[PATCH 0/9] More index cleanups

This is based on ds/cache-tree-basics.

Here are a few more cleanups that are vaguely related to the index. I
discovered these while preparing my sparse-index RFC that I intend to send
early next week.

The biggest patch is the final one, which creates a test script for
comparing sparse-checkouts to full checkouts. There are some commands that
do not behave similarly. This script will be the backbone of my testing
strategy for the sparse-index by adding a new mode to compare
sparse-checkouts with the two index types (full and sparse).

Thanks, -Stolee

Derrick Stolee (9):
  cache-tree: clean up cache_tree_update()
  cache-tree: extract subtree_pos()
  fsmonitor: de-duplicate BUG()s around dirty bits
  repository: add repo reference to index_state
  name-hash: use trace2 regions for init
  sparse-checkout: load sparse-checkout patterns
  sparse-checkout: hold pattern list in index
  test-lib: test_region looks for trace2 regions
  t1092: test interesting sparse-checkout scenarios

 builtin/sparse-checkout.c                |  22 +-
 cache-tree.c                             |  20 +-
 cache-tree.h                             |   2 +
 cache.h                                  |   3 +
 dir.c                                    |  17 ++
 dir.h                                    |   2 +
 fsmonitor.c                              |  27 +-
 name-hash.c                              |   3 +
 repository.c                             |   4 +
 t/t0500-progress-display.sh              |   3 +-
 t/t1092-sparse-checkout-compatibility.sh | 323 +++++++++++++++++++++++
 t/test-lib-functions.sh                  |  40 +++
 unpack-trees.c                           |   6 +-
 13 files changed, 431 insertions(+), 41 deletions(-)
 create mode 100755 t/t1092-sparse-checkout-compatibility.sh


base-commit: a4b6d202caad83c6dc29abe9b17e53a1b3fb54a0
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-839%2Fderrickstolee%2Fmore-index-cleanups-v1
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-839/derrickstolee/more-index-cleanups-v1
Pull-Request: https://github.com/gitgitgadget/git/pull/839
-- 
gitgitgadget

From: Derrick Stolee <redacted>

Make the method safer by allocating a cache_tree member for the given
index_state if it is not already present.

Also drop local variables that are used exactly once and can be found
directly from the 'istate' parameter.

Signed-off-by: Derrick Stolee <redacted>
---
 cache-tree.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/cache-tree.c b/cache-tree.c
index 3f1a8d4f1b7..c1e49901c17 100644
--- a/cache-tree.c
+++ b/cache-tree.c

@@ -436,16 +436,20 @@ static int update_one(struct cache_tree *it,
 
 int cache_tree_update(struct index_state *istate, int flags)
 {
-	struct cache_tree *it = istate->cache_tree;
-	struct cache_entry **cache = istate->cache;
-	int entries = istate->cache_nr;
-	int skip, i = verify_cache(cache, entries, flags);
+	int skip, i;
+
+	i = verify_cache(istate->cache, istate->cache_nr, flags);
 
 	if (i)
 		return i;
+
+	if (!istate->cache_tree)
+		istate->cache_tree = cache_tree();
+
 	trace_performance_enter();
 	trace2_region_enter("cache_tree", "update", the_repository);
-	i = update_one(it, cache, entries, "", 0, &skip, flags);
+	i = update_one(istate->cache_tree, istate->cache, istate->cache_nr,
+		       "", 0, &skip, flags);
 	trace2_region_leave("cache_tree", "update", the_repository);
 	trace_performance_leave("cache_tree_update");
 	if (i < 0)

-- 
gitgitgadget

From: Derrick Stolee <redacted>

The index has an fsmonitor_dirty bitmap that records which index entries
are "dirty" based on the response from the FSMonitor. If this bitmap
ever grows larger than the index, then there was an error in how it was
constructed, and it was probably a developer's bug.

There are several BUG() statements that are very similar, so replace
these uses with a simpler assert_index_minimum(). Since there is one
caller that uses a custom 'pos' value instead of the bit_size member, we
cannot simplify it too much. However, the error string is identical in
each, so this simplifies things.

The end result is that the code is simpler to read while also preserving
these assertions for developers in the FSMonitor space.

Signed-off-by: Derrick Stolee <redacted>
---
 fsmonitor.c | 27 +++++++++++++--------------
 1 file changed, 13 insertions(+), 14 deletions(-)

diff --git a/fsmonitor.c b/fsmonitor.c
index ca031c3abb8..52a50a9545a 100644
--- a/fsmonitor.c
+++ b/fsmonitor.c

@@ -13,14 +13,19 @@
 
 struct trace_key trace_fsmonitor = TRACE_KEY_INIT(FSMONITOR);
 
+static void assert_index_minimum(struct index_state *istate, size_t pos)
+{
+	if (pos > istate->cache_nr)
+		BUG("fsmonitor_dirty has more entries than the index (%"PRIuMAX" > %u)",
+		    (uintmax_t)pos, istate->cache_nr);
+}
+
 static void fsmonitor_ewah_callback(size_t pos, void *is)
 {
 	struct index_state *istate = (struct index_state *)is;
 	struct cache_entry *ce;
 
-	if (pos >= istate->cache_nr)
-		BUG("fsmonitor_dirty has more entries than the index (%"PRIuMAX" >= %u)",
-		    (uintmax_t)pos, istate->cache_nr);
+	assert_index_minimum(istate, pos);
 
 	ce = istate->cache[pos];
 	ce->ce_flags &= ~CE_FSMONITOR_VALID;

@@ -82,10 +87,8 @@ int read_fsmonitor_extension(struct index_state *istate, const void *data,
 	}
 	istate->fsmonitor_dirty = fsmonitor_dirty;
 
-	if (!istate->split_index &&
-	    istate->fsmonitor_dirty->bit_size > istate->cache_nr)
-		BUG("fsmonitor_dirty has more entries than the index (%"PRIuMAX" > %u)",
-		    (uintmax_t)istate->fsmonitor_dirty->bit_size, istate->cache_nr);
+	if (!istate->split_index)
+		assert_index_minimum(istate, istate->fsmonitor_dirty->bit_size);
 
 	trace_printf_key(&trace_fsmonitor, "read fsmonitor extension successful");
 	return 0;

@@ -110,10 +113,8 @@ void write_fsmonitor_extension(struct strbuf *sb, struct index_state *istate)
 	uint32_t ewah_size = 0;
 	int fixup = 0;
 
