &it);
-	err = reftable_iterator_seek_ref(&it, "");
-	check(!err);
-
-	/* Set up a second stack for the same directory and compact it. */
-	err = reftable_new_stack(&st2, dir, &opts);
-	check(!err);
-	check_int(st2->merged->readers_len, ==, 2);
-	err = reftable_stack_compact_all(st2, NULL);
-	check(!err);
-	check_int(st2->merged->readers_len, ==, 1);
-
-	/*
-	 * Verify that we can continue to use the old iterator even after we
-	 * have reloaded its stack.
-	 */
-	err = reftable_stack_reload(st1);
-	check(!err);
-	check_int(st1->merged->readers_len, ==, 1);
-	err = reftable_iterator_next_ref(&it, &rec);
-	check(!err);
-	check_str(rec.refname, "refs/heads/branch-0000");
-	err = reftable_iterator_next_ref(&it, &rec);
-	check(!err);
-	check_str(rec.refname, "refs/heads/branch-0001");
-	err = reftable_iterator_next_ref(&it, &rec);
-	check_int(err, >, 0);
-
-	reftable_ref_record_release(&rec);
-	reftable_iterator_destroy(&it);
-	reftable_stack_destroy(st1);
-	reftable_stack_destroy(st2);
-	clear_dir(dir);
-}
-
-static void t_reftable_stack_reload_with_missing_table(void)
-{
-	struct reftable_write_options opts = { 0 };
-	struct reftable_stack *st = NULL;
-	struct reftable_ref_record rec = { 0 };
-	struct reftable_iterator it = { 0 };
-	struct reftable_buf table_path = REFTABLE_BUF_INIT, content = REFTABLE_BUF_INIT;
-	char *dir = get_tmp_dir(__LINE__);
-	int err;
-
-	/* Create a first stack and set up an iterator for it. */
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-	write_n_ref_tables(st, 2);
-	check_int(st->merged->readers_len, ==, 2);
-	reftable_stack_init_ref_iterator(st, &it);
-	err = reftable_iterator_seek_ref(&it, "");
-	check(!err);
-
-	/*
-	 * Update the tables.list file with some garbage data, while reusing
-	 * our old readers. This should trigger a partial reload of the stack,
-	 * where we try to reuse our old readers.
-	*/
-	check(!reftable_buf_addstr(&content, st->readers[0]->name));
-	check(!reftable_buf_addstr(&content, "\n"));
-	check(!reftable_buf_addstr(&content, st->readers[1]->name));
-	check(!reftable_buf_addstr(&content, "\n"));
-	check(!reftable_buf_addstr(&content, "garbage\n"));
-	check(!reftable_buf_addstr(&table_path, st->list_file));
-	check(!reftable_buf_addstr(&table_path, ".lock"));
-	write_file_buf(table_path.buf, content.buf, content.len);
-	err = rename(table_path.buf, st->list_file);
-	check(!err);
-
-	err = reftable_stack_reload(st);
-	check_int(err, ==, -4);
-	check_int(st->merged->readers_len, ==, 2);
-
-	/*
-	 * Even though the reload has failed, we should be able to continue
-	 * using the iterator.
-	*/
-	err = reftable_iterator_next_ref(&it, &rec);
-	check(!err);
-	check_str(rec.refname, "refs/heads/branch-0000");
-	err = reftable_iterator_next_ref(&it, &rec);
-	check(!err);
-	check_str(rec.refname, "refs/heads/branch-0001");
-	err = reftable_iterator_next_ref(&it, &rec);
-	check_int(err, >, 0);
-
-	reftable_ref_record_release(&rec);
-	reftable_iterator_destroy(&it);
-	reftable_stack_destroy(st);
-	reftable_buf_release(&table_path);
-	reftable_buf_release(&content);
-	clear_dir(dir);
-}
-
-static int write_limits_after_ref(struct reftable_writer *wr, void *arg)
-{
-	struct reftable_ref_record *ref = arg;
-	check(!reftable_writer_set_limits(wr, ref->update_index, ref->update_index));
-	check(!reftable_writer_add_ref(wr, ref));
-	return reftable_writer_set_limits(wr, ref->update_index, ref->update_index);
-}
-
-static void t_reftable_invalid_limit_updates(void)
-{
-	struct reftable_ref_record ref = {
-		.refname = (char *) "HEAD",
-		.update_index = 1,
-		.value_type = REFTABLE_REF_SYMREF,
-		.value.symref = (char *) "master",
-	};
-	struct reftable_write_options opts = {
-		.default_permissions = 0660,
-	};
-	struct reftable_addition *add = NULL;
-	char *dir = get_tmp_dir(__LINE__);
-	struct reftable_stack *st = NULL;
-	int err;
-
-	err = reftable_new_stack(&st, dir, &opts);
-	check(!err);
-
-	reftable_addition_destroy(add);
-
-	err = reftable_stack_new_addition(&add, st, 0);
-	check(!err);
-
-	/*
-	 * write_limits_after_ref also updates the update indexes after adding
-	 * the record. This should cause an err to be returned, since the limits
-	 * must be set at the start.
-	 */
-	err = reftable_addition_add(add, write_limits_after_ref, &ref);
-	check_int(err, ==, REFTABLE_API_ERROR);
-
-	reftable_addition_destroy(add);
-	reftable_stack_destroy(st);
-	clear_dir(dir);
-}
-
-int cmd_main(int argc UNUSED, const char *argv[] UNUSED)
-{
-	TEST(t_empty_add(), "empty addition to stack");
-	TEST(t_read_file(), "read_lines works");
-	TEST(t_reflog_expire(), "expire reflog entries");
-	TEST(t_reftable_invalid_limit_updates(), "prevent limit updates after adding records");
-	TEST(t_reftable_stack_add(), "add multiple refs and logs to stack");
-	TEST(t_reftable_stack_add_one(), "add a single ref record to stack");
-	TEST(t_reftable_stack_add_performs_auto_compaction(), "addition to stack triggers auto-compaction");
-	TEST(t_reftable_stack_auto_compaction(), "stack must form geometric sequence after compaction");
-	TEST(t_reftable_stack_auto_compaction_factor(), "auto-compaction with non-default geometric factor");
-	TEST(t_reftable_stack_auto_compaction_fails_gracefully(), "failure on auto-compaction");
-	TEST(t_reftable_stack_auto_compaction_with_locked_tables(), "auto compaction with locked tables");
-	TEST(t_reftable_stack_compaction_concurrent(), "compaction with concurrent stack");
-	TEST(t_reftable_stack_compaction_concurrent_clean(), "compaction with unclean stack shutdown");
-	TEST(t_reftable_stack_compaction_with_locked_tables(), "compaction with locked tables");
-	TEST(t_reftable_stack_hash_id(), "read stack with wrong hash ID");
-	TEST(t_reftable_stack_iterator(), "log and ref iterator for reftable stack");
-	TEST(t_reftable_stack_lock_failure(), "stack addition with lockfile failure");
-	TEST(t_reftable_stack_log_normalize(), "log messages should be normalized");
-	TEST(t_reftable_stack_read_across_reload(), "stack iterators work across reloads");
-	TEST(t_reftable_stack_reload_with_missing_table(), "stack iteration with garbage tables");
-	TEST(t_reftable_stack_tombstone(), "'tombstone' refs in stack");
-	TEST(t_reftable_stack_transaction_api(), "update transaction to stack");
-	TEST(t_reftable_stack_transaction_with_reload(), "transaction with reload");
-	TEST(t_reftable_stack_transaction_api_performs_auto_compaction(), "update transaction triggers auto-compaction");
-	TEST(t_reftable_stack_update_index_check(), "update transactions with equal update indices");
-	TEST(t_reftable_stack_uptodate(), "stack must be reloaded before ref update");
