Merge remote-tracking branch 'arrbee/git-attributes' into development

Conflicts:
	tests-clay/clay_main.c

include/git2/attr.h

+/*
+ * Copyright (C) 2009-2011 the libgit2 contributors
+ *
+ * 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_attr_h__
+#define INCLUDE_git_attr_h__
+
+#include "common.h"
+#include "types.h"
+
+/**
+ * @file git2/attr.h
+ * @brief Git attribute management routines
+ * @defgroup git_attr Git attribute management routines
+ * @ingroup Git
+ * @{
+ */
+GIT_BEGIN_DECL
+
+#define GIT_ATTR_TRUE			git_attr__true
+#define GIT_ATTR_FALSE			git_attr__false
+#define GIT_ATTR_UNSPECIFIED	NULL
+
+GIT_EXTERN(const char *)git_attr__true;
+GIT_EXTERN(const char *)git_attr__false;
+
+
+/**
+ * Lookup attribute for path returning string caller must free
+ */
+GIT_EXTERN(int) git_attr_get(
+    git_repository *repo, const char *path, const char *name,
+	const char **value);
+
+/**
+ * Lookup list of attributes for path, populating array of strings
+ */
+GIT_EXTERN(int) git_attr_get_many(
+    git_repository *repo, const char *path,
+    size_t num_attr, const char **names,
+	const char **values);
+
+/**
+ * Perform an operation on each attribute of a path.
+ */
+GIT_EXTERN(int) git_attr_foreach(
+    git_repository *repo, const char *path,
+	int (*callback)(const char *name, const char *value, void *payload),
+	void *payload);
+
+/**
+ * Flush the gitattributes cache.
+ *
+ * Call this if you have reason to believe that the attributes files
+ * on disk no longer match the cached contents of memory.
+ */
+GIT_EXTERN(void) git_attr_cache_flush(
+	git_repository *repo);
+
+/**
+ * Add a macro definition.
+ *
+ * Macros will automatically be loaded from the top level .gitattributes
+ * file of the repository (plus the build-in "binary" macro).  This
+ * function allows you to add others.  For example, to add the default
+ * macro, you would call:
+ *
+ *    git_attr_add_macro(repo, "binary", "-diff -crlf");
+ */
+GIT_EXTERN(int) git_attr_add_macro(
+	git_repository *repo,
+	const char *name,
+	const char *values);
+
+/** @} */
+GIT_END_DECL
+#endif
+

