patch in sprase-checkout from libgit2#6394

Author: jochenhz

include/git2.h

@@ -60,6 +60,7 @@
 #include "git2/revparse.h"
 #include "git2/revwalk.h"
 #include "git2/signature.h"
+#include "git2/sparse.h"
 #include "git2/stash.h"
 #include "git2/status.h"
 #include "git2/submodule.h"

include/git2/checkout.h

@@ -186,6 +186,12 @@ typedef enum {
 	/** Include common ancestor data in zdiff3 format for conflicts */
 	GIT_CHECKOUT_CONFLICT_STYLE_ZDIFF3 = (1u << 25),
 
+	/**
+	 *  Remove files that are excluded by the sparse-checkout ruleset.
+	 * 	Does nothing when GIT_CHECKOUT_SAFE is set.
+	 */
+	GIT_CHECKOUT_REMOVE_SPARSE_FILES = (1u << 26),
+
 	/**
 	 * THE FOLLOWING OPTIONS ARE NOT YET IMPLEMENTED
 	 */

include/git2/diff.h

@@ -445,6 +445,11 @@ typedef struct {
 	 * Defaults to "b".
 	 */
 	const char *new_prefix;
+
+	/** Skip files in the diff that are excluded by the `sparse-checkout` file.
+	 * Set to 1 to skip sparse files, 0 otherwise
+	 */
+	int skip_sparse_files;
 } git_diff_options;
 
 /* The current version of the diff options structure */

include/git2/sparse.h

@@ -0,0 +1,114 @@
+/*
+ * Copyright (C) the libgit2 contributors. All rights reserved.
+ *
+ * This file is part of libgit2, distributed under the GNU GPL v2 with
+ * a Linking Exception. For full terms see the included COPYING file.
+ */
+#ifndef INCLUDE_git_sparse_h__
+#define INCLUDE_git_sparse_h__
+
+#include "common.h"
+#include "types.h"
+
+GIT_BEGIN_DECL
+
+typedef struct {
+	unsigned int version; /**< The version */
+
+    /**
+     * Set to zero (false) to consider sparse-checkout patterns as
+     * full patterns, or non-zero for cone patterns.
+     */
+    /* int cone; */
+} git_sparse_checkout_init_options;
+
+#define GIT_SPARSE_CHECKOUT_INIT_OPTIONS_VERSION 1
+#define GIT_SPARSE_CHECKOUT_INIT_OPTIONS_INIT {GIT_SPARSE_CHECKOUT_INIT_OPTIONS_VERSION};
+
+/**
+ * Enable the core.sparseCheckout setting. If the sparse-checkout
+ * file does not exist, then populate it with patterns that match
+ * every file in the root directory and no other directories.
+ *
+ * @param repo Repository where to find the sparse-checkout file
+ * @param opts The `git_sparse_checkout_init_options` when
+ *      initializing the sparse-checkout file
+ * @return 0 or an error code
+ */
+GIT_EXTERN(int) git_sparse_checkout_init(
+		git_repository *repo,
+		git_sparse_checkout_init_options *opts);
+
+/**
+ * Fill a list with all the patterns in the sparse-checkout file
+ *
+ * @param patterns Pointer to a git_strarray structure where
+ *		the patterns will be stored
+ * @param repo Repository where to find the sparse-checkout file
+ * @return 0 or an error code
+ */
+GIT_EXTERN(int) git_sparse_checkout_list(
+        git_strarray *patterns,
+        git_repository *repo);
+
+/**
+ * Write a set of patterns to the sparse-checkout file.
+ * Update the working directory to match the new patterns.
+ * Enable the core.sparseCheckout config setting if it is not
+ * already enabled.
+ *
+ * @param repo Repository where to find the sparse-checkout file
+ * @param patterns Pointer to a git_strarray structure where
+ *      the patterns to set can be found
+ * @return 0 or an error code
+ */
+GIT_EXTERN(int) git_sparse_checkout_set(
+        git_repository *repo,
+        git_strarray *patterns);
+
+/**
+ * Update the sparse-checkout file to include additional patterns.
+ *
+ * @param repo Repository where to find the sparse-checkout file
+ * @param patterns Pointer to a git_strarray structure where
+ *      the patterns to set can be found
+ * @return 0 or an error code
+ */
+GIT_EXTERN(int) git_sparse_checkout_add(
+        git_repository *repo,
+        git_strarray *patterns);
+
+GIT_EXTERN(int) git_sparse_checkout_reapply(git_repository *repo);
+
+/**
+ * Disable the core.sparseCheckout config setting, and restore the
+ * working directory to include all files. Leaves the sparse-checkout
+ * file intact so a later git sparse-checkout init command may return
+ * the working directory to the same state.
+ *
+ * @param repo Repository where to find the sparse-checkout file
+ * @return 0 or an error code
+ */
+GIT_EXTERN(int) git_sparse_checkout_disable(git_repository *repo);
+
+/**
+ * Test if the sparse-checkout rules apply to a given path.
+ *
+ * This function checks the sparse-checkout rules to see if they would apply
+ * to the given path. This indicates if the path would be included on checkout.
+ *
+ * @param checkout boolean returning 1 if the sparse-checkout rules apply
+ * 		(the file will be checked out), 0 if they do not
+ * @param repo Repository where to find the sparse-checkout file
+ * @param path the path to check sparse-checkout rules for, relative to the repo's workdir.
+ * @return 0 if sparse-checkout rules could be processed for the path
+ * 		(regardless of whether it exists or not), or an error < 0 if they could not.
+ */
+GIT_EXTERN(int) git_sparse_check_path(
+	int *checkout,
+	git_repository *repo,
+	const char *path);
+
+GIT_END_DECL
+
+#endif