-	if (!istate->split_index &&
-	    istate->fsmonitor_dirty->bit_size > istate->cache_nr)
-		BUG("fsmonitor_dirty has more entries than the index (%"PRIuMAX" > %u)",
-		    (uintmax_t)istate->fsmonitor_dirty->bit_size, istate->cache_nr);
+	if (!istate->split_index)
+		assert_index_minimum(istate, istate->fsmonitor_dirty->bit_size);
 
 	put_be32(&hdr_version, INDEX_EXTENSION_VERSION2);
 	strbuf_add(sb, &hdr_version, sizeof(uint32_t));

@@ -335,9 +336,7 @@ void tweak_fsmonitor(struct index_state *istate)
 			}
 
 			/* Mark all previously saved entries as dirty */
-			if (istate->fsmonitor_dirty->bit_size > istate->cache_nr)
-				BUG("fsmonitor_dirty has more entries than the index (%"PRIuMAX" > %u)",
-				    (uintmax_t)istate->fsmonitor_dirty->bit_size, istate->cache_nr);
+			assert_index_minimum(istate, istate->fsmonitor_dirty->bit_size);
 			ewah_each_bit(istate->fsmonitor_dirty, fsmonitor_ewah_callback, istate);
 
 			refresh_fsmonitor(istate);

-- 
gitgitgadget

From: Derrick Stolee <redacted>

It will be helpful to add behavior to index opertations that might
trigger an object lookup. Since each index belongs to a specific
repository, add a 'repo' pointer to struct index_state that allows
access to this repository.

This will prevent future changes from needing to pass an additional
'struct repository *repo' parameter and instead rely only on the 'struct
index_state *istate' parameter.

Signed-off-by: Derrick Stolee <redacted>
---
 cache.h      | 1 +
 repository.c | 4 ++++
 2 files changed, 5 insertions(+)

diff --git a/cache.h b/cache.h
index 71097657489..f9c7a603841 100644
--- a/cache.h
+++ b/cache.h

@@ -328,6 +328,7 @@ struct index_state {
 	struct ewah_bitmap *fsmonitor_dirty;
 	struct mem_pool *ce_mem_pool;
 	struct progress *progress;
+	struct repository *repo;
 };
 
 /* Name hashing */

diff --git a/repository.c b/repository.c
index a4174ddb062..67a4c1da2d9 100644
--- a/repository.c
+++ b/repository.c

@@ -264,6 +264,10 @@ int repo_read_index(struct repository *repo)
 	if (!repo->index)
 		repo->index = xcalloc(1, sizeof(*repo->index));
 
+	/* Complete the double-reference */
+	if (!repo->index->repo)
+		repo->index->repo = repo;
+
 	return read_index_from(repo->index, repo->index_file, repo->gitdir);
 }
 

-- 
gitgitgadget

From: Derrick Stolee <redacted>

The lazy_init_name_hash() populates a hashset with all filenames and
another with all directories represented in the index. This is run only
if we need to use the hashsets to check for existence or case-folding
renames.

Place trace2 regions where there is already a performance trace.

Signed-off-by: Derrick Stolee <redacted>
---
 name-hash.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/name-hash.c b/name-hash.c
index 5d3c7b12c18..4e03fac9bb1 100644
--- a/name-hash.c
+++ b/name-hash.c

@@ -7,6 +7,7 @@
  */
 #include "cache.h"
 #include "thread-utils.h"
+#include "trace2.h"
 
 struct dir_entry {
 	struct hashmap_entry ent;

@@ -577,6 +578,7 @@ static void lazy_init_name_hash(struct index_state *istate)
 	if (istate->name_hash_initialized)
 		return;
 	trace_performance_enter();
+	trace2_region_enter("index", "name-hash-init", istate->repo);
 	hashmap_init(&istate->name_hash, cache_entry_cmp, NULL, istate->cache_nr);
 	hashmap_init(&istate->dir_hash, dir_entry_cmp, NULL, istate->cache_nr);
 

@@ -597,6 +599,7 @@ static void lazy_init_name_hash(struct index_state *istate)
 	}
 
 	istate->name_hash_initialized = 1;
+	trace2_region_leave("index", "name-hash-init", istate->repo);
 	trace_performance_leave("initialize name hash");
 }
 

-- 
gitgitgadget

From: Derrick Stolee <redacted>

A future feature will want to load the sparse-checkout patterns into a
pattern_list, but the current mechanism to do so is a bit complicated.
This is made difficult due to needing to find the sparse-checkout file
in different ways throughout the codebase.

The logic implemented in the new get_sparse_checkout_patterns() was
duplicated in populate_from_existing_patterns() in unpack-trees.c. Use
the new method instead, keeping the logic around handling the struct
unpack_trees_options.

The callers to get_sparse_checkout_filename() in
builtin/sparse-checkout.c manipulate the sparse-checkout file directly,
so it is not appropriate to replace logic in that file with
get_sparse_checkout_patterns().

Signed-off-by: Derrick Stolee <redacted>
---
 builtin/sparse-checkout.c |  5 -----
 dir.c                     | 17 +++++++++++++++++
 dir.h                     |  2 ++
 unpack-trees.c            |  6 +-----
 4 files changed, 20 insertions(+), 10 deletions(-)

diff --git a/builtin/sparse-checkout.c b/builtin/sparse-checkout.c
index e3140db2a0a..2306a9ad98e 100644
--- a/builtin/sparse-checkout.c
+++ b/builtin/sparse-checkout.c

@@ -22,11 +22,6 @@ static char const * const builtin_sparse_checkout_usage[] = {
 	NULL
 };
 
-static char *get_sparse_checkout_filename(void)
-{
-	return git_pathdup("info/sparse-checkout");
-}
-
 static void write_patterns_to_file(FILE *fp, struct pattern_list *pl)
 {
 	int i;

diff --git a/dir.c b/dir.c
index d637461da5c..d153a63bbd1 100644
--- a/dir.c
+++ b/dir.c

@@ -2998,6 +2998,23 @@ void setup_standard_excludes(struct dir_struct *dir)
 	}
 }
 