src/attr.c

+#include "repository.h"
+#include "fileops.h"
+#include "config.h"
+#include <ctype.h>
+
+#define GIT_ATTR_FILE_INREPO	"info/attributes"
+#define GIT_ATTR_FILE			".gitattributes"
+#define GIT_ATTR_FILE_SYSTEM	"gitattributes"
+
+static int collect_attr_files(
+	git_repository *repo, const char *path, git_vector *files);
+
+static int attr_cache_init(git_repository *repo);
+
+
+int git_attr_get(
+    git_repository *repo, const char *pathname,
+	const char *name, const char **value)
+{
+	int error;
+	git_attr_path path;
+	git_vector files = GIT_VECTOR_INIT;
+	unsigned int i, j;
+	git_attr_file *file;
+	git_attr_name attr;
+	git_attr_rule *rule;
+
+	*value = NULL;
+
+	if ((error = git_attr_path__init(&path, pathname)) < GIT_SUCCESS ||
+		(error = collect_attr_files(repo, pathname, &files)) < GIT_SUCCESS)
+		return git__rethrow(error, "Could not get attribute for %s", pathname);
+
+	attr.name = name;
+	attr.name_hash = git_attr_file__name_hash(name);
+
+	git_vector_foreach(&files, i, file) {
+
+		git_attr_file__foreach_matching_rule(file, &path, j, rule) {
+			int pos = git_vector_bsearch(&rule->assigns, &attr);
+			git_clearerror(); /* okay if search failed */
+
+			if (pos >= 0) {
+				*value = ((git_attr_assignment *)git_vector_get(
+							  &rule->assigns, pos))->value;
+				goto found;
+			}
+		}
+	}
+
+found:
+	git_vector_free(&files);
+
+	return error;
+}
+
+
+typedef struct {
+	git_attr_name name;
+	git_attr_assignment *found;
+} attr_get_many_info;
+
+int git_attr_get_many(
+    git_repository *repo, const char *pathname,
+    size_t num_attr, const char **names, const char **values)
+{
+	int error;
+	git_attr_path path;
+	git_vector files = GIT_VECTOR_INIT;
+	unsigned int i, j, k;
+	git_attr_file *file;
+	git_attr_rule *rule;
+	attr_get_many_info *info = NULL;
+	size_t num_found = 0;
+
+	memset(values, 0, sizeof(const char *) * num_attr);
+
+	if ((error = git_attr_path__init(&path, pathname)) < GIT_SUCCESS ||
+		(error = collect_attr_files(repo, pathname, &files)) < GIT_SUCCESS)
+		return git__rethrow(error, "Could not get attributes for %s", pathname);
+
+	if ((info = git__calloc(num_attr, sizeof(attr_get_many_info))) == NULL) {
+		git__rethrow(GIT_ENOMEM, "Could not get attributes for %s", pathname);
+		goto cleanup;
+	}
+
+	git_vector_foreach(&files, i, file) {
+
+		git_attr_file__foreach_matching_rule(file, &path, j, rule) {
+
+			for (k = 0; k < num_attr; k++) {
+				int pos;
+
+				if (info[k].found != NULL) /* already found assignment */
+					continue;
+
+				if (!info[k].name.name) {
+					info[k].name.name = names[k];
+					info[k].name.name_hash = git_attr_file__name_hash(names[k]);
+				}
+
+				pos = git_vector_bsearch(&rule->assigns, &info[k].name);
+				git_clearerror(); /* okay if search failed */
+
+				if (pos >= 0) {
+					info[k].found = (git_attr_assignment *)
+						git_vector_get(&rule->assigns, pos);
+					values[k] = info[k].found->value;
+
+					if (++num_found == num_attr)
+						goto cleanup;
+				}
+			}
+		}
+	}
+
+cleanup:
+	git_vector_free(&files);
+	git__free(info);
+
+	return error;
+}
+
+
+int git_attr_foreach(
+    git_repository *repo, const char *pathname,
+	int (*callback)(const char *name, const char *value, void *payload),
+	void *payload)
+{
+	int error;
+	git_attr_path path;
+	git_vector files = GIT_VECTOR_INIT;
+	unsigned int i, j, k;
+	git_attr_file *file;
+	git_attr_rule *rule;
+	git_attr_assignment *assign;
+	git_hashtable *seen = NULL;
+
+	if ((error = git_attr_path__init(&path, pathname)) < GIT_SUCCESS ||
+		(error = collect_attr_files(repo, pathname, &files)) < GIT_SUCCESS)
+		return git__rethrow(error, "Could not get attributes for %s", pathname);
+
+	seen = git_hashtable_alloc(8, git_hash__strhash_cb, git_hash__strcmp_cb);
+	if (!seen) {
+		error = GIT_ENOMEM;
+		goto cleanup;
+	}
+
+	git_vector_foreach(&files, i, file) {
+
+		git_attr_file__foreach_matching_rule(file, &path, j, rule) {
+
+			git_vector_foreach(&rule->assigns, k, assign) {
+				/* skip if higher priority assignment was already seen */
+				if (git_hashtable_lookup(seen, assign->name))
+					continue;
+
+				error = git_hashtable_insert(seen, assign->name, assign);
+				if (error != GIT_SUCCESS)
+					goto cleanup;
+
+				error = callback(assign->name, assign->value, payload);
+				if (error != GIT_SUCCESS)
+					goto cleanup;
+			}
+		}
+	}
+
+cleanup:
+	if (seen)
+		git_hashtable_free(seen);
+	git_vector_free(&files);
+
+	if (error != GIT_SUCCESS)