src/libgit2/attr_file.c

@@ -1025,3 +1025,152 @@ void git_attr_session__free(git_attr_session *session)
 
 	memset(session, 0, sizeof(git_attr_session));
 }
+
+
+/**
+ * A negative ignore pattern can negate a positive one without
+ * wildcards if it is a basename only and equals the basename of
+ * the positive pattern. Thus
+ *
+ * foo/bar
+ * !bar
+ *
+ * would result in foo/bar being unignored again while
+ *
+ * moo/foo/bar
+ * !foo/bar
+ *
+ * would do nothing. The reverse also holds true: a positive
+ * basename pattern can be negated by unignoring the basename in
+ * subdirectories. Thus
+ *
+ * bar
+ * !foo/bar
+ *
+ * would result in foo/bar being unignored again. As with the
+ * first case,
+ *
+ * foo/bar
+ * !moo/foo/bar
+ *
+ * would do nothing, again.
+ */
+static int does_negate_pattern(git_attr_fnmatch *rule, git_attr_fnmatch *neg)
+{
+	int (*cmp)(const char *, const char *, size_t);
+	git_attr_fnmatch *longer, *shorter;
+	char *p;
+
+	if ((rule->flags & GIT_ATTR_FNMATCH_NEGATIVE) != 0
+		|| (neg->flags & GIT_ATTR_FNMATCH_NEGATIVE) == 0)
+		return false;
+
+	if (neg->flags & GIT_ATTR_FNMATCH_ICASE)
+		cmp = git__strncasecmp;
+	else
+		cmp = git__strncmp;
+
+	/* If lengths match we need to have an exact match */
+	if (rule->length == neg->length) {
+		return cmp(rule->pattern, neg->pattern, rule->length) == 0;
+	} else if (rule->length < neg->length) {
+		shorter = rule;
+		longer = neg;
+	} else {
+		shorter = neg;
+		longer = rule;
+	}
+
+	/* Otherwise, we need to check if the shorter
+	 * rule is a basename only (that is, it contains
+	 * no path separator) and, if so, if it
+	 * matches the tail of the longer rule */
+	p = longer->pattern + longer->length - shorter->length;
+
+	if (p[-1] != '/')
+		return false;
+	if (memchr(shorter->pattern, '/', shorter->length) != NULL)
+		return false;
+
+	return cmp(p, shorter->pattern, shorter->length) == 0;
+}
+
+/**
+ * A negative ignore can only unignore a file which is given explicitly before, thus
+ *
+ *    foo
+ *    !foo/bar
+ *
+ * does not unignore 'foo/bar' as it's not in the list. However
+ *
+ *    foo/<star>
+ *    !foo/bar
+ *
+ * does unignore 'foo/bar', as it is contained within the 'foo/<star>' rule.
+ */
+int git_attr__does_negate_rule(int *out, git_vector *rules, git_attr_fnmatch *match)
+{
+	int error = 0, wildmatch_flags, effective_flags;
+	size_t i;
+	git_attr_fnmatch *rule;
+	char *path;
+	git_str buf = GIT_STR_INIT;
+
+	*out = 0;
+
+	wildmatch_flags = WM_PATHNAME;
+	if (match->flags & GIT_ATTR_FNMATCH_ICASE)
+		wildmatch_flags |= WM_CASEFOLD;
+
+	/* path of the file relative to the workdir, so we match the rules in subdirs */
+	if (match->containing_dir) {
+		git_str_puts(&buf, match->containing_dir);
+	}
+	if (git_str_puts(&buf, match->pattern) < 0)
+		return -1;
+
+	path = git_str_detach(&buf);
+
+	git_vector_foreach(rules, i, rule) {
+		if (!(rule->flags & GIT_ATTR_FNMATCH_HASWILD)) {
+			if (does_negate_pattern(rule, match)) {
+				error = 0;
+				*out = 1;
+				goto out;
+			}
+			else
+				continue;
+		}
+
+		git_str_clear(&buf);
+		if (rule->containing_dir)
+			git_str_puts(&buf, rule->containing_dir);
+		git_str_puts(&buf, rule->pattern);
+
+		if (git_str_oom(&buf))
+			goto out;
+
+		/*
+		 * if rule isn't for full path we match without PATHNAME flag
+		 * as lines like *.txt should match something like dir/test.txt
+		 * requiring * to also match /
+		 */
+		effective_flags = wildmatch_flags;
+		if (!(rule->flags & GIT_ATTR_FNMATCH_FULLPATH))
+			effective_flags &= ~WM_PATHNAME;
+
+		/* if we found a match, we want to keep this rule */
+		if ((wildmatch(git_str_cstr(&buf), path, effective_flags)) == WM_MATCH) {
+			*out = 1;
+			error = 0;
+			goto out;
+		}
+	}
+
+	error = 0;
+
+	out:
+	git__free(path);
+	git_str_dispose(&buf);
+	return error;
+}

src/libgit2/attr_file.h

@@ -238,4 +238,6 @@ extern int git_attr_assignment__parse(
 	git_vector *assigns,
 	const char **scan);
 
