Make diff and status perform soft index reload

This changes `git_index_read` to have two modes - a hard index
reload that always resets the index to match the on-disk data
(which was the old behavior) and a soft index reload that uses
the timestamp / file size information and only replaces the index
data if the file on disk has been modified.

This then updates the git_status code to do a soft reload unless
the new GIT_STATUS_OPT_NO_REFRESH flag is passed in.

This also changes the behavior of the git_diff functions that use
the index so that when an index is not explicitly passed in (i.e.
when the functions call git_repository_index for you), they will
also do a soft reload for you.

This intentionally breaks the file signature of git_index_read
because there has been some confusion about the behavior previously
and it seems like all existing uses of the API should probably be
examined to select the desired behavior.

5 files changed, 42 insertions, 18 deletions

diff --git a/src/checkout.c b/src/checkout.c
index 76edc6a72..94968e378 100644
--- a/src/checkout.c
+++ b/src/checkout.c
@@ -1837,7 +1837,7 @@ static int checkout_data_init(
 		} else {
 			/* otherwise, grab and reload the index */
 			if ((error = git_repository_index(&data->index, data->repo)) < 0 ||
-				(error = git_index_read(data->index)) < 0)
+				(error = git_index_read(data->index, false)) < 0)
 				goto cleanup;
 
 			/* cannot checkout if unresolved conflicts exist */
diff --git a/src/diff.c b/src/diff.c
index b1f64e6a3..d1ff04b52 100644
--- a/src/diff.c
+++ b/src/diff.c
@@ -1184,6 +1184,17 @@ int git_diff_tree_to_tree(
 	return error;
 }
 
+static int diff_load_index(git_index **index, git_repository *repo)
+{
+	int error = git_repository_index__weakptr(index, repo);
+
+	/* reload the repository index when user did not pass one in */
+	if (!error && git_index_read(*index, true) < 0)
+		giterr_clear();
+
+	return error;
+}
+
 int git_diff_tree_to_index(
 	git_diff **diff,
 	git_repository *repo,
@@ -1196,7 +1207,7 @@ int git_diff_tree_to_index(
 
 	assert(diff && repo);
 
-	if (!index && (error = git_repository_index__weakptr(&index, repo)) < 0)
+	if (!index && (error = diff_load_index(&index, repo)) < 0)
 		return error;
 
 	if (index->ignore_case) {
@@ -1239,7 +1250,7 @@ int git_diff_index_to_workdir(
 
 	assert(diff && repo);
 
-	if (!index && (error = git_repository_index__weakptr(&index, repo)) < 0)
+	if (!index && (error = diff_load_index(&index, repo)) < 0)
 		return error;
 
 	DIFF_FROM_ITERATORS(
@@ -1278,11 +1289,15 @@ int git_diff_tree_to_workdir_with_index(
 {
 	int error = 0;
 	git_diff *d1 = NULL, *d2 = NULL;
+	git_index *index = NULL;
 
 	assert(diff && repo);
 
-	if (!(error = git_diff_tree_to_index(&d1, repo, old_tree, NULL, opts)) &&
-		!(error = git_diff_index_to_workdir(&d2, repo, NULL, opts)))
+	if ((error = diff_load_index(&index, repo)) < 0)
+		return error;
+
+	if (!(error = git_diff_tree_to_index(&d1, repo, old_tree, index, opts)) &&
+		!(error = git_diff_index_to_workdir(&d2, repo, index, opts)))
 		error = git_diff_merge(d1, d2);
 
 	git_diff_free(d2);
diff --git a/src/index.c b/src/index.c
index 5cdd40aaa..19de43d29 100644
--- a/src/index.c
+++ b/src/index.c
@@ -349,7 +349,7 @@ int git_index_open(git_index **index_out, const char *index_path)
 	*index_out = index;
 	GIT_REFCOUNT_INC(index);
 
-	return (index_path != NULL) ? git_index_read(index) : 0;
+	return (index_path != NULL) ? git_index_read(index, false) : 0;
 }
 
 int git_index_new(git_index **out)
@@ -451,11 +451,11 @@ unsigned int git_index_caps(const git_index *index)
 			(index->no_symlinks ? GIT_INDEXCAP_NO_SYMLINKS : 0));
 }
 