-	TEST(t_suggest_compaction_segment(), "suggest_compaction_segment with basic input");
-	TEST(t_suggest_compaction_segment_nothing(), "suggest_compaction_segment with pre-compacted input");
-
-	return test_done();
-}
diff --git a/t/unit-tests/u-reftable-stack.c b/t/unit-tests/u-reftable-stack.c
new file mode 100644
index 0000000000..1f41d8c4ca
--- /dev/null
+++ b/t/unit-tests/u-reftable-stack.c
@@ -0,0 +1,1247 @@
+/*
+Copyright 2020 Google LLC
+
+Use of this source code is governed by a BSD-style
+license that can be found in the LICENSE file or at
+https://developers.google.com/open-source/licenses/bsd
+*/
+
+#define DISABLE_SIGN_COMPARE_WARNINGS
+
+#include "unit-test.h"
+#include "lib-reftable.h"
+#include "dir.h"
+#include "reftable/merged.h"
+#include "reftable/reader.h"
+#include "reftable/reftable-error.h"
+#include "reftable/stack.h"
+#include "strbuf.h"
+#include "tempfile.h"
+#include <dirent.h>
+
+static void clear_dir(const char *dirname)
+{
+	struct strbuf path = REFTABLE_BUF_INIT;
+	strbuf_addstr(&path, dirname);
+	remove_dir_recursively(&path, 0);
+	strbuf_release(&path);
+}
+
+static int count_dir_entries(const char *dirname)
+{
+	DIR *dir = opendir(dirname);
+	int len = 0;
+	struct dirent *d;
+	if (!dir)
+		return 0;
+
+	while ((d = readdir(dir))) {
+		/*
+		 * Besides skipping over "." and "..", we also need to
+		 * skip over other files that have a leading ".". This
+		 * is due to behaviour of NFS, which will rename files
+		 * to ".nfs*" to emulate delete-on-last-close.
+		 *
+		 * In any case this should be fine as the reftable
+		 * library will never write files with leading dots
+		 * anyway.
+		 */
+		if (starts_with(d->d_name, "."))
+			continue;
+		len++;
+	}
+	closedir(dir);
+	return len;
+}
+
+/*
+ * Work linenumber into the tempdir, so we can see which tests forget to
+ * cleanup.
+ */
+static char *get_tmp_template(int linenumber)
+{
+	const char *tmp = getenv("TMPDIR");
+	static char template[1024];
+	snprintf(template, sizeof(template) - 1, "%s/stack_test-%d.XXXXXX",
+		 tmp ? tmp : "/tmp", linenumber);
+	return template;
+}
+
+static char *get_tmp_dir(int linenumber)
+{
+	char *dir = get_tmp_template(linenumber);
+	cl_assert(mkdtemp(dir) != NULL);
+	return dir;
+}
+
+void test_reftable_stack__read_file(void)
+{
+	char *fn = get_tmp_template(__LINE__);
+	struct tempfile *tmp = mks_tempfile(fn);
+	int fd = get_tempfile_fd(tmp);
+	char out[1024] = "line1\n\nline2\nline3";
+	int n, err;
+	char **names = NULL;
+	const char *want[] = { "line1", "line2", "line3" };
+
+	cl_assert(fd > 0);
+	n = write_in_full(fd, out, strlen(out));
+	cl_assert_equal_i(n, strlen(out));
+	err = close(fd);
+	cl_assert(err >= 0);
+
+	err = read_lines(fn, &names);
+	cl_assert(err == 0);
+
+	for (size_t i = 0; names[i]; i++)
+		cl_assert_equal_s(want[i], names[i]);
+	free_names(names);
+	(void) remove(fn);
+	delete_tempfile(&tmp);
+}
+
+static int write_test_ref(struct reftable_writer *wr, void *arg)
+{
+	struct reftable_ref_record *ref = arg;
+	cl_assert(reftable_writer_set_limits(wr, ref->update_index,
+										 ref->update_index) == 0);
+	return reftable_writer_add_ref(wr, ref);
+}
+
+static void write_n_ref_tables(struct reftable_stack *st,
+			       size_t n)
+{
+	int disable_auto_compact;
+
+	disable_auto_compact = st->opts.disable_auto_compact;
+	st->opts.disable_auto_compact = 1;
+
+	for (size_t i = 0; i < n; i++) {
+		struct reftable_ref_record ref = {
+			.update_index = reftable_stack_next_update_index(st),
+			.value_type = REFTABLE_REF_VAL1,
+		};
+		char buf[128];
+
+		snprintf(buf, sizeof(buf), "refs/heads/branch-%04"PRIuMAX, (uintmax_t)i);
+		ref.refname = buf;
+		cl_reftable_set_hash(ref.value.val1, i, REFTABLE_HASH_SHA1);
+
+		cl_assert(reftable_stack_add(st, &write_test_ref, &ref) == 0);
+	}
+
+	st->opts.disable_auto_compact = disable_auto_compact;
+}
+
+struct write_log_arg {
+	struct reftable_log_record *log;
+	uint64_t update_index;
+};
+
+static int write_test_log(struct reftable_writer *wr, void *arg)
+{
+	struct write_log_arg *wla = arg;
+
+	cl_assert(reftable_writer_set_limits(wr, wla->update_index,
+										 wla->update_index) == 0);
+	return reftable_writer_add_log(wr, wla->log);
+}
+
+void test_reftable_stack__add_one(void)
+{
+	char *dir = get_tmp_dir(__LINE__);
+	struct reftable_buf scratch = REFTABLE_BUF_INIT;
+	int mask = umask(002);
+	struct reftable_write_options opts = {
+		.default_permissions = 0660,
+	};
+	struct reftable_stack *st = NULL;
+	struct reftable_ref_record ref = {
+		.refname = (char *) "HEAD",
+		.update_index = 1,
+		.value_type = REFTABLE_REF_SYMREF,
+		.value.symref = (char *) "master",
+	};
+	struct reftable_ref_record dest = { 0 };
+	struct stat stat_result = { 0 };
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+	cl_assert(reftable_stack_add(st, write_test_ref, &ref) == 0);
+	cl_assert(reftable_stack_read_ref(st, ref.refname, &dest) == 0);
+	cl_assert(reftable_ref_record_equal(&ref, &dest, REFTABLE_HASH_SIZE_SHA1) != 0);
+	cl_assert(st->readers_len > 0);
+
+#ifndef GIT_WINDOWS_NATIVE
+	cl_assert(reftable_buf_addstr(&scratch, dir) == 0);
+	cl_assert(reftable_buf_addstr(&scratch, "/tables.list") == 0);
+	cl_assert(stat(scratch.buf, &stat_result) == 0);
+	cl_assert_equal_i((stat_result.st_mode & 0777), opts.default_permissions);
+
+	reftable_buf_reset(&scratch);
+	cl_assert(reftable_buf_addstr(&scratch, dir) == 0);
+	cl_assert(reftable_buf_addstr(&scratch, "/") == 0);
+	/* do not try at home; not an external API for reftable. */
+	cl_assert(reftable_buf_addstr(&scratch, st->readers[0]->name) == 0);
+	cl_assert(stat(scratch.buf, &stat_result) == 0);
+	cl_assert_equal_i((stat_result.st_mode & 0777), opts.default_permissions);
+#else
+	(void) stat_result;
+#endif
+
+	reftable_ref_record_release(&dest);
+	reftable_stack_destroy(st);
+	reftable_buf_release(&scratch);
+	clear_dir(dir);
+	umask(mask);
+}
+
+void test_reftable_stack__uptodate(void)
+{
+	struct reftable_write_options opts = { 0 };
+	struct reftable_stack *st1 = NULL;
+	struct reftable_stack *st2 = NULL;
+	char *dir = get_tmp_dir(__LINE__);
+
+	struct reftable_ref_record ref1 = {
+		.refname = (char *) "HEAD",
+		.update_index = 1,
+		.value_type = REFTABLE_REF_SYMREF,
+		.value.symref = (char *) "master",
+	};
+	struct reftable_ref_record ref2 = {
+		.refname = (char *) "branch2",
+		.update_index = 2,
+		.value_type = REFTABLE_REF_SYMREF,
+		.value.symref = (char *) "master",
+	};
+
+
+	/* simulate multi-process access to the same stack
+	   by creating two stacks for the same directory.