+char *get_sparse_checkout_filename(void)
+{
+	return git_pathdup("info/sparse-checkout");
+}
+
+int get_sparse_checkout_patterns(struct pattern_list *pl)
+{
+	int res;
+	char *sparse_filename = get_sparse_checkout_filename();
+
+	pl->use_cone_patterns = core_sparse_checkout_cone;
+	res = add_patterns_from_file_to_list(sparse_filename, "", 0, pl, NULL);
+
+	free(sparse_filename);
+	return res;
+}
+
 int remove_path(const char *name)
 {
 	char *slash;

diff --git a/dir.h b/dir.h
index a3c40dec516..facfae47402 100644
--- a/dir.h
+++ b/dir.h

@@ -448,6 +448,8 @@ int is_empty_dir(const char *dir);
 
 void setup_standard_excludes(struct dir_struct *dir);
 
+char *get_sparse_checkout_filename(void);
+int get_sparse_checkout_patterns(struct pattern_list *pl);
 
 /* Constants for remove_dir_recursively: */
 

diff --git a/unpack-trees.c b/unpack-trees.c
index af6e9b9c2fd..837b8bb42fb 100644
--- a/unpack-trees.c
+++ b/unpack-trees.c

@@ -1549,14 +1549,10 @@ static void mark_new_skip_worktree(struct pattern_list *pl,
 static void populate_from_existing_patterns(struct unpack_trees_options *o,
 					    struct pattern_list *pl)
 {
-	char *sparse = git_pathdup("info/sparse-checkout");
-
-	pl->use_cone_patterns = core_sparse_checkout_cone;
-	if (add_patterns_from_file_to_list(sparse, "", 0, pl, NULL) < 0)
+	if (get_sparse_checkout_patterns(pl) < 0)
 		o->skip_sparse_checkout = 1;
 	else
 		o->pl = pl;
-	free(sparse);
 }
 
 

-- 
gitgitgadget

From: Derrick Stolee <redacted>

As we modify the sparse-checkout definition, we perform index operations
on a pattern_list that only exists in-memory. This allows easy backing
out in case the index update fails.

However, if the index write itself cares about the sparse-checkout
pattern set, we need access to that in-memory copy. Place a pointer to
a 'struct pattern_list' in the index so we can access this on-demand.

Signed-off-by: Derrick Stolee <redacted>
---
 builtin/sparse-checkout.c | 17 ++++++++++-------
 cache.h                   |  2 ++
 2 files changed, 12 insertions(+), 7 deletions(-)

diff --git a/builtin/sparse-checkout.c b/builtin/sparse-checkout.c
index 2306a9ad98e..e00b82af727 100644
--- a/builtin/sparse-checkout.c
+++ b/builtin/sparse-checkout.c

@@ -110,6 +110,8 @@ static int update_working_directory(struct pattern_list *pl)
 	if (is_index_unborn(r->index))
 		return UPDATE_SPARSITY_SUCCESS;
 
+	r->index->sparse_checkout_patterns = pl;
+
 	memset(&o, 0, sizeof(o));
 	o.verbose_update = isatty(2);
 	o.update = 1;

@@ -138,6 +140,7 @@ static int update_working_directory(struct pattern_list *pl)
 	else
 		rollback_lock_file(&lock_file);
 
+	r->index->sparse_checkout_patterns = NULL;
 	return result;
 }
 

@@ -517,19 +520,18 @@ static int modify_pattern_list(int argc, const char **argv, enum modify_type m)
 {
 	int result;
 	int changed_config = 0;
-	struct pattern_list pl;
-	memset(&pl, 0, sizeof(pl));
+	struct pattern_list *pl = xcalloc(1, sizeof(*pl));
 
 	switch (m) {
 	case ADD:
 		if (core_sparse_checkout_cone)
-			add_patterns_cone_mode(argc, argv, &pl);
+			add_patterns_cone_mode(argc, argv, pl);
 		else
-			add_patterns_literal(argc, argv, &pl);
+			add_patterns_literal(argc, argv, pl);
 		break;
 
 	case REPLACE:
-		add_patterns_from_input(&pl, argc, argv);
+		add_patterns_from_input(pl, argc, argv);
 		break;
 	}
 

@@ -539,12 +541,13 @@ static int modify_pattern_list(int argc, const char **argv, enum modify_type m)
 		changed_config = 1;
 	}
 
-	result = write_patterns_and_update(&pl);
+	result = write_patterns_and_update(pl);
 
 	if (result && changed_config)
 		set_config(MODE_NO_PATTERNS);
 
-	clear_pattern_list(&pl);
+	clear_pattern_list(pl);
+	free(pl);
 	return result;
 }
 

diff --git a/cache.h b/cache.h
index f9c7a603841..bf4ec3de4b0 100644
--- a/cache.h
+++ b/cache.h

@@ -305,6 +305,7 @@ static inline unsigned int canon_mode(unsigned int mode)
 struct split_index;
 struct untracked_cache;
 struct progress;
+struct pattern_list;
 
 struct index_state {
 	struct cache_entry **cache;

@@ -329,6 +330,7 @@ struct index_state {
 	struct mem_pool *ce_mem_pool;
 	struct progress *progress;
 	struct repository *repo;
+	struct pattern_list *sparse_checkout_patterns;
 };
 
 /* Name hashing */

-- 
gitgitgadget

From: Derrick Stolee <redacted>

Most test cases can verify Git's behavior using input/output
expectations or changes to the .git directory. However, sometimes we
want to check that Git did or did not run a certain section of code.
This is particularly important for performance-only features that we
want to ensure have been enabled in certain cases.

Add a new 'test_region' function that checks if a trace2 region was
entered and left in a given trace2 event log.

There is one existing test (t0500-progress-display.sh) that performs
this check already, so use the helper function instead. More uses will
be added in a later change.

t6423-merge-rename-directories.sh also greps for region_enter lines, but
it verifies the number of such lines, which is not the same as an
existence check.

