Merge branch 'sparse-index-stuff'

This is random stuff that probably all got upstream in the meantime.

Documentation/config/index.txt

@@ -1,3 +1,9 @@
+index.deleteSparseDirectories::
+	When enabled, the cone mode sparse-checkout feature will delete
+	directories that are outside of the sparse-checkout cone, unless
+	such a directory contains an untracked, non-ignored file. Defaults
+	to true.
+
 index.recordEndOfIndexEntries::
 	Specifies whether the index file should include an "End Of Index
 	Entry" section. This reduces index load time on multiprocessor

builtin/add.c

@@ -47,6 +47,7 @@ static int chmod_pathspec(struct pathspec *pathspec, char flip, int show_only)
 		int err;
 
 		if (!include_sparse &&
+		    !core_virtualfilesystem &&
 		    (ce_skip_worktree(ce) ||
 		     !path_in_sparse_checkout(ce->name, &the_index)))
 			continue;
@@ -97,7 +98,8 @@ static void update_callback(struct diff_queue_struct *q,
 		struct diff_filepair *p = q->queue[i];
 		const char *path = p->one->path;
 
-		if (!include_sparse && !path_in_sparse_checkout(path, &the_index))
+		if (!include_sparse && !core_virtualfilesystem &&
+		    !path_in_sparse_checkout(path, &the_index))
 			continue;
 
 		switch (fix_unmerged_status(p, data)) {
@@ -215,8 +217,9 @@ static int refresh(int verbose, const struct pathspec *pathspec)
 		if (!seen[i]) {
 			const char *path = pathspec->items[i].original;
 
-			if (matches_skip_worktree(pathspec, i, &skip_worktree_seen) ||
-			    !path_in_sparse_checkout(path, &the_index)) {
+			if (!core_virtualfilesystem &&
+			    (matches_skip_worktree(pathspec, i, &skip_worktree_seen) ||
+			     !path_in_sparse_checkout(path, &the_index))) {
 				string_list_append(&only_match_skip_worktree,
 						   pathspec->items[i].original);
 			} else {
@@ -226,7 +229,11 @@ static int refresh(int verbose, const struct pathspec *pathspec)
 		}
 	}
 
-	if (only_match_skip_worktree.nr) {
+	/*
+	 * When using a virtual filesystem, we might re-add a path
+	 * that is currently virtual and we want that to succeed.
+	 */
+	if (!core_virtualfilesystem && only_match_skip_worktree.nr) {
 		advise_on_updating_sparse_paths(&only_match_skip_worktree);
 		ret = 1;
 	}
@@ -644,7 +651,11 @@ int cmd_add(int argc, const char **argv, const char *prefix)
 			if (seen[i])
 				continue;
 
-			if (!include_sparse &&
+			/*
+			 * When using a virtual filesystem, we might re-add a path
+			 * that is currently virtual and we want that to succeed.
+			 */
+			if (!include_sparse && !core_virtualfilesystem &&
 			    matches_skip_worktree(&pathspec, i, &skip_worktree_seen)) {
 				string_list_append(&only_match_skip_worktree,
 						   pathspec.items[i].original);
@@ -668,7 +679,6 @@ int cmd_add(int argc, const char **argv, const char *prefix)
 			}
 		}
 
-
 		if (only_match_skip_worktree.nr) {
 			advise_on_updating_sparse_paths(&only_match_skip_worktree);
 			exit_status = 1;

builtin/gc.c

@@ -1285,6 +1285,8 @@ static int maintenance_run_tasks(struct maintenance_run_opts *opts)
 	char *lock_path = xstrfmt("%s/maintenance", r->objects->odb->path);
 