+	 */
+	cl_assert(reftable_new_stack(&st1, dir, &opts) == 0);
+	cl_assert(reftable_new_stack(&st2, dir, &opts) == 0);
+	cl_assert(reftable_stack_add(st1, write_test_ref, &ref1) == 0);
+	cl_assert_equal_i(reftable_stack_add(st2, write_test_ref, &ref2),
+					  REFTABLE_OUTDATED_ERROR);
+	cl_assert(reftable_stack_reload(st2) == 0);
+	cl_assert(reftable_stack_add(st2, write_test_ref, &ref2) == 0);
+	reftable_stack_destroy(st1);
+	reftable_stack_destroy(st2);
+	clear_dir(dir);
+}
+
+void test_reftable_stack__transaction_api(void)
+{
+	char *dir = get_tmp_dir(__LINE__);
+	struct reftable_write_options opts = { 0 };
+	struct reftable_stack *st = NULL;
+	struct reftable_addition *add = NULL;
+
+	struct reftable_ref_record ref = {
+		.refname = (char *) "HEAD",
+		.update_index = 1,
+		.value_type = REFTABLE_REF_SYMREF,
+		.value.symref = (char *) "master",
+	};
+	struct reftable_ref_record dest = { 0 };
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+
+	reftable_addition_destroy(add);
+
+	cl_assert(reftable_stack_new_addition(&add, st, 0) == 0);
+	cl_assert(reftable_addition_add(add, write_test_ref, &ref) == 0);
+	cl_assert(reftable_addition_commit(add) == 0);
+
+	reftable_addition_destroy(add);
+
+	cl_assert(reftable_stack_read_ref(st, ref.refname, &dest) == 0);
+	cl_assert_equal_i(REFTABLE_REF_SYMREF, dest.value_type);
+	cl_assert(reftable_ref_record_equal(&ref, &dest, REFTABLE_HASH_SIZE_SHA1) != 0);
+
+	reftable_ref_record_release(&dest);
+	reftable_stack_destroy(st);
+	clear_dir(dir);
+}
+
+void test_reftable_stack__transaction_with_reload(void)
+{
+	char *dir = get_tmp_dir(__LINE__);
+	struct reftable_stack *st1 = NULL, *st2 = NULL;
+	struct reftable_addition *add = NULL;
+	struct reftable_ref_record refs[2] = {
+		{
+			.refname = (char *) "refs/heads/a",
+			.update_index = 1,
+			.value_type = REFTABLE_REF_VAL1,
+			.value.val1 = { '1' },
+		},
+		{
+			.refname = (char *) "refs/heads/b",
+			.update_index = 2,
+			.value_type = REFTABLE_REF_VAL1,
+			.value.val1 = { '1' },
+		},
+	};
+	struct reftable_ref_record ref = { 0 };
+
+	cl_assert(reftable_new_stack(&st1, dir, NULL) == 0);
+	cl_assert(reftable_new_stack(&st2, dir, NULL) == 0);
+	cl_assert(reftable_stack_new_addition(&add, st1, 0) == 0);
+	cl_assert(reftable_addition_add(add, write_test_ref, &refs[0]) == 0);
+	cl_assert(reftable_addition_commit(add) == 0);
+	reftable_addition_destroy(add);
+
+	/*
+	 * The second stack is now outdated, which we should notice. We do not
+	 * create the addition and lock the stack by default, but allow the
+	 * reload to happen when REFTABLE_STACK_NEW_ADDITION_RELOAD is set.
+	 */
+	cl_assert_equal_i(reftable_stack_new_addition(&add, st2, 0),
+					  REFTABLE_OUTDATED_ERROR);
+	cl_assert(reftable_stack_new_addition(&add, st2,
+										  REFTABLE_STACK_NEW_ADDITION_RELOAD) == 0);
+	cl_assert(reftable_addition_add(add, write_test_ref, &refs[1]) == 0);
+	cl_assert(reftable_addition_commit(add) == 0);
+	reftable_addition_destroy(add);
+
+	for (size_t i = 0; i < ARRAY_SIZE(refs); i++) {
+		cl_assert(reftable_stack_read_ref(st2, refs[i].refname, &ref) == 0);
+		cl_assert(reftable_ref_record_equal(&refs[i], &ref,
+											REFTABLE_HASH_SIZE_SHA1) != 0);
+	}
+
+	reftable_ref_record_release(&ref);
+	reftable_stack_destroy(st1);
+	reftable_stack_destroy(st2);
+	clear_dir(dir);
+}
+
+void test_reftable_stack__transaction_api_performs_auto_compaction(void)
+{
+	char *dir = get_tmp_dir(__LINE__);
+	struct reftable_write_options opts = {0};
+	struct reftable_addition *add = NULL;
+	struct reftable_stack *st = NULL;
+	size_t n = 20;
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+
+	for (size_t i = 0; i <= n; i++) {
+		struct reftable_ref_record ref = {
+			.update_index = reftable_stack_next_update_index(st),
+			.value_type = REFTABLE_REF_SYMREF,
+			.value.symref = (char *) "master",
+		};
+		char name[100];
+
+		snprintf(name, sizeof(name), "branch%04"PRIuMAX, (uintmax_t)i);
+		ref.refname = name;
+
+		/*
+		 * Disable auto-compaction for all but the last runs. Like this
+		 * we can ensure that we indeed honor this setting and have
+		 * better control over when exactly auto compaction runs.
+		 */
+		st->opts.disable_auto_compact = i != n;
+
+		cl_assert(reftable_stack_new_addition(&add, st, 0) == 0);
+		cl_assert(reftable_addition_add(add, write_test_ref, &ref) == 0);
+		cl_assert(reftable_addition_commit(add) == 0);
+
+		reftable_addition_destroy(add);
+
+		/*
+		 * The stack length should grow continuously for all runs where
+		 * auto compaction is disabled. When enabled, we should merge
+		 * all tables in the stack.