Signed-off-by: Derrick Stolee <redacted>
---
 t/t0500-progress-display.sh |  3 +--
 t/test-lib-functions.sh     | 40 +++++++++++++++++++++++++++++++++++++
 2 files changed, 41 insertions(+), 2 deletions(-)

diff --git a/t/t0500-progress-display.sh b/t/t0500-progress-display.sh
index 1ed1df351cb..c461b89dfaf 100755
--- a/t/t0500-progress-display.sh
+++ b/t/t0500-progress-display.sh

@@ -303,8 +303,7 @@ test_expect_success 'progress generates traces' '
 		"Working hard" <in 2>stderr &&
 
 	# t0212/parse_events.perl intentionally omits regions and data.
-	grep -e "region_enter" -e "\"category\":\"progress\"" trace.event &&
-	grep -e "region_leave" -e "\"category\":\"progress\"" trace.event &&
+	test_region category progress trace.event &&
 	grep "\"key\":\"total_objects\",\"value\":\"40\"" trace.event &&
 	grep "\"key\":\"total_bytes\",\"value\":\"409600\"" trace.event
 '

diff --git a/t/test-lib-functions.sh b/t/test-lib-functions.sh
index 999982fe4a9..c878db93013 100644
--- a/t/test-lib-functions.sh
+++ b/t/test-lib-functions.sh

@@ -1655,3 +1655,43 @@ test_subcommand () {
 		grep "\[$expr\]"
 	fi
 }
+
+# Check that the given command was invoked as part of the
+# trace2-format trace on stdin.
+#
+#	test_region [!] <category> <label> git <command> <args>...
+#
+# For example, to look for trace2_region_enter("index", "do_read_index", repo)
+# in an invocation of "git checkout HEAD~1", run
+#
+#	GIT_TRACE2_EVENT="$(pwd)/trace.txt" GIT_TRACE2_EVENT_NESTING=10 \
+#		git checkout HEAD~1 &&
+#	test_region index do_read_index <trace.txt
+#
+# If the first parameter passed is !, this instead checks that
+# the given region was not entered.
+#
+test_region () {
+	local expect_exit=0
+	if test "$1" = "!"
+	then
+		expect_exit=1
+		shift
+	fi
+
+	grep -e "region_enter" -e "\"category\":\"$1\",\"label\":\"$2\"" "$3"
+	exitcode=$?
+
+	if test $exitcode != $expect_exit
+	then
+		return 1
+	fi
+
+	grep -e "region_leave" -e "\"category\":\"$1\",\"label\":\"$2\"" "$3"
+	exitcode=$?
+
+	if test $exitcode != $expect_exit
+	then
+		return 1
+	fi
+}

-- 
gitgitgadget

From: Derrick Stolee <redacted>

This method will be helpful to use outside of cache-tree.c in a later
feature. The implementation is subtle due to subtree_name_cmp() sorting
by length and then lexicographically.

Signed-off-by: Derrick Stolee <redacted>
---
 cache-tree.c | 6 +++---
 cache-tree.h | 2 ++
 2 files changed, 5 insertions(+), 3 deletions(-)

diff --git a/cache-tree.c b/cache-tree.c
index c1e49901c17..2b130dd5e19 100644
--- a/cache-tree.c
+++ b/cache-tree.c

@@ -45,7 +45,7 @@ static int subtree_name_cmp(const char *one, int onelen,
 	return memcmp(one, two, onelen);
 }
 
-static int subtree_pos(struct cache_tree *it, const char *path, int pathlen)
+int cache_tree_subtree_pos(struct cache_tree *it, const char *path, int pathlen)
 {
 	struct cache_tree_sub **down = it->down;
 	int lo, hi;

@@ -72,7 +72,7 @@ static struct cache_tree_sub *find_subtree(struct cache_tree *it,
 					   int create)
 {
 	struct cache_tree_sub *down;
-	int pos = subtree_pos(it, path, pathlen);
+	int pos = cache_tree_subtree_pos(it, path, pathlen);
 	if (0 <= pos)
 		return it->down[pos];
 	if (!create)

@@ -123,7 +123,7 @@ static int do_invalidate_path(struct cache_tree *it, const char *path)
 	it->entry_count = -1;
 	if (!*slash) {
 		int pos;
-		pos = subtree_pos(it, path, namelen);
+		pos = cache_tree_subtree_pos(it, path, namelen);
 		if (0 <= pos) {
 			cache_tree_free(&it->down[pos]->cache_tree);
 			free(it->down[pos]);

diff --git a/cache-tree.h b/cache-tree.h
index 639bfa5340e..8efeccebfc9 100644
--- a/cache-tree.h
+++ b/cache-tree.h

@@ -27,6 +27,8 @@ void cache_tree_free(struct cache_tree **);
 void cache_tree_invalidate_path(struct index_state *, const char *);
 struct cache_tree_sub *cache_tree_sub(struct cache_tree *, const char *);
 
+int cache_tree_subtree_pos(struct cache_tree *it, const char *path, int pathlen);
+
 void cache_tree_write(struct strbuf *, struct cache_tree *root);
 struct cache_tree *cache_tree_read(const char *buffer, unsigned long size);
 

-- 
gitgitgadget

From: Derrick Stolee <redacted>

These also document some behaviors that differ from a full checkout, and
possibly in a way that is not intended.

Signed-off-by: Derrick Stolee <redacted>
---
 t/t1092-sparse-checkout-compatibility.sh | 323 +++++++++++++++++++++++
 1 file changed, 323 insertions(+)
 create mode 100755 t/t1092-sparse-checkout-compatibility.sh

diff --git a/t/t1092-sparse-checkout-compatibility.sh b/t/t1092-sparse-checkout-compatibility.sh
new file mode 100755
index 00000000000..46f9dc2cdf3
--- /dev/null
+++ b/t/t1092-sparse-checkout-compatibility.sh