 	if (hold_lock_file_for_update(&lk, lock_path, LOCK_NO_DEREF) < 0) {
+		struct stat st;
+		struct strbuf lock_dot_lock = STRBUF_INIT;
 		/*
 		 * Another maintenance command is running.
 		 *
@@ -1295,6 +1297,25 @@ static int maintenance_run_tasks(struct maintenance_run_opts *opts)
 		if (!opts->auto_flag && !opts->quiet)
 			warning(_("lock file '%s' exists, skipping maintenance"),
 				lock_path);
+
+		/*
+		 * Check timestamp on .lock file to see if we should
+		 * delete it to recover from a fail state.
+		 */
+		strbuf_addstr(&lock_dot_lock, lock_path);
+		strbuf_addstr(&lock_dot_lock, ".lock");
+		if (lstat(lock_dot_lock.buf, &st))
+			warning_errno(_("unable to stat '%s'"), lock_dot_lock.buf);
+		else {
+			if (st.st_mtime < time(NULL) - (6 * 60 * 60)) {
+				if (unlink(lock_dot_lock.buf))
+					warning_errno(_("unable to delete stale lock file"));
+				else
+					warning(_("deleted stale lock file"));
+			}
+		}
+
+		strbuf_release(&lock_dot_lock);
 		free(lock_path);
 		return 0;
 	}

builtin/reset.c

@@ -28,6 +28,8 @@
 #include "dir.h"
 #include "strbuf.h"
 #include "quote.h"
+#include "dir.h"
+#include "entry.h"
 
 #define REFRESH_INDEX_DELAY_WARNING_IN_MS (2 * 1000)
 
@@ -141,9 +143,47 @@ static void update_index_from_diff(struct diff_queue_struct *q,
 
 	for (i = 0; i < q->nr; i++) {
 		int pos;
+		int respect_skip_worktree = 1;
 		struct diff_filespec *one = q->queue[i]->one;
+		struct diff_filespec *two = q->queue[i]->two;
 		int is_in_reset_tree = one->mode && !is_null_oid(&one->oid);
+		int is_missing = !(one->mode && !is_null_oid(&one->oid));
+		int was_missing = !two->mode && is_null_oid(&two->oid);
 		struct cache_entry *ce;
+		struct cache_entry *ceBefore;
+		struct checkout state = CHECKOUT_INIT;
+
+		/*
+		 * When using the virtual filesystem feature, the cache entries that are
+		 * added here will not have the skip-worktree bit set.
+		 *
+		 * Without this code there is data that is lost because the files that
+		 * would normally be in the working directory are not there and show as
+		 * deleted for the next status or in the case of added files just disappear.
+		 * We need to create the previous version of the files in the working
+		 * directory so that they will have the right content and the next
+		 * status call will show modified or untracked files correctly.
+		 */
+		if (core_virtualfilesystem && !file_exists(two->path))
+		{
+			pos = cache_name_pos(two->path, strlen(two->path));
+			if ((pos >= 0 && ce_skip_worktree(active_cache[pos])) &&
+			    (is_missing || !was_missing))
+			{
+				state.force = 1;
+				state.refresh_cache = 1;
+				state.istate = &the_index;
+				ceBefore = make_cache_entry(&the_index, two->mode,
+							    &two->oid, two->path,
+							    0, 0);
+				if (!ceBefore)
+					die(_("make_cache_entry failed for path '%s'"),
+						two->path);
+
+				checkout_entry(ceBefore, &state, NULL, NULL);
+				respect_skip_worktree = 0;
+			}
+		}
 
 		if (!is_in_reset_tree && !intent_to_add) {
 			remove_file_from_cache(one->path);
@@ -162,8 +202,14 @@ static void update_index_from_diff(struct diff_queue_struct *q,
 		 * to properly construct the reset sparse directory.
 		 */
 		pos = cache_name_pos(one->path, strlen(one->path));
-		if ((pos >= 0 && ce_skip_worktree(active_cache[pos])) ||
-		    (pos < 0 && !path_in_sparse_checkout(one->path, &the_index)))
+
+		/*
+		 * Do not add the SKIP_WORKTREE bit back if we populated the
+		 * file on purpose in a virtual filesystem scenario.
+		 */
+		if (respect_skip_worktree &&
+		    ((pos >= 0 && ce_skip_worktree(active_cache[pos])) ||
+		     (pos < 0 && !path_in_sparse_checkout(one->path, &the_index))))
 			ce->ce_flags |= CE_SKIP_WORKTREE;
 
 		if (!ce)

builtin/rm.c

@@ -304,7 +304,7 @@ int cmd_rm(int argc, const char **argv, const char *prefix)
 	for (i = 0; i < active_nr; i++) {
 		const struct cache_entry *ce = active_cache[i];
 