+		 */
+		if (i != n)
+			cl_assert_equal_i(st->merged->readers_len, i + 1);
+		else
+			cl_assert_equal_i(st->merged->readers_len, 1);
+	}
+
+	reftable_stack_destroy(st);
+	clear_dir(dir);
+}
+
+void test_reftable_stack__auto_compaction_fails_gracefully(void)
+{
+	struct reftable_ref_record ref = {
+		.refname = (char *) "refs/heads/master",
+		.update_index = 1,
+		.value_type = REFTABLE_REF_VAL1,
+		.value.val1 = {0x01},
+	};
+	struct reftable_write_options opts = { 0 };
+	struct reftable_stack *st;
+	struct reftable_buf table_path = REFTABLE_BUF_INIT;
+	char *dir = get_tmp_dir(__LINE__);
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+	cl_assert(reftable_stack_add(st, write_test_ref, &ref) == 0);
+	cl_assert_equal_i(st->merged->readers_len, 1);
+	cl_assert_equal_i(st->stats.attempts, 0);
+	cl_assert_equal_i(st->stats.failures, 0);
+
+	/*
+	 * Lock the newly written table such that it cannot be compacted.
+	 * Adding a new table to the stack should not be impacted by this, even
+	 * though auto-compaction will now fail.
+	 */
+	cl_assert(reftable_buf_addstr(&table_path, dir) == 0);
+	cl_assert(reftable_buf_addstr(&table_path, "/") == 0);
+	cl_assert(reftable_buf_addstr(&table_path, st->readers[0]->name) == 0);
+	cl_assert(reftable_buf_addstr(&table_path, ".lock") == 0);
+	write_file_buf(table_path.buf, "", 0);
+
+	ref.update_index = 2;
+	cl_assert(reftable_stack_add(st, write_test_ref, &ref) == 0);
+	cl_assert_equal_i(st->merged->readers_len, 2);
+	cl_assert_equal_i(st->stats.attempts, 1);
+	cl_assert_equal_i(st->stats.failures, 1);
+
+	reftable_stack_destroy(st);
+	reftable_buf_release(&table_path);
+	clear_dir(dir);
+}
+
+static int write_error(struct reftable_writer *wr UNUSED, void *arg)
+{
+	return *((int *)arg);
+}
+
+void test_reftable_stack__update_index_check(void)
+{
+	char *dir = get_tmp_dir(__LINE__);
+	struct reftable_write_options opts = { 0 };
+	struct reftable_stack *st = NULL;
+	struct reftable_ref_record ref1 = {
+		.refname = (char *) "name1",
+		.update_index = 1,
+		.value_type = REFTABLE_REF_SYMREF,
+		.value.symref = (char *) "master",
+	};
+	struct reftable_ref_record ref2 = {
+		.refname = (char *) "name2",
+		.update_index = 1,
+		.value_type = REFTABLE_REF_SYMREF,
+		.value.symref = (char *) "master",
+	};
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+	cl_assert(reftable_stack_add(st, write_test_ref, &ref1) == 0);
+	cl_assert_equal_i(reftable_stack_add(st, write_test_ref,
+										 &ref2), REFTABLE_API_ERROR);
+	reftable_stack_destroy(st);
+	clear_dir(dir);
+}
+
+void test_reftable_stack__lock_failure(void)
+{
+	char *dir = get_tmp_dir(__LINE__);
+	struct reftable_write_options opts = { 0 };
+	struct reftable_stack *st = NULL;
+	int i;
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+	for (i = -1; i != REFTABLE_EMPTY_TABLE_ERROR; i--)
+		cl_assert_equal_i(reftable_stack_add(st, write_error, &i), i);
+
+	reftable_stack_destroy(st);
+	clear_dir(dir);
+}
+
+void test_reftable_stack__add(void)
+{
+	struct reftable_write_options opts = {
+		.exact_log_message = 1,
+		.default_permissions = 0660,
+		.disable_auto_compact = 1,
+	};
+	struct reftable_stack *st = NULL;
+	char *dir = get_tmp_dir(__LINE__);
+	struct reftable_ref_record refs[2] = { 0 };
+	struct reftable_log_record logs[2] = { 0 };
+	struct reftable_buf path = REFTABLE_BUF_INIT;
+	struct stat stat_result;
+	size_t i, N = ARRAY_SIZE(refs);
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+
+	for (i = 0; i < N; i++) {
+		char buf[256];
+		snprintf(buf, sizeof(buf), "branch%02"PRIuMAX, (uintmax_t)i);
+		refs[i].refname = xstrdup(buf);
+		refs[i].update_index = i + 1;
+		refs[i].value_type = REFTABLE_REF_VAL1;
+		cl_reftable_set_hash(refs[i].value.val1, i, REFTABLE_HASH_SHA1);
+
+		logs[i].refname = xstrdup(buf);
+		logs[i].update_index = N + i + 1;
+		logs[i].value_type = REFTABLE_LOG_UPDATE;
+		logs[i].value.update.email = xstrdup("identity@invalid");
+		cl_reftable_set_hash(logs[i].value.update.new_hash, i, REFTABLE_HASH_SHA1);
+	}
+
+	for (i = 0; i < N; i++)
+		cl_assert(reftable_stack_add(st, write_test_ref, &refs[i]) == 0);
+
+	for (i = 0; i < N; i++) {
+		struct write_log_arg arg = {
+			.log = &logs[i],
+			.update_index = reftable_stack_next_update_index(st),
+		};
+		cl_assert(reftable_stack_add(st, write_test_log, &arg) == 0);
+	}
+
+	cl_assert(reftable_stack_compact_all(st, NULL) == 0);
+
+	for (i = 0; i < N; i++) {
+		struct reftable_ref_record dest = { 0 };
+
+		cl_assert(reftable_stack_read_ref(st, refs[i].refname, &dest) == 0);
+		cl_assert(reftable_ref_record_equal(&dest, refs + i,
+											REFTABLE_HASH_SIZE_SHA1) != 0);
+		reftable_ref_record_release(&dest);
+	}
+
+	for (i = 0; i < N; i++) {
+		struct reftable_log_record dest = { 0 };
+		cl_assert(reftable_stack_read_log(st, refs[i].refname, &dest) == 0);
+		cl_assert(reftable_log_record_equal(&dest, logs + i,
+											REFTABLE_HASH_SIZE_SHA1) != 0);
+		reftable_log_record_release(&dest);
+	}
+
+#ifndef GIT_WINDOWS_NATIVE
+	cl_assert(reftable_buf_addstr(&path, dir) == 0);
+	cl_assert(reftable_buf_addstr(&path, "/tables.list") == 0);
+	cl_assert(stat(path.buf, &stat_result) == 0);
+	cl_assert_equal_i((stat_result.st_mode & 0777), opts.default_permissions);
+
+	reftable_buf_reset(&path);
+	cl_assert(reftable_buf_addstr(&path, dir) == 0);