@@ -0,0 +1,323 @@
+#!/bin/sh
+
+test_description='compare full workdir to sparse workdir'
+
+. ./test-lib.sh
+
+test_expect_success 'setup' '
+	git init initial-repo &&
+	(
+		cd initial-repo &&
+		echo a >a &&
+		echo "after deep" >e &&
+		echo "after folder1" >g &&
+		mkdir folder1 folder2 deep x &&
+		mkdir deep/deeper1 deep/deeper2 &&
+		mkdir deep/deeper1/deepest &&
+		echo "after deeper1" >deep/e &&
+		echo "after deepest" >deep/deeper1/e &&
+		cp a folder1 &&
+		cp a folder2 &&
+		cp a deep &&
+		cp a deep/deeper1 &&
+		cp a deep/deeper2 &&
+		cp a deep/deeper1/deepest &&
+		git add . &&
+		git commit -m "initial commit" &&
+		git checkout -b base &&
+		for dir in folder1 folder2 deep
+		do
+			git checkout -b update-$dir &&
+			echo "updated $dir" >$dir/a &&
+			git commit -a -m "update $dir" || return 1
+		done &&
+
+		git checkout -b rename-base base &&
+		echo >folder1/larger-content <<-\EOF &&
+		matching
+		lines
+		help
+		inexact
+		renames
+		EOF
+		cp folder1/larger-content folder2/ &&
+		cp folder1/larger-content deep/deeper1/ &&
+		git add . &&
+		git commit -m "add interesting rename content" &&
+
+		git checkout -b rename-out-to-out rename-base &&
+		mv folder1/a folder2/b &&
+		mv folder1/larger-content folder2/edited-content &&
+		echo >>folder2/edited-content &&
+		git add . &&
+		git commit -m "rename folder1/... to folder2/..." &&
+
+		git checkout -b rename-out-to-in rename-base &&
+		mv folder1/a deep/deeper1/b &&
+		mv folder1/larger-content deep/deeper1/edited-content &&
+		echo >>deep/deeper1/edited-content &&
+		git add . &&
+		git commit -m "rename folder1/... to deep/deeper1/..." &&
+
+		git checkout -b rename-in-to-out rename-base &&
+		mv deep/deeper1/a folder1/b &&
+		mv deep/deeper1/larger-content folder1/edited-content &&
+		echo >>folder1/edited-content &&
+		git add . &&
+		git commit -m "rename deep/deeper1/... to folder1/..." &&
+
+		git checkout -b deepest base &&
+		echo "updated deepest" >deep/deeper1/deepest/a &&
+		git commit -a -m "update deepest" &&
+
+		git checkout -f base &&
+		git reset --hard
+	)
+'
+
+init_repos () {
+	rm -rf full-checkout sparse-checkout sparse-index &&
+
+	# create repos in initial state
+	cp -r initial-repo full-checkout &&
+	git -C full-checkout reset --hard &&
+
+	cp -r initial-repo sparse-checkout &&
+	git -C sparse-checkout reset --hard &&
+	git -C sparse-checkout sparse-checkout init --cone &&
+
+	# initialize sparse-checkout definitions
+	git -C sparse-checkout sparse-checkout set deep
+}
+
+run_on_sparse () {
+	(
+		cd sparse-checkout &&
+		$* >../sparse-checkout-out 2>../sparse-checkout-err
+	)
+}
+
+run_on_all () {
+	(
+		cd full-checkout &&
+		$* >../full-checkout-out 2>../full-checkout-err
+	) &&
+	run_on_sparse $*
+}
+
+test_all_match () {
+	run_on_all $* &&
+	test_cmp full-checkout-out sparse-checkout-out &&
+	test_cmp full-checkout-err sparse-checkout-err
+}
+
+test_expect_success 'status with options' '
+	init_repos &&
+	test_all_match git status --porcelain=v2 &&
+	test_all_match git status --porcelain=v2 -z -u &&
+	test_all_match git status --porcelain=v2 -uno &&
+	run_on_all "touch README.md" &&
+	test_all_match git status --porcelain=v2 &&
+	test_all_match git status --porcelain=v2 -z -u &&
+	test_all_match git status --porcelain=v2 -uno &&
+	test_all_match git add README.md &&
+	test_all_match git status --porcelain=v2 &&
+	test_all_match git status --porcelain=v2 -z -u &&
+	test_all_match git status --porcelain=v2 -uno
+'
+
+test_expect_success 'add, commit, checkout' '
+	init_repos &&
+
+	write_script edit-contents <<-\EOF &&
+	echo text >>README.md
+	EOF
+	run_on_all "../edit-contents" &&
+
+	test_all_match git add README.md &&
+	test_all_match git status --porcelain=v2 &&
+	test_all_match git commit -m "Add README.md" &&
+
+	test_all_match git checkout HEAD~1 &&
+	test_all_match git checkout - &&
+
+	run_on_all "../edit-contents" &&
+
+	test_all_match git add -A &&
+	test_all_match git status --porcelain=v2 &&
+	test_all_match git commit -m "Extend README.md" &&
+
+	test_all_match git checkout HEAD~1 &&
+	test_all_match git checkout -
+'
+
+test_expect_success 'add, commit, checkout' '
+	init_repos &&
+
+	write_script edit-contents <<-\EOF &&
+	echo text >>$1
+	EOF
+	run_on_all "../edit-contents README.md" &&
+
+	test_all_match git add README.md &&
+	test_all_match git status --porcelain=v2 &&
+	test_all_match git commit -m "Add README.md" &&
+
+	test_all_match git checkout HEAD~1 &&
+	test_all_match git checkout - &&
+
+	run_on_all "../edit-contents README.md" &&
+
+	test_all_match git add -A &&
+	test_all_match git status --porcelain=v2 &&
+	test_all_match git commit -m "Extend README.md" &&
+
+	test_all_match git checkout HEAD~1 &&
+	test_all_match git checkout - &&
+
+	run_on_all "../edit-contents deep/newfile" &&
+
+	test_all_match git status --porcelain=v2 -uno &&
+	test_all_match git status --porcelain=v2 &&
+	test_all_match git add . &&
+	test_all_match git status --porcelain=v2 &&
+	test_all_match git commit -m "add deep/newfile" &&
+
+	test_all_match git checkout HEAD~1 &&
+	test_all_match git checkout -
+'
+
+test_expect_success 'checkout and reset --hard' '
+	init_repos &&
+
+	test_all_match git checkout update-folder1 &&
+	test_all_match git status --porcelain=v2 &&
+
+	test_all_match git checkout update-deep &&
+	test_all_match git status --porcelain=v2 &&
+
+	test_all_match git checkout -b reset-test &&
+	test_all_match git reset --hard deepest &&
+	test_all_match git reset --hard update-folder1 &&
+	test_all_match git reset --hard update-folder2
+'
+
+test_expect_success 'diff --staged' '
+	init_repos &&
+
+	write_script edit-contents <<-\EOF &&
+	echo text >>README.md
+	EOF
+	run_on_all "../edit-contents" &&
+
+	test_all_match git diff &&
+	test_all_match git diff --staged &&
+	test_all_match git add README.md &&
+	test_all_match git diff &&
+	test_all_match git diff --staged
+'
+
+test_expect_success 'diff with renames' '
+	init_repos &&
+
+	for branch in rename-out-to-out rename-out-to-in rename-in-to-out
+	do
+		test_all_match git checkout rename-base &&
+		test_all_match git checkout $branch -- .&&
+		test_all_match git diff --staged &&
+		test_all_match git diff --staged --find-renames || return 1
+	done
+'
+
+test_expect_success 'log with pathspec outside sparse definition' '
+	init_repos &&
+
+	test_all_match git log -- a &&
+	test_all_match git log -- folder1/a &&
+	test_all_match git log -- folder2/a &&
+	test_all_match git log -- deep/a &&
+	test_all_match git log -- deep/deeper1/a &&
+	test_all_match git log -- deep/deeper1/deepest/a &&
+
+	test_all_match git checkout update-folder1 &&
+	test_all_match git log -- folder1/a
+'
+
+test_expect_success 'blame with pathspec inside sparse definition' '
+	init_repos &&
+
+	test_all_match git blame a &&
+	test_all_match git blame deep/a &&
+	test_all_match git blame deep/deeper1/a &&
+	test_all_match git blame deep/deeper1/deepest/a
+'
+
+# TODO: blame currently does not support blaming files outside of the
+# sparse definition. It complains that the file doesn't exist locally.
+test_expect_failure 'blame with pathspec outside sparse definition' '
+	init_repos &&
+
+	test_all_match git blame folder1/a &&
+	test_all_match git blame folder2/a &&
+	test_all_match git blame deep/deeper2/a &&
+	test_all_match git blame deep/deeper2/deepest/a
+'
+
+# TODO: reset currently does not behave as expected when in a
+# sparse-checkout.
+test_expect_failure 'checkout and reset (mixed)' '
+	init_repos &&
+
+	test_all_match git checkout -b reset-test update-deep &&
+	test_all_match git reset deepest &&
+	test_all_match git reset update-folder1 &&
+	test_all_match git reset update-folder2
+'
+
+test_expect_success 'merge' '
+	init_repos &&
+
+	test_all_match git checkout -b merge update-deep &&
+	test_all_match git merge -m "folder1" update-folder1 &&
+	test_all_match git rev-parse HEAD^{tree} &&
+	test_all_match git merge -m "folder2" update-folder2 &&
+	test_all_match git rev-parse HEAD^{tree}
+'
+
+test_expect_success 'merge with outside renames' '
+	init_repos &&
+
+	for type in out-to-out out-to-in in-to-out
+	do
+		test_all_match git reset --hard &&
+		test_all_match git checkout -f -b merge-$type update-deep &&
+		test_all_match git merge -m "$type" rename-$type &&
+		test_all_match git rev-parse HEAD^{tree} || return 1
+	done
+'
+
+test_expect_success 'clean' '
+	init_repos &&
+
+	echo bogus >>.gitignore &&
+	run_on_all cp ../.gitignore . &&
+	test_all_match git add .gitignore &&
+	test_all_match git commit -m ignore-bogus-files &&
+
+	run_on_sparse mkdir folder1 &&
+	run_on_all touch folder1/bogus &&
+
+	test_all_match git status --porcelain=v2 &&
+	test_all_match git clean -f &&
+	test_all_match git status --porcelain=v2 &&
+
+	test_all_match git clean -xf &&
+	test_all_match git status --porcelain=v2 &&
+
+	test_all_match git clean -xdf &&
+	test_all_match git status --porcelain=v2 &&
+
+	test_path_is_dir sparse-checkout/folder1
+'
+
+test_done