-		if (!include_sparse &&
+		if (!include_sparse && !core_virtualfilesystem &&
 		    (ce_skip_worktree(ce) ||
 		     !path_in_sparse_checkout(ce->name, &the_index)))
 			continue;
@@ -341,7 +341,11 @@ int cmd_rm(int argc, const char **argv, const char *prefix)
 				    *original ? original : ".");
 		}
 
-		if (only_match_skip_worktree.nr) {
+		/*
+		 * When using a virtual filesystem, we might re-add a path
+		 * that is currently virtual and we want that to succeed.
+		 */
+		if (!core_virtualfilesystem && only_match_skip_worktree.nr) {
 			advise_on_updating_sparse_paths(&only_match_skip_worktree);
 			ret = 1;
 		}

builtin/sparse-checkout.c

@@ -106,7 +106,7 @@ static int sparse_checkout_list(int argc, const char **argv, const char *prefix)
 
 static void clean_tracked_sparse_directories(struct repository *r)
 {
-	int i, was_full = 0;
+	int i, value, was_full = 0;
 	struct strbuf path = STRBUF_INIT;
 	size_t pathlen;
 	struct string_list_item *item;
@@ -122,6 +122,13 @@ static void clean_tracked_sparse_directories(struct repository *r)
 	    !r->index->sparse_checkout_patterns->use_cone_patterns)
 		return;
 
+	/*
+	 * Users can disable this behavior.
+	 */
+	if (!repo_config_get_bool(r, "index.deletesparsedirectories", &value) &&
+	    !value)
+		return;
+
 	/*
 	 * Use the sparse index as a data structure to assist finding
 	 * directories that are safe to delete. This conversion to a

diff.c

@@ -3922,6 +3922,13 @@ static int reuse_worktree_file(struct index_state *istate,
 	if (!FAST_WORKING_DIRECTORY && !want_file && has_object_pack(oid))
 		return 0;
 
+	/*
+	 * If this path does not match our sparse-checkout definition,
+	 * then the file will not be in the working directory.
+	 */
+	if (!path_in_sparse_checkout(name, istate))
+		return 0;
+
 	/*
 	 * Similarly, if we'd have to convert the file contents anyway, that
 	 * makes the optimization not worthwhile.

dir.c

@@ -1398,46 +1398,16 @@ static struct path_pattern *last_matching_pattern_from_list(const char *pathname
 	return res;
 }
 
-/*
- * Scan the list of patterns to determine if the ordered list
- * of patterns matches on 'pathname'.
- *
- * Return 1 for a match, 0 for not matched and -1 for undecided.
- */
-enum pattern_match_result path_matches_pattern_list(
+enum pattern_match_result path_matches_cone_mode_pattern_list(
 				const char *pathname, int pathlen,
-				const char *basename, int *dtype,
-				struct pattern_list *pl,
-				struct index_state *istate)
+				struct pattern_list *pl)
 {
-	struct path_pattern *pattern;
 	struct strbuf parent_pathname = STRBUF_INIT;
 	int result = NOT_MATCHED;
 	size_t slash_pos;
 
-	/*
-	 * The virtual file system data is used to prevent git from traversing
-	 * any part of the tree that is not in the virtual file system.  Return
-	 * 1 to exclude the entry if it is not found in the virtual file system,
-	 * else fall through to the regular excludes logic as it may further exclude.
-	 */
-	if (*dtype == DT_UNKNOWN)
-		*dtype = resolve_dtype(DT_UNKNOWN, istate, pathname, pathlen);
-	if (is_excluded_from_virtualfilesystem(pathname, pathlen, *dtype) > 0)
-		return 1;
-
-	if (!pl->use_cone_patterns) {
-		pattern = last_matching_pattern_from_list(pathname, pathlen, basename,
-							dtype, pl, istate);
-		if (pattern) {
-			if (pattern->flags & PATTERN_FLAG_NEGATIVE)
-				return NOT_MATCHED;
-			else
-				return MATCHED;
-		}
-
-		return UNDECIDED;
-	}
+	if (!pl->use_cone_patterns)
+		BUG("path_matches_cone_mode_pattern_list requires cone mode patterns");
 
 	if (pl->full_cone)
 		return MATCHED;
@@ -1490,6 +1460,46 @@ enum pattern_match_result path_matches_pattern_list(
 	return result;
 }
 