+	cl_assert(reftable_buf_addstr(&path, "/") == 0);
+	/* do not try at home; not an external API for reftable. */
+	cl_assert(reftable_buf_addstr(&path, st->readers[0]->name) == 0);
+	cl_assert(stat(path.buf, &stat_result) == 0);
+	cl_assert_equal_i((stat_result.st_mode & 0777), opts.default_permissions);
+#else
+	(void) stat_result;
+#endif
+
+	/* cleanup */
+	reftable_stack_destroy(st);
+	for (i = 0; i < N; i++) {
+		reftable_ref_record_release(&refs[i]);
+		reftable_log_record_release(&logs[i]);
+	}
+	reftable_buf_release(&path);
+	clear_dir(dir);
+}
+
+void test_reftable_stack__iterator(void)
+{
+	struct reftable_write_options opts = { 0 };
+	struct reftable_stack *st = NULL;
+	char *dir = get_tmp_dir(__LINE__);
+	struct reftable_ref_record refs[10] = { 0 };
+	struct reftable_log_record logs[10] = { 0 };
+	struct reftable_iterator it = { 0 };
+	size_t N = ARRAY_SIZE(refs), i;
+	int err;
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+
+	for (i = 0; i < N; i++) {
+		refs[i].refname = xstrfmt("branch%02"PRIuMAX, (uintmax_t)i);
+		refs[i].update_index = i + 1;
+		refs[i].value_type = REFTABLE_REF_VAL1;
+		cl_reftable_set_hash(refs[i].value.val1, i, REFTABLE_HASH_SHA1);
+
+		logs[i].refname = xstrfmt("branch%02"PRIuMAX, (uintmax_t)i);
+		logs[i].update_index = i + 1;
+		logs[i].value_type = REFTABLE_LOG_UPDATE;
+		logs[i].value.update.email = xstrdup("johndoe@invalid");
+		logs[i].value.update.message = xstrdup("commit\n");
+		cl_reftable_set_hash(logs[i].value.update.new_hash, i, REFTABLE_HASH_SHA1);
+	}
+
+	for (i = 0; i < N; i++)
+		cl_assert(reftable_stack_add(st, write_test_ref, &refs[i]) == 0);
+
+	for (i = 0; i < N; i++) {
+		struct write_log_arg arg = {
+			.log = &logs[i],
+			.update_index = reftable_stack_next_update_index(st),
+		};
+
+		cl_assert(reftable_stack_add(st, write_test_log, &arg) == 0);
+	}
+
+	reftable_stack_init_ref_iterator(st, &it);
+	reftable_iterator_seek_ref(&it, refs[0].refname);
+	for (i = 0; ; i++) {
+		struct reftable_ref_record ref = { 0 };
+
+		err = reftable_iterator_next_ref(&it, &ref);
+		if (err > 0)
+			break;
+		cl_assert(err == 0);
+		cl_assert(reftable_ref_record_equal(&ref, &refs[i],
+											REFTABLE_HASH_SIZE_SHA1) != 0);
+		reftable_ref_record_release(&ref);
+	}
+	cl_assert_equal_i(i, N);
+
+	reftable_iterator_destroy(&it);
+
+	cl_assert(reftable_stack_init_log_iterator(st, &it) == 0);
+
+	reftable_iterator_seek_log(&it, logs[0].refname);
+	for (i = 0; ; i++) {
+		struct reftable_log_record log = { 0 };
+
+		err = reftable_iterator_next_log(&it, &log);
+		if (err > 0)
+			break;
+		cl_assert(err == 0);
+		cl_assert(reftable_log_record_equal(&log, &logs[i],
+											REFTABLE_HASH_SIZE_SHA1) != 0);
+		reftable_log_record_release(&log);
+	}
+	cl_assert_equal_i(i, N);
+
+	reftable_stack_destroy(st);
+	reftable_iterator_destroy(&it);
+	for (i = 0; i < N; i++) {
+		reftable_ref_record_release(&refs[i]);
+		reftable_log_record_release(&logs[i]);
+	}
+	clear_dir(dir);
+}
+
+void test_reftable_stack__log_normalize(void)
+{
+	struct reftable_write_options opts = {
+		0,
+	};
+	struct reftable_stack *st = NULL;
+	char *dir = get_tmp_dir(__LINE__);
+	struct reftable_log_record input = {
+		.refname = (char *) "branch",
+		.update_index = 1,
+		.value_type = REFTABLE_LOG_UPDATE,
+		.value = {
+			.update = {
+				.new_hash = { 1 },
+				.old_hash = { 2 },
+			},
+		},
+	};
+	struct reftable_log_record dest = {
+		.update_index = 0,
+	};
+	struct write_log_arg arg = {
+		.log = &input,
+		.update_index = 1,
+	};
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+
+	input.value.update.message = (char *) "one\ntwo";
+	cl_assert_equal_i(reftable_stack_add(st, write_test_log,
+										 &arg), REFTABLE_API_ERROR);
+
+	input.value.update.message = (char *) "one";
+	cl_assert(reftable_stack_add(st, write_test_log, &arg) == 0);
+	cl_assert(reftable_stack_read_log(st, input.refname, &dest) == 0);
+	cl_assert_equal_s(dest.value.update.message, "one\n");
+
+	input.value.update.message = (char *) "two\n";
+	arg.update_index = 2;
+	cl_assert(reftable_stack_add(st, write_test_log, &arg) == 0);
+	cl_assert(reftable_stack_read_log(st, input.refname, &dest) == 0);
+	cl_assert_equal_s(dest.value.update.message, "two\n");
+
+	/* cleanup */
+	reftable_stack_destroy(st);
+	reftable_log_record_release(&dest);
+	clear_dir(dir);
+}
+
+void test_reftable_stack__tombstone(void)
+{
+	char *dir = get_tmp_dir(__LINE__);
+	struct reftable_write_options opts = { 0 };
+	struct reftable_stack *st = NULL;
+	struct reftable_ref_record refs[2] = { 0 };
+	struct reftable_log_record logs[2] = { 0 };
+	size_t i, N = ARRAY_SIZE(refs);
+	struct reftable_ref_record dest = { 0 };
+	struct reftable_log_record log_dest = { 0 };
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+
+	/* even entries add the refs, odd entries delete them. */
+	for (i = 0; i < N; i++) {
+		const char *buf = "branch";
+		refs[i].refname = xstrdup(buf);
+		refs[i].update_index = i + 1;
+		if (i % 2 == 0) {
+			refs[i].value_type = REFTABLE_REF_VAL1;
+			cl_reftable_set_hash(refs[i].value.val1, i,
+					    REFTABLE_HASH_SHA1);
+		}
+
+		logs[i].refname = xstrdup(buf);
+		/*
+		 * update_index is part of the key so should be constant.
+		 * The value itself should be less than the writer's upper
+		 * limit.