-- 
gitgitgadget

On Wed, Jan 20, 2021 at 8:54 AM Derrick Stolee via GitGitGadget
[off-list ref] wrote:

All mechanical changes so far; these look obviously correct.

        if (i)
                return i;
+
+       if (!istate->cache_tree)
+               istate->cache_tree = cache_tree();

This is the only substantive change.  It seems fairly innocuous, but
it makes me wonder the reasoning...I don't know/remember enough about
cache_tree handling to know when this would or wouldn't have already
been allocated.  It seems that this would have had to segfault below
if istate->cache_tree were ever NULL, and I don't see you mentioning
any bug you are fixing, so I presume this means you are going to be
adding new codepaths somewhere that cause this function to be reached
under different circumstances than previously had been and you need it
to be more safe for those.  Is that correct?  Or is it just an
abundance of caution thing that you're adding?  If the latter, any
reason you chose to allocate one rather than assume it's a violation
of design invariants and BUG() instead?  (Perhaps the commit message
could add a sentence about the rationale for the extra safety?)

+
        trace_performance_enter();
        trace2_region_enter("cache_tree", "update", the_repository);
-       i = update_one(it, cache, entries, "", 0, &skip, flags);
+       i = update_one(istate->cache_tree, istate->cache, istate->cache_nr,
+                      "", 0, &skip, flags);

Another mechanical update; looks good.

        trace2_region_leave("cache_tree", "update", the_repository);
        trace_performance_leave("cache_tree_update");
        if (i < 0)
--
gitgitgadget

On Wed, Jan 20, 2021 at 8:54 AM Derrick Stolee via GitGitGadget
[off-list ref] wrote:

Simple, straight-forward patch for exposing the function outside the
file scope; looks good.

On Wed, Jan 20, 2021 at 8:54 AM Derrick Stolee via GitGitGadget
[off-list ref] wrote:

From: Derrick Stolee <redacted>

The index has an fsmonitor_dirty bitmap that records which index entries
are "dirty" based on the response from the FSMonitor. If this bitmap
ever grows larger than the index, then there was an error in how it was
constructed, and it was probably a developer's bug.

There are several BUG() statements that are very similar, so replace
these uses with a simpler assert_index_minimum(). Since there is one
caller that uses a custom 'pos' value instead of the bit_size member, we
cannot simplify it too much. However, the error string is identical in
each, so this simplifies things.

The end result is that the code is simpler to read while also preserving
these assertions for developers in the FSMonitor space.

Indeed, looking through the patch, the end result is simpler to read.
Nice cleanup.

On Wed, Jan 20, 2021 at 8:54 AM Derrick Stolee via GitGitGadget
[off-list ref] wrote:

From: Derrick Stolee <redacted>

It will be helpful to add behavior to index opertations that might

s/opertations/operations/

Since we have repo->index and we have index->repo, which are intended
to be circular...what if they aren't?  Do we want or need to add
assertions anywhere that repo == repo->index->repo or that index ==
index->repo->index ?