+/*
+ * Scan the list of patterns to determine if the ordered list
+ * of patterns matches on 'pathname'.
+ *
+ * Return 1 for a match, 0 for not matched and -1 for undecided.
+ */
+enum pattern_match_result path_matches_pattern_list(
+				const char *pathname, int pathlen,
+				const char *basename, int *dtype,
+				struct pattern_list *pl,
+				struct index_state *istate)
+{
+	/*
+	 * The virtual file system data is used to prevent git from traversing
+	 * any part of the tree that is not in the virtual file system.  Return
+	 * 1 to exclude the entry if it is not found in the virtual file system,
+	 * else fall through to the regular excludes logic as it may further exclude.
+	 */
+	if (*dtype == DT_UNKNOWN)
+		*dtype = resolve_dtype(DT_UNKNOWN, istate, pathname, pathlen);
+	if (is_excluded_from_virtualfilesystem(pathname, pathlen, *dtype) > 0)
+		return 1;
+
+	if (!pl->use_cone_patterns) {
+		struct path_pattern *pattern = last_matching_pattern_from_list(
+							pathname, pathlen, basename,
+							dtype, pl, istate);
+		if (pattern) {
+			if (pattern->flags & PATTERN_FLAG_NEGATIVE)
+				return NOT_MATCHED;
+			else
+				return MATCHED;
+		}
+
+		return UNDECIDED;
+	}
+
+	return path_matches_cone_mode_pattern_list(pathname, pathlen, pl);
+}
+
 int init_sparse_checkout_patterns(struct index_state *istate)
 {
 	if (!core_apply_sparse_checkout)
@@ -1515,6 +1525,13 @@ static int path_in_sparse_checkout_1(const char *path,
 	enum pattern_match_result match = UNDECIDED;
 	const char *end, *slash;
 
+	/*
+	 * When using a virtual filesystem, there aren't really patterns
+	 * to follow, but be extra careful to skip this check.
+	 */
+	if (core_virtualfilesystem)
+		return 1;
+
 	/*
 	 * We default to accepting a path if the path is empty, there are no
 	 * patterns, or the patterns are of the wrong type.

dir.h

@@ -383,6 +383,15 @@ enum pattern_match_result {
 	MATCHED_RECURSIVE = 2,
 };
 
+/*
+ * Test if a given path is contained in the given pattern list.
+ *
+ * The given pattern list _must_ use cone mode patterns.
+ */
+enum pattern_match_result path_matches_cone_mode_pattern_list(
+				const char *pathname, int pathlen,
+				struct pattern_list *pl);
+
 /*
  * Scan the list of patterns to determine if the ordered list
  * of patterns matches on 'pathname'.

repo-settings.c

@@ -84,7 +84,7 @@ void prepare_repo_settings(struct repository *r)
 	/* Boolean config or default, does not cascade (simple)  */
 	repo_cfg_bool(r, "pack.usesparse", &r->settings.pack_use_sparse, 1);
 	repo_cfg_bool(r, "core.multipackindex", &r->settings.core_multi_pack_index, 1);
-	repo_cfg_bool(r, "index.sparse", &r->settings.sparse_index, 0);
+	repo_cfg_bool(r, "index.sparse", &r->settings.sparse_index, 1);
 
 	/*
 	 * The GIT_TEST_MULTI_PACK_INDEX variable is special in that

sequencer.c

@@ -683,7 +683,7 @@ static int do_recursive_merge(struct repository *r,
 	o.branch2 = next ? next_label : "(empty tree)";
 	if (is_rebase_i(opts))
 		o.buffer_output = 2;
-	o.show_rename_progress = 1;
+	o.show_rename_progress = isatty(2);
 
 	head_tree = parse_tree_indirect(head);
 	next_tree = next ? get_commit_tree(next) : empty_tree(r);