+		 */
+		logs[i].update_index = 1;
+		if (i % 2 == 0) {
+			logs[i].value_type = REFTABLE_LOG_UPDATE;
+			cl_reftable_set_hash(logs[i].value.update.new_hash, i,
+					    REFTABLE_HASH_SHA1);
+			logs[i].value.update.email =
+				xstrdup("identity@invalid");
+		}
+	}
+	for (i = 0; i < N; i++)
+		cl_assert(reftable_stack_add(st, write_test_ref, &refs[i]) == 0);
+
+	for (i = 0; i < N; i++) {
+		struct write_log_arg arg = {
+			.log = &logs[i],
+			.update_index = reftable_stack_next_update_index(st),
+		};
+		cl_assert(reftable_stack_add(st, write_test_log, &arg) == 0);
+	}
+
+	cl_assert_equal_i(reftable_stack_read_ref(st, "branch", &dest), 1);
+	reftable_ref_record_release(&dest);
+
+	cl_assert_equal_i(reftable_stack_read_log(st, "branch", &log_dest), 1);
+	reftable_log_record_release(&log_dest);
+
+	cl_assert(reftable_stack_compact_all(st, NULL) == 0);
+	cl_assert_equal_i(reftable_stack_read_ref(st, "branch", &dest), 1);
+	cl_assert_equal_i(reftable_stack_read_log(st, "branch", &log_dest), 1);
+	reftable_ref_record_release(&dest);
+	reftable_log_record_release(&log_dest);
+
+	/* cleanup */
+	reftable_stack_destroy(st);
+	for (i = 0; i < N; i++) {
+		reftable_ref_record_release(&refs[i]);
+		reftable_log_record_release(&logs[i]);
+	}
+	clear_dir(dir);
+}
+
+void test_reftable_stack__hash_id(void)
+{
+	char *dir = get_tmp_dir(__LINE__);
+	struct reftable_write_options opts = { 0 };
+	struct reftable_stack *st = NULL;
+
+	struct reftable_ref_record ref = {
+		.refname = (char *) "master",
+		.value_type = REFTABLE_REF_SYMREF,
+		.value.symref = (char *) "target",
+		.update_index = 1,
+	};
+	struct reftable_write_options opts32 = { .hash_id = REFTABLE_HASH_SHA256 };
+	struct reftable_stack *st32 = NULL;
+	struct reftable_write_options opts_default = { 0 };
+	struct reftable_stack *st_default = NULL;
+	struct reftable_ref_record dest = { 0 };
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+	cl_assert(reftable_stack_add(st, write_test_ref, &ref) == 0);
+
+	/* can't read it with the wrong hash ID. */
+	cl_assert_equal_i(reftable_new_stack(&st32, dir, &opts32), REFTABLE_FORMAT_ERROR);
+
+	/* check that we can read it back with default opts too. */
+	cl_assert(reftable_new_stack(&st_default, dir, &opts_default) == 0);
+	cl_assert(reftable_stack_read_ref(st_default, "master", &dest) == 0);
+	cl_assert(reftable_ref_record_equal(&ref, &dest, REFTABLE_HASH_SIZE_SHA1) != 0);
+	reftable_ref_record_release(&dest);
+	reftable_stack_destroy(st);
+	reftable_stack_destroy(st_default);
+	clear_dir(dir);
+}
+
+void test_reftable_stack__suggest_compaction_segment(void)
+{
+	uint64_t sizes[] = { 512, 64, 17, 16, 9, 9, 9, 16, 2, 16 };
+	struct segment min =
+		suggest_compaction_segment(sizes, ARRAY_SIZE(sizes), 2);
+	cl_assert_equal_i(min.start, 1);
+	cl_assert_equal_i(min.end, 10);
+}
+
+void test_reftable_stack__suggest_compaction_segment_nothing(void)
+{
+	uint64_t sizes[] = { 64, 32, 16, 8, 4, 2 };
+	struct segment result =
+		suggest_compaction_segment(sizes, ARRAY_SIZE(sizes), 2);
+	cl_assert_equal_i(result.start, result.end);
+}
+
+void test_reftable_stack__reflog_expire(void)
+{
+	char *dir = get_tmp_dir(__LINE__);
+	struct reftable_write_options opts = { 0 };
+	struct reftable_stack *st = NULL;
+	struct reftable_log_record logs[20] = { 0 };
+	size_t i, N = ARRAY_SIZE(logs) - 1;
+	struct reftable_log_expiry_config expiry = {
+		.time = 10,
+	};
+	struct reftable_log_record log = { 0 };
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+
+	for (i = 1; i <= N; i++) {
+		char buf[256];
+		snprintf(buf, sizeof(buf), "branch%02"PRIuMAX, (uintmax_t)i);
+
+		logs[i].refname = xstrdup(buf);
+		logs[i].update_index = i;
+		logs[i].value_type = REFTABLE_LOG_UPDATE;
+		logs[i].value.update.time = i;
+		logs[i].value.update.email = xstrdup("identity@invalid");
+		cl_reftable_set_hash(logs[i].value.update.new_hash, i,
+				    REFTABLE_HASH_SHA1);
+	}
+
+	for (i = 1; i <= N; i++) {
+		struct write_log_arg arg = {
+			.log = &logs[i],
+			.update_index = reftable_stack_next_update_index(st),
+		};
+		cl_assert(reftable_stack_add(st, write_test_log, &arg) == 0);
+	}
+
+	cl_assert(reftable_stack_compact_all(st, NULL) == 0);
+	cl_assert(reftable_stack_compact_all(st, &expiry) == 0);
+	cl_assert_equal_i(reftable_stack_read_log(st, logs[9].refname, &log), 1);
+	cl_assert(reftable_stack_read_log(st, logs[11].refname, &log) == 0);
+
+	expiry.min_update_index = 15;
+	cl_assert(reftable_stack_compact_all(st, &expiry) == 0);
+	cl_assert_equal_i(reftable_stack_read_log(st, logs[14].refname, &log), 1);
+	cl_assert(reftable_stack_read_log(st, logs[16].refname, &log) == 0);
+
+	/* cleanup */
+	reftable_stack_destroy(st);
+	for (i = 0; i <= N; i++)
+		reftable_log_record_release(&logs[i]);
+	clear_dir(dir);
+	reftable_log_record_release(&log);
+}
+
+static int write_nothing(struct reftable_writer *wr, void *arg UNUSED)
+{
+	cl_assert(reftable_writer_set_limits(wr, 1, 1) == 0);
+	return 0;
+}
+
+void test_reftable_stack__empty_add(void)
+{
+	struct reftable_write_options opts = { 0 };
+	struct reftable_stack *st = NULL;
+	char *dir = get_tmp_dir(__LINE__);
+	struct reftable_stack *st2 = NULL;
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+	cl_assert(reftable_stack_add(st, write_nothing, NULL) == 0);
+	cl_assert(reftable_new_stack(&st2, dir, &opts) == 0);
+	clear_dir(dir);