My initial implementations of --remerge-diff[1] played around with
creating a second repo, with a different primary object store but
everything else the same.  The index for the two repository objects
was thus the same, and thus clearly would have violated this assumed
invariant for one of the two repos.  I discarded that initial
implementation (which I didn't quite have working) because I
discovered tmp-objdir.h and was able to add some
tmp_objdir_make_primary() and tmp_objdir_remove_as_primary() functions
that merely altered the existing repo's primary object store, but I'm
curious if there might be other cases of folks doing stuff that might
have weird failures with this new invariant.

It's entirely possible that --remerge-diff was just so different, and
I was so unfamiliar with repo objects (and still kind of am) that I
was just doing weird stuff no one has done before, so perhaps no
additional checks are needed -- I'm just throwing my gut question out
there as food for thought.



[1] I have not yet submitted `--remerge-diff` to the list; you haven't
missed anything.  I'm waiting for merge-ort to be submitted, reviewed,
and merged first.  It's the remerge-diff branch in my fork on GitHub
if anyone is curious, though.

On Wed, Jan 20, 2021 at 8:54 AM Derrick Stolee via GitGitGadget
[off-list ref] wrote:

Yaay for trace2.  :-)

Total sidenote: Are we ever going to get rid of the trace1 stuff?  I'm
sure this was discussed somewhere, but I don't know where.

On Wed, Jan 20, 2021 at 8:54 AM Derrick Stolee via GitGitGadget
[off-list ref] wrote:

Looks straightforward and well motivated to me.

But the cherry on top that really sells this patch is that more lines
of dir.c will blame to someone besides me.  Win-win!

On Wed, Jan 20, 2021 at 8:54 AM Derrick Stolee via GitGitGadget
[off-list ref] wrote:

The setting back to NULL made me curious; we don't want this
information to remain available later?  Is it only going to be used
for the updating of the working directory?

I dug a bit into the callers, and didn't find the answer to my
question...but I did notice that modify_pattern_list() will correctly
free the patterns after write_patterns_and_update() via calling
clear_pattern_list(&pl), but sparse_checkout_init() appears to leak
the patterns it allocates.  That's a separate issue from this patch,
but do you want to fix that up while working in this area (so I avoid
stepping on your toes with all your other patches)?

Slightly annoying that the other functions are (argc, argv, pl) and
this one is (pl, argc, argv).  But again, that's outside the scope of
this patch and might not be worth the churn to fix.

Otherwise, this patch looks good to me.

On Wed, Jan 20, 2021 at 8:54 AM Derrick Stolee via GitGitGadget
[off-list ref] wrote:

From: Derrick Stolee <redacted>

Most test cases can verify Git's behavior using input/output
expectations or changes to the .git directory. However, sometimes we
want to check that Git did or did not run a certain section of code.
This is particularly important for performance-only features that we
want to ensure have been enabled in certain cases.

Add a new 'test_region' function that checks if a trace2 region was
entered and left in a given trace2 event log.

Ooh, others do this too?  Sounds like a helpful function to add, but
just checking for entered and left means that...

There is one existing test (t0500-progress-display.sh) that performs
this check already, so use the helper function instead. More uses will
be added in a later change.

t6423-merge-rename-directories.sh also greps for region_enter lines, but
it verifies the number of such lines, which is not the same as an
existence check.

...yeah, won't cover the case that I added.  That's fine, since it
appears to be a one-off for now and we don't know of any other cases,
current or planned, that want to do something like that yet.

Sidenote: Hmm...about 40% of my region labels in merge-ort.c and 90%
in diffcore-rename.c have spaces in them.  This function could still
be used, but I'm curious if I should change the labels (but then
again, they are testing logical regions rather than individual
functions, and the spaces instead of underscores kind of convey
that...)

This patch looks good to me.

On 1/20/2021 12:21 PM, Elijah Newren wrote:

On Wed, Jan 20, 2021 at 8:54 AM Derrick Stolee via GitGitGadget
[off-list ref] wrote:...

quoted

+
+       if (!istate->cache_tree)
+               istate->cache_tree = cache_tree();

This is the only substantive change.  It seems fairly innocuous, but
it makes me wonder the reasoning...I don't know/remember enough about
cache_tree handling to know when this would or wouldn't have already
been allocated.  It seems that this would have had to segfault below
if istate->cache_tree were ever NULL, and I don't see you mentioning
any bug you are fixing, so I presume this means you are going to be
adding new codepaths somewhere that cause this function to be reached
under different circumstances than previously had been and you need it
to be more safe for those.  Is that correct?  Or is it just an
abundance of caution thing that you're adding?  If the latter, any
reason you chose to allocate one rather than assume it's a violation
of design invariants and BUG() instead?  (Perhaps the commit message
could add a sentence about the rationale for the extra safety?)

It's something I need in the future when I use the cache_tree_update()
in more places. I think I call it two times, and either I need to
initialize the cache_tree member outside of both, or just make it a
feature of the method that it will re-initialize the cache-tree.

Note: the implementation treats an initialized, but empty cache-tree
as "invalid" so update_one() correctly populates the full tree.

Thanks,
-Stolee

On 1/20/2021 12:46 PM, Elijah Newren wrote:

On Wed, Jan 20, 2021 at 8:54 AM Derrick Stolee via GitGitGadget
[off-list ref] wrote:

quoted

From: Derrick Stolee <redacted>

It will be helpful to add behavior to index opertations that might

s/opertations/operations/

Thanks.

quoted

trigger an object lookup. Since each index belongs to a specific
repository, add a 'repo' pointer to struct index_state that allows
access to this repository.

This will prevent future changes from needing to pass an additional
'struct repository *repo' parameter and instead rely only on the 'struct
index_state *istate' parameter.

Signed-off-by: Derrick Stolee <redacted>
---
 cache.h      | 1 +
 repository.c | 4 ++++
 2 files changed, 5 insertions(+)

diff --git a/cache.h b/cache.h
index 71097657489..f9c7a603841 100644
--- a/cache.h
+++ b/cache.h

@@ -328,6 +328,7 @@ struct index_state {
        struct ewah_bitmap *fsmonitor_dirty;
        struct mem_pool *ce_mem_pool;
        struct progress *progress;
+       struct repository *repo;
 };

 /* Name hashing */

diff --git a/repository.c b/repository.c
index a4174ddb062..67a4c1da2d9 100644
--- a/repository.c
+++ b/repository.c

@@ -264,6 +264,10 @@ int repo_read_index(struct repository *repo)
        if (!repo->index)
                repo->index = xcalloc(1, sizeof(*repo->index));

+       /* Complete the double-reference */
+       if (!repo->index->repo)
+               repo->index->repo = repo;
+
        return read_index_from(repo->index, repo->index_file, repo->gitdir);
 }

--

gitgitgadget

Since we have repo->index and we have index->repo, which are intended
to be circular...what if they aren't?  Do we want or need to add
assertions anywhere that repo == repo->index->repo or that index ==
index->repo->index ?