+		(void)git__rethrow(error, "Could not get attributes for %s", pathname);
+
+	return error;
+}
+
+
+int git_attr_add_macro(
+	git_repository *repo,
+	const char *name,
+	const char *values)
+{
+	int error;
+	git_attr_rule *macro = NULL;
+
+	if ((error = attr_cache_init(repo)) < GIT_SUCCESS)
+		return error;
+
+	macro = git__calloc(1, sizeof(git_attr_rule));
+	if (!macro)
+		return GIT_ENOMEM;
+
+	macro->match.pattern = git__strdup(name);
+	if (!macro->match.pattern) {
+		git__free(macro);
+		return GIT_ENOMEM;
+	}
+
+	macro->match.length = strlen(macro->match.pattern);
+	macro->match.flags = GIT_ATTR_FNMATCH_MACRO;
+
+	error = git_attr_assignment__parse(repo, &macro->assigns, &values);
+
+	if (error == GIT_SUCCESS)
+		error = git_attr_cache__insert_macro(repo, macro);
+
+	if (error < GIT_SUCCESS)
+		git_attr_rule__free(macro);
+
+	return error;
+}
+
+
+/* add git_attr_file to vector of files, loading if needed */
+static int push_attrs(
+	git_repository *repo,
+	git_vector     *files,
+	const char     *base,
+	const char     *filename)
+{
+	int error = GIT_SUCCESS;
+	git_attr_cache *cache = &repo->attrcache;
+	git_buf path = GIT_BUF_INIT;
+	git_attr_file *file;
+	int add_to_cache = 0;
+
+	if ((error = git_path_prettify(&path, filename, base)) < GIT_SUCCESS) {
+		if (error == GIT_EOSERR)
+			/* file was not found -- ignore error */
+			error = GIT_SUCCESS;
+		goto cleanup;
+	}
+
+	/* either get attr_file from cache or read from disk */
+	file = git_hashtable_lookup(cache->files, path.ptr);
+	if (file == NULL) {
+		error = git_attr_file__from_file(repo, path.ptr, &file);
+		add_to_cache = (error == GIT_SUCCESS);
+	}
+
+	if (file != NULL) {
+		/* add file to vector, if we found it */
+		error = git_vector_insert(files, file);
+
+		/* add file to cache, if it is new */
+		/* do this after above step b/c it is not critical */
+		if (error == GIT_SUCCESS && add_to_cache && file->path != NULL)
+			error = git_hashtable_insert(cache->files, file->path, file);
+	}
+
+cleanup:
+	git_buf_free(&path);
+	return error;
+}
+
+
+static int collect_attr_files(
+	git_repository *repo, const char *path, git_vector *files)
+{
+	int error = GIT_SUCCESS;
+	git_buf dir = GIT_BUF_INIT;
+	git_config *cfg;
+	const char *workdir = git_repository_workdir(repo);
+
+	if ((error = attr_cache_init(repo)) < GIT_SUCCESS)
+		goto cleanup;
+
+	if ((error = git_vector_init(files, 4, NULL)) < GIT_SUCCESS)
+		goto cleanup;
+
+	if ((error = git_path_prettify(&dir, path, workdir)) < GIT_SUCCESS)
+		goto cleanup;
+
+	if (git_futils_isdir(dir.ptr) != GIT_SUCCESS) {
+		git_path_dirname_r(&dir, dir.ptr);
+		git_path_to_dir(&dir);
+		if ((error = git_buf_lasterror(&dir)) < GIT_SUCCESS)
+			goto cleanup;
+	}
+
+	/* in precendence order highest to lowest:
+	 * - $GIT_DIR/info/attributes
+	 * - path components with .gitattributes
+	 * - config core.attributesfile
+	 * - $GIT_PREFIX/etc/gitattributes
+	 */
+
+	error = push_attrs(repo, files, repo->path_repository, GIT_ATTR_FILE_INREPO);
+	if (error < GIT_SUCCESS)
+		goto cleanup;
+
+	if (workdir && git__prefixcmp(dir.ptr, workdir) == 0) {
+		ssize_t rootlen = (ssize_t)strlen(workdir);
+
+		do {
+			error = push_attrs(repo, files, dir.ptr, GIT_ATTR_FILE);
+			if (error == GIT_SUCCESS) {
+				git_path_dirname_r(&dir, dir.ptr);
+				git_path_to_dir(&dir);
+				error = git_buf_lasterror(&dir);
+			}
+		} while (!error && dir.size >= rootlen);
+	} else {
+		error = push_attrs(repo, files, dir.ptr, GIT_ATTR_FILE);
+	}
+	if (error < GIT_SUCCESS)
+		goto cleanup;
+
+	if (git_repository_config(&cfg, repo) == GIT_SUCCESS) {
+		const char *core_attribs = NULL;
+		git_config_get_string(cfg, "core.attributesfile", &core_attribs);
+		git_clearerror(); /* don't care if attributesfile is not set */
+		if (core_attribs)
+			error = push_attrs(repo, files, NULL, core_attribs);
+		git_config_free(cfg);
+	}
+
+	if (error == GIT_SUCCESS) {
+		error = git_futils_find_system_file(&dir, GIT_ATTR_FILE_SYSTEM);
+		if (error == GIT_SUCCESS)
+			error = push_attrs(repo, files, NULL, dir.ptr);
+		else if (error == GIT_ENOTFOUND)
+			error = GIT_SUCCESS;
+	}
+
+ cleanup:
+	if (error < GIT_SUCCESS) {
+		git__rethrow(error, "Could not get attributes for '%s'", path);
+		git_vector_free(files);
+	}
+	git_buf_free(&dir);
+
+	return error;
+}
+
+
+static int attr_cache_init(git_repository *repo)
+{
+	int error = GIT_SUCCESS;
+	git_attr_cache *cache = &repo->attrcache;
+
+	if (cache->initialized)
+		return GIT_SUCCESS;