+	reftable_stack_destroy(st);
+	reftable_stack_destroy(st2);
+}
+
+static int fastlogN(uint64_t sz, uint64_t N)
+{
+	int l = 0;
+	if (sz == 0)
+		return 0;
+	for (; sz; sz /= N)
+		l++;
+	return l - 1;
+}
+
+void test_reftable_stack__auto_compaction(void)
+{
+	struct reftable_write_options opts = {
+		.disable_auto_compact = 1,
+	};
+	struct reftable_stack *st = NULL;
+	char *dir = get_tmp_dir(__LINE__);
+	size_t i, N = 100;
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+
+	for (i = 0; i < N; i++) {
+		char name[100];
+		struct reftable_ref_record ref = {
+			.refname = name,
+			.update_index = reftable_stack_next_update_index(st),
+			.value_type = REFTABLE_REF_SYMREF,
+			.value.symref = (char *) "master",
+		};
+		snprintf(name, sizeof(name), "branch%04"PRIuMAX, (uintmax_t)i);
+
+		cl_assert(reftable_stack_add(st, write_test_ref, &ref) == 0);
+		cl_assert(reftable_stack_auto_compact(st) == 0);
+		cl_assert(i < 2 || st->merged->readers_len < 2 * fastlogN(i, 2));
+	}
+
+	cl_assert(reftable_stack_compaction_stats(st)->entries_written <
+			  (uint64_t)(N * fastlogN(N, 2)));
+
+	reftable_stack_destroy(st);
+	clear_dir(dir);
+}
+
+void test_reftable_stack__auto_compaction_factor(void)
+{
+	struct reftable_write_options opts = {
+		.auto_compaction_factor = 5,
+	};
+	struct reftable_stack *st = NULL;
+	char *dir = get_tmp_dir(__LINE__);
+	size_t N = 100;
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+
+	for (size_t i = 0; i < N; i++) {
+		char name[20];
+		struct reftable_ref_record ref = {
+			.refname = name,
+			.update_index = reftable_stack_next_update_index(st),
+			.value_type = REFTABLE_REF_VAL1,
+		};
+		xsnprintf(name, sizeof(name), "branch%04"PRIuMAX, (uintmax_t)i);
+
+		cl_assert(reftable_stack_add(st, &write_test_ref, &ref) == 0);
+		cl_assert(i < 5 || st->merged->readers_len < 5 * fastlogN(i, 5));
+	}
+
+	reftable_stack_destroy(st);
+	clear_dir(dir);
+}
+
+void test_reftable_stack__auto_compaction_with_locked_tables(void)
+{
+	struct reftable_write_options opts = {
+		.disable_auto_compact = 1,
+	};
+	struct reftable_stack *st = NULL;
+	struct reftable_buf buf = REFTABLE_BUF_INIT;
+	char *dir = get_tmp_dir(__LINE__);
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+
+	write_n_ref_tables(st, 5);
+	cl_assert_equal_i(st->merged->readers_len, 5);
+
+	/*
+	 * Given that all tables we have written should be roughly the same
+	 * size, we expect that auto-compaction will want to compact all of the
+	 * tables. Locking any of the tables will keep it from doing so.
+	 */
+	cl_assert(reftable_buf_addstr(&buf, dir) == 0);
+	cl_assert(reftable_buf_addstr(&buf, "/") == 0);
+	cl_assert(reftable_buf_addstr(&buf, st->readers[2]->name) == 0);
+	cl_assert(reftable_buf_addstr(&buf, ".lock") == 0);
+	write_file_buf(buf.buf, "", 0);
+
+	/*
+	 * When parts of the stack are locked, then auto-compaction does a best
+	 * effort compaction of those tables which aren't locked. So while this
+	 * would in theory compact all tables, due to the preexisting lock we
+	 * only compact the newest two tables.
+	 */
+	cl_assert(reftable_stack_auto_compact(st) == 0);
+	cl_assert_equal_i(st->stats.failures, 0);
+	cl_assert_equal_i(st->merged->readers_len, 4);
+
+	reftable_stack_destroy(st);
+	reftable_buf_release(&buf);
+	clear_dir(dir);
+}
+
+void test_reftable_stack__add_performs_auto_compaction(void)
+{
+	struct reftable_write_options opts = { 0 };
+	struct reftable_stack *st = NULL;
+	char *dir = get_tmp_dir(__LINE__);
+	size_t i, n = 20;
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+
+	for (i = 0; i <= n; i++) {
+		struct reftable_ref_record ref = {
+			.update_index = reftable_stack_next_update_index(st),
+			.value_type = REFTABLE_REF_SYMREF,
+			.value.symref = (char *) "master",
+		};
+		char buf[128];
+
+		/*
+		 * Disable auto-compaction for all but the last runs. Like this
+		 * we can ensure that we indeed honor this setting and have
+		 * better control over when exactly auto compaction runs.
+		 */
+		st->opts.disable_auto_compact = i != n;
+
+		snprintf(buf, sizeof(buf), "branch-%04"PRIuMAX, (uintmax_t)i);
+		ref.refname = buf;
+
+		cl_assert(reftable_stack_add(st, write_test_ref, &ref) == 0);
+
+		/*
+		 * The stack length should grow continuously for all runs where
+		 * auto compaction is disabled. When enabled, we should merge
+		 * all tables in the stack.
+		 */
+		if (i != n)
+			cl_assert_equal_i(st->merged->readers_len, i + 1);
+		else
+			cl_assert_equal_i(st->merged->readers_len, 1);
+	}
+
+	reftable_stack_destroy(st);
+	clear_dir(dir);
+}
+
+void test_reftable_stack__compaction_with_locked_tables(void)
+{
+	struct reftable_write_options opts = {
+		.disable_auto_compact = 1,
+	};
+	struct reftable_stack *st = NULL;
+	struct reftable_buf buf = REFTABLE_BUF_INIT;
+	char *dir = get_tmp_dir(__LINE__);
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+
+	write_n_ref_tables(st, 3);
+	cl_assert_equal_i(st->merged->readers_len, 3);
+
+	/* Lock one of the tables that we're about to compact. */
+	cl_assert(reftable_buf_addstr(&buf, dir) == 0);
+	cl_assert(reftable_buf_addstr(&buf, "/") == 0);
+	cl_assert(reftable_buf_addstr(&buf, st->readers[1]->name) == 0);
+	cl_assert(reftable_buf_addstr(&buf, ".lock") == 0);
+	write_file_buf(buf.buf, "", 0);
+
+	/*
+	 * Compaction is expected to fail given that we were not able to
+	 * compact all tables.