Here, we are pairing them together and the loop is complete. I don't
view that as a permanent thing. This only initializes istate->repo
when we are parsing an index from a file, but not when we create one
in memory.

I imagine it will be likely in some cases to have multiple index_state
instances for a single repository. However, having the pointer "this
index belongs to this repository" seems helpful (to me).

My initial implementations of --remerge-diff[1] played around with
creating a second repo, with a different primary object store but
everything else the same.  The index for the two repository objects
was thus the same, and thus clearly would have violated this assumed
invariant for one of the two repos.  I discarded that initial
implementation (which I didn't quite have working) because I
discovered tmp-objdir.h and was able to add some
tmp_objdir_make_primary() and tmp_objdir_remove_as_primary() functions
that merely altered the existing repo's primary object store, but I'm
curious if there might be other cases of folks doing stuff that might
have weird failures with this new invariant.

This is an interesting concept, and definitely violates my expectations
that an index belongs to only one repository. I'd need to know more
about why this was a good design decision before being convinced that
the relationship should not be many-to-one (index-to-repo).

It's entirely possible that --remerge-diff was just so different, and
I was so unfamiliar with repo objects (and still kind of am) that I
was just doing weird stuff no one has done before, so perhaps no
additional checks are needed -- I'm just throwing my gut question out
there as food for thought.

[1] I have not yet submitted `--remerge-diff` to the list; you haven't
missed anything.  I'm waiting for merge-ort to be submitted, reviewed,
and merged first.  It's the remerge-diff branch in my fork on GitHub
if anyone is curious, though.

 
I'm interested in what others might say about this idea. I'd be able
to do most of what I want to do without this patch, but it just gets
a lot messier. (istate->repo is used in the very next patch in a way
that would be less clean without it.)

Thanks,
-Stolee

On 1/20/2021 1:03 PM, Elijah Newren wrote:

On Wed, Jan 20, 2021 at 8:54 AM Derrick Stolee via GitGitGadget
[off-list ref] wrote:

quoted

From: Derrick Stolee <redacted>

As we modify the sparse-checkout definition, we perform index operations
on a pattern_list that only exists in-memory. This allows easy backing
out in case the index update fails.

However, if the index write itself cares about the sparse-checkout
pattern set, we need access to that in-memory copy. Place a pointer to
a 'struct pattern_list' in the index so we can access this on-demand.

Signed-off-by: Derrick Stolee <redacted>
---
 builtin/sparse-checkout.c | 17 ++++++++++-------
 cache.h                   |  2 ++
 2 files changed, 12 insertions(+), 7 deletions(-)

diff --git a/builtin/sparse-checkout.c b/builtin/sparse-checkout.c
index 2306a9ad98e..e00b82af727 100644
--- a/builtin/sparse-checkout.c
+++ b/builtin/sparse-checkout.c

@@ -110,6 +110,8 @@ static int update_working_directory(struct pattern_list *pl)
        if (is_index_unborn(r->index))
                return UPDATE_SPARSITY_SUCCESS;

+       r->index->sparse_checkout_patterns = pl;
+
        memset(&o, 0, sizeof(o));
        o.verbose_update = isatty(2);
        o.update = 1;

@@ -138,6 +140,7 @@ static int update_working_directory(struct pattern_list *pl)
        else
                rollback_lock_file(&lock_file);

+       r->index->sparse_checkout_patterns = NULL;
        return result;

The setting back to NULL made me curious; we don't want this
information to remain available later?  Is it only going to be used
for the updating of the working directory?

I dug a bit into the callers, and didn't find the answer to my
question...but I did notice that modify_pattern_list() will correctly
free the patterns after write_patterns_and_update() via calling
clear_pattern_list(&pl), but sparse_checkout_init() appears to leak
the patterns it allocates.  That's a separate issue from this patch,
but do you want to fix that up while working in this area (so I avoid
stepping on your toes with all your other patches)?

The thing that caught me here is that update_working_directory() uses
an in-memory pattern_list that hasn't been committed to the
sparse-checkout file yet. This means we need to (temporarily) point
to this pattern_list.

Perhaps this patch is premature, since nothing actually _uses_
sparse_checkout_patterns yet. When we do add such a use, it will
initialize a NULL value with the patterns in the sparse-checkout
file. In that case, we definitely want to inject our in-memory
patterns instead.

Thanks,
-Stolee

On 1/20/2021 1:20 PM, Elijah Newren wrote:

On Wed, Jan 20, 2021 at 8:54 AM Derrick Stolee via GitGitGadget
[off-list ref] wrote:

...

quoted

        # t0212/parse_events.perl intentionally omits regions and data.
-       grep -e "region_enter" -e "\"category\":\"progress\"" trace.event &&
-       grep -e "region_leave" -e "\"category\":\"progress\"" trace.event &&
+       test_region category progress trace.event &&

Sidenote: Hmm...about 40% of my region labels in merge-ort.c and 90%
in diffcore-rename.c have spaces in them.  This function could still
be used, but I'm curious if I should change the labels (but then
again, they are testing logical regions rather than individual
functions, and the spaces instead of underscores kind of convey
that...)

You should be able to use

	test_region "category with spaces" "progress with spaces" trace

but if not, then the test_region helper could be improved to match.

I do think that it's better to avoid spaces in these identifiers.

Thanks,
-Stolee

On Wed, Jan 20, 2021 at 8:54 AM Derrick Stolee via GitGitGadget
[off-list ref] wrote:

From: Derrick Stolee <redacted>

These also document some behaviors that differ from a full checkout, and
possibly in a way that is not intended.

I'm in favor.  I should turn some of my noted weird behaviors from [1]
into testcases as well.

[1] https://lore.kernel.org/git/CABPp-BGJ_Nvi5TmgriD9Bh6eNXE2EDq2f8e8QKXAeYG3BxZafA@mail.gmail.com/