+extern int git_attr__does_negate_rule(int *out, git_vector *rules, git_attr_fnmatch *match);
+
 #endif

src/libgit2/checkout.c

@@ -390,8 +390,15 @@ static int checkout_action_wd_only(
 
 		if (wd->mode != GIT_FILEMODE_TREE) {
 			if (!error) { /* found by git_index__find_pos call */
-				notify = GIT_CHECKOUT_NOTIFY_DIRTY;
-				remove = ((data->strategy & GIT_CHECKOUT_FORCE) != 0);
+
+				/* Sparse checkout will set the SKIP_WORKTREE bit if a file should be skipped */
+				const git_index_entry *e = git_index_get_byindex(data->index, pos);
+				if (e == NULL ||
+						(e->flags_extended & GIT_INDEX_ENTRY_SKIP_WORKTREE) == 0 ||
+						(data->strategy & GIT_CHECKOUT_REMOVE_SPARSE_FILES) != 0) {
+					notify = GIT_CHECKOUT_NOTIFY_DIRTY;
+					remove = ((data->strategy & GIT_CHECKOUT_FORCE) != 0);
+				}
 			} else if (error != GIT_ENOTFOUND)
 				return error;
 			else
@@ -2571,6 +2578,8 @@ int git_checkout_iterator(
 		GIT_DIFF_INCLUDE_TYPECHANGE_TREES |
 		GIT_DIFF_SKIP_BINARY_CHECK |
 		GIT_DIFF_INCLUDE_CASECHANGE;
+	diff_opts.skip_sparse_files = 1;
+
 	if (data.opts.checkout_strategy & GIT_CHECKOUT_DISABLE_PATHSPEC_MATCH)
 		diff_opts.flags |= GIT_DIFF_DISABLE_PATHSPEC_MATCH;
 	if (data.opts.paths.count > 0)

src/libgit2/config_cache.c

@@ -87,6 +87,7 @@ static struct map_data _configmaps[] = {
 	{"core.protectntfs", NULL, 0, GIT_PROTECTNTFS_DEFAULT },
 	{"core.fsyncobjectfiles", NULL, 0, GIT_FSYNCOBJECTFILES_DEFAULT },
 	{"core.longpaths", NULL, 0, GIT_LONGPATHS_DEFAULT },
+	{"core.sparsecheckout", NULL, 0, GIT_SPARSECHECKOUT_DEFAULT },
 };
 
 int git_config__configmap_lookup(int *out, git_config *config, git_configmap_item item)

src/libgit2/diff_generate.c

@@ -799,6 +799,12 @@ static int maybe_modified(
 	if (!diff_pathspec_match(&matched_pathspec, diff, oitem))
 		return 0;
 
+	if (diff->base.opts.skip_sparse_files &&
+			git_iterator_current_skip_checkout(info->new_iter))
+		return 0;
+	
+	memset(&noid, 0, sizeof(noid));
+
 	/* on platforms with no symlinks, preserve mode of existing symlinks */
 	if (S_ISLNK(omode) && S_ISREG(nmode) && new_is_workdir &&
 		!(diff->diffcaps & GIT_DIFFCAPS_HAS_SYMLINKS))
@@ -1028,6 +1034,11 @@ static int handle_unmatched_new_item(
 	const git_index_entry *nitem = info->nitem;
 	git_delta_t delta_type = GIT_DELTA_UNTRACKED;
 	bool contains_oitem;
+	
+	/* check if this item should be skipped due to sparse checkout */
+	if (diff->base.opts.skip_sparse_files &&
+			git_iterator_current_skip_checkout(info->new_iter))
+		return iterator_advance(&info->nitem, info->new_iter);
 
 	/* check if this is a prefix of the other side */
 	contains_oitem = entry_is_prefixed(diff, info->oitem, nitem);
@@ -1188,7 +1199,12 @@ static int handle_unmatched_old_item(
 	if (git_index_entry_is_conflict(info->oitem))
 		delta_type = GIT_DELTA_CONFLICTED;
 
-	if ((error = diff_delta__from_one(diff, delta_type, info->oitem, NULL)) < 0)
+	if ((diff->base.opts.skip_sparse_files &&
+			 git_iterator_current_skip_checkout(info->new_iter)) ||
+			(info->oitem->flags_extended & GIT_INDEX_ENTRY_SKIP_WORKTREE) != 0)
+		delta_type = GIT_DELTA_UNMODIFIED;
+	
+	else if ((error = diff_delta__from_one(diff, delta_type, info->oitem, NULL)) < 0)
 		return error;
 
 	/* if we are generating TYPECHANGE records then check for that