+	 */
+	cl_assert_equal_i(reftable_stack_compact_all(st, NULL), REFTABLE_LOCK_ERROR);
+	cl_assert_equal_i(st->stats.failures, 1);
+	cl_assert_equal_i(st->merged->readers_len, 3);
+
+	reftable_stack_destroy(st);
+	reftable_buf_release(&buf);
+	clear_dir(dir);
+}
+
+void test_reftable_stack__compaction_concurrent(void)
+{
+	struct reftable_write_options opts = { 0 };
+	struct reftable_stack *st1 = NULL, *st2 = NULL;
+	char *dir = get_tmp_dir(__LINE__);
+
+	cl_assert(reftable_new_stack(&st1, dir, &opts) == 0);
+	write_n_ref_tables(st1, 3);
+
+	cl_assert(reftable_new_stack(&st2, dir, &opts) == 0);
+	cl_assert(reftable_stack_compact_all(st1, NULL) == 0);
+
+	reftable_stack_destroy(st1);
+	reftable_stack_destroy(st2);
+
+	cl_assert_equal_i(count_dir_entries(dir), 2);
+	clear_dir(dir);
+}
+
+static void unclean_stack_close(struct reftable_stack *st)
+{
+	/* break abstraction boundary to simulate unclean shutdown. */
+	for (size_t i = 0; i < st->readers_len; i++)
+		reftable_reader_decref(st->readers[i]);
+	st->readers_len = 0;
+	REFTABLE_FREE_AND_NULL(st->readers);
+}
+
+void test_reftable_stack__compaction_concurrent_clean(void)
+{
+	struct reftable_write_options opts = { 0 };
+	struct reftable_stack *st1 = NULL, *st2 = NULL, *st3 = NULL;
+	char *dir = get_tmp_dir(__LINE__);
+
+	cl_assert(reftable_new_stack(&st1, dir, &opts) == 0);
+	write_n_ref_tables(st1, 3);
+
+	cl_assert(reftable_new_stack(&st2, dir, &opts) == 0);
+	cl_assert(reftable_stack_compact_all(st1, NULL) == 0);
+
+	unclean_stack_close(st1);
+	unclean_stack_close(st2);
+
+	cl_assert(reftable_new_stack(&st3, dir, &opts) == 0);
+	cl_assert(reftable_stack_clean(st3) == 0);
+	cl_assert_equal_i(count_dir_entries(dir), 2);
+
+	reftable_stack_destroy(st1);
+	reftable_stack_destroy(st2);
+	reftable_stack_destroy(st3);
+
+	clear_dir(dir);
+}
+
+void test_reftable_stack__read_across_reload(void)
+{
+	struct reftable_write_options opts = { 0 };
+	struct reftable_stack *st1 = NULL, *st2 = NULL;
+	struct reftable_ref_record rec = { 0 };
+	struct reftable_iterator it = { 0 };
+	char *dir = get_tmp_dir(__LINE__);
+
+	/* Create a first stack and set up an iterator for it. */
+	cl_assert(reftable_new_stack(&st1, dir, &opts) == 0);
+	write_n_ref_tables(st1, 2);
+	cl_assert_equal_i(st1->merged->readers_len, 2);
+	reftable_stack_init_ref_iterator(st1, &it);
+	cl_assert(reftable_iterator_seek_ref(&it, "") == 0);
+
+	/* Set up a second stack for the same directory and compact it. */
+	cl_assert(reftable_new_stack(&st2, dir, &opts) == 0);
+	cl_assert_equal_i(st2->merged->readers_len, 2);
+	cl_assert(reftable_stack_compact_all(st2, NULL) == 0);
+	cl_assert_equal_i(st2->merged->readers_len, 1);
+
+	/*
+	 * Verify that we can continue to use the old iterator even after we
+	 * have reloaded its stack.
+	 */
+	cl_assert(reftable_stack_reload(st1) == 0);
+	cl_assert_equal_i(st1->merged->readers_len, 1);
+	cl_assert(reftable_iterator_next_ref(&it, &rec) == 0);
+	cl_assert_equal_s(rec.refname, "refs/heads/branch-0000");
+	cl_assert(reftable_iterator_next_ref(&it, &rec) == 0);
+	cl_assert_equal_s(rec.refname, "refs/heads/branch-0001");
+	cl_assert(reftable_iterator_next_ref(&it, &rec) > 0);
+
+	reftable_ref_record_release(&rec);
+	reftable_iterator_destroy(&it);
+	reftable_stack_destroy(st1);
+	reftable_stack_destroy(st2);
+	clear_dir(dir);
+}
+
+void test_reftable_stack__reload_with_missing_table(void)
+{
+	struct reftable_write_options opts = { 0 };
+	struct reftable_stack *st = NULL;
+	struct reftable_ref_record rec = { 0 };
+	struct reftable_iterator it = { 0 };
+	struct reftable_buf table_path = REFTABLE_BUF_INIT, content = REFTABLE_BUF_INIT;
+	char *dir = get_tmp_dir(__LINE__);
+
+	/* Create a first stack and set up an iterator for it. */
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+	write_n_ref_tables(st, 2);
+	cl_assert_equal_i(st->merged->readers_len, 2);
+	reftable_stack_init_ref_iterator(st, &it);
+	cl_assert(reftable_iterator_seek_ref(&it, "") == 0);
+
+	/*
+	 * Update the tables.list file with some garbage data, while reusing
+	 * our old readers. This should trigger a partial reload of the stack,
+	 * where we try to reuse our old readers.
+	*/
+	cl_assert(reftable_buf_addstr(&content, st->readers[0]->name) == 0);
+	cl_assert(reftable_buf_addstr(&content, "\n") == 0);
+	cl_assert(reftable_buf_addstr(&content, st->readers[1]->name) == 0);
+	cl_assert(reftable_buf_addstr(&content, "\n") == 0);
+	cl_assert(reftable_buf_addstr(&content, "garbage\n") == 0);
+	cl_assert(reftable_buf_addstr(&table_path, st->list_file) == 0);
+	cl_assert(reftable_buf_addstr(&table_path, ".lock") == 0);
+	write_file_buf(table_path.buf, content.buf, content.len);
+	cl_assert(rename(table_path.buf, st->list_file) == 0);
+
+	cl_assert_equal_i(reftable_stack_reload(st), -4);
+	cl_assert_equal_i(st->merged->readers_len, 2);
+
+	/*
+	 * Even though the reload has failed, we should be able to continue
+	 * using the iterator.
+	*/
+	cl_assert(reftable_iterator_next_ref(&it, &rec) == 0);
+	cl_assert_equal_s(rec.refname, "refs/heads/branch-0000");
+	cl_assert(reftable_iterator_next_ref(&it, &rec) == 0);
+	cl_assert_equal_s(rec.refname, "refs/heads/branch-0001");
+	cl_assert(reftable_iterator_next_ref(&it, &rec) > 0);
+
+	reftable_ref_record_release(&rec);
+	reftable_iterator_destroy(&it);
+	reftable_stack_destroy(st);
+	reftable_buf_release(&table_path);
+	reftable_buf_release(&content);
+	clear_dir(dir);
+}
+
+static int write_limits_after_ref(struct reftable_writer *wr, void *arg)
+{
+	struct reftable_ref_record *ref = arg;
+	cl_assert(reftable_writer_set_limits(wr, ref->update_index,
+										 ref->update_index) == 0);
+	cl_assert(reftable_writer_add_ref(wr, ref) == 0);
+	return reftable_writer_set_limits(wr, ref->update_index, ref->update_index);
+}
+
+void test_reftable_stack__invalid_limit_updates(void)
+{
+	struct reftable_ref_record ref = {
+		.refname = (char *) "HEAD",
+		.update_index = 1,
+		.value_type = REFTABLE_REF_SYMREF,
+		.value.symref = (char *) "master",
+	};
+	struct reftable_write_options opts = {
+		.default_permissions = 0660,
+	};
+	struct reftable_addition *add = NULL;
+	char *dir = get_tmp_dir(__LINE__);
+	struct reftable_stack *st = NULL;
+
+	cl_assert(reftable_new_stack(&st, dir, &opts) == 0);
+
+	reftable_addition_destroy(add);
+
+	cl_assert(reftable_stack_new_addition(&add, st, 0) == 0);
+
+	/*
+	 * write_limits_after_ref also updates the update indexes after adding
+	 * the record. This should cause an err to be returned, since the limits
+	 * must be set at the start.
+	 */
+	cl_assert_equal_i(reftable_addition_add(add, write_limits_after_ref, &ref),
+					  REFTABLE_API_ERROR);
+
+	reftable_addition_destroy(add);
+	reftable_stack_destroy(st);
+	clear_dir(dir);
+}
-- 
2.43.0