-int git_index_read(git_index *index)
+int git_index_read(git_index *index, int only_if_changed)
 {
 	int error = 0, updated;
 	git_buf buffer = GIT_BUF_INIT;
-	git_futils_filestamp stamp = {0};
+	git_futils_filestamp stamp = index->stamp;
 
 	if (!index->index_file_path)
 		return create_index_error(-1,
@@ -464,12 +464,13 @@ int git_index_read(git_index *index)
 	index->on_disk = git_path_exists(index->index_file_path);
 
 	if (!index->on_disk) {
-		git_index_clear(index);
+		if (!only_if_changed)
+			git_index_clear(index);
 		return 0;
 	}
 
 	updated = git_futils_filestamp_check(&stamp, index->index_file_path);
-	if (updated <= 0)
+	if (updated < 0 || (only_if_changed && !updated))
 		return updated;
 
 	error = git_futils_readbuffer(&buffer, index->index_file_path);
diff --git a/src/status.c b/src/status.c
index 2b84794b5..74bccf7a1 100644
--- a/src/status.c
+++ b/src/status.c
@@ -251,12 +251,17 @@ int git_status_list_new(
 		return error;
 
 	/* if there is no HEAD, that's okay - we'll make an empty iterator */
-	if (((error = git_repository_head_tree(&head, repo)) < 0) &&
-		error != GIT_ENOTFOUND && error != GIT_EUNBORNBRANCH) {
-		git_index_free(index); /* release index */
-		return error;
+	if ((error = git_repository_head_tree(&head, repo)) < 0) {
+		if (error != GIT_ENOTFOUND && error != GIT_EUNBORNBRANCH)
+			goto done;
+		giterr_clear();
 	}
 
+	/* refresh index from disk unless prevented */
+	if ((flags & GIT_STATUS_OPT_NO_REFRESH) == 0 &&
+		git_index_read(index, true) < 0)
+		giterr_clear();
+
 	status = git_status_list_alloc(index);
 	GITERR_CHECK_ALLOC(status);
 
@@ -291,7 +296,7 @@ int git_status_list_new(
 
 	if (show != GIT_STATUS_SHOW_WORKDIR_ONLY) {
 		if ((error = git_diff_tree_to_index(
-				&status->head2idx, repo, head, NULL, &diffopt)) < 0)
+				&status->head2idx, repo, head, index, &diffopt)) < 0)
 			goto done;
 
 		if ((flags & GIT_STATUS_OPT_RENAMES_HEAD_TO_INDEX) != 0 &&
@@ -301,7 +306,7 @@ int git_status_list_new(
 
 	if (show != GIT_STATUS_SHOW_INDEX_ONLY) {
 		if ((error = git_diff_index_to_workdir(
-				&status->idx2wd, repo, NULL, &diffopt)) < 0)
+				&status->idx2wd, repo, index, &diffopt)) < 0)
 			goto done;
 
 		if ((flags & GIT_STATUS_OPT_RENAMES_INDEX_TO_WORKDIR) != 0 &&
diff --git a/src/submodule.c b/src/submodule.c
index 18d80f0a9..586494fed 100644
--- a/src/submodule.c
+++ b/src/submodule.c
@@ -1527,6 +1527,7 @@ static void submodule_get_wd_status(
 	const git_oid *wd_oid =
 		(sm->flags & GIT_SUBMODULE_STATUS__WD_OID_VALID) ? &sm->wd_oid : NULL;
 	git_tree *sm_head = NULL;
+	git_index *index = NULL;
 	git_diff_options opt = GIT_DIFF_OPTIONS_INIT;
 	git_diff *diff;
 
@@ -1558,12 +1559,14 @@ static void submodule_get_wd_status(
 	if (ign == GIT_SUBMODULE_IGNORE_NONE)
 		opt.flags |= GIT_DIFF_INCLUDE_UNTRACKED;
 
+	(void)git_repository_index__weakptr(&index, sm_repo);
+
 	/* if we don't have an unborn head, check diff with index */
 	if (git_repository_head_tree(&sm_head, sm_repo) < 0)
 		giterr_clear();
 	else {
 		/* perform head to index diff on submodule */
-		if (git_diff_tree_to_index(&diff, sm_repo, sm_head, NULL, &opt) < 0)
+		if (git_diff_tree_to_index(&diff, sm_repo, sm_head, index, &opt) < 0)
 			giterr_clear();
 		else {
 			if (git_diff_num_deltas(diff) > 0)
@@ -1576,7 +1579,7 @@ static void submodule_get_wd_status(
 	}
 
 	/* perform index-to-workdir diff on submodule */
-	if (git_diff_index_to_workdir(&diff, sm_repo, NULL, &opt) < 0)
+	if (git_diff_index_to_workdir(&diff, sm_repo, index, &opt) < 0)
 		giterr_clear();
 	else {
 		size_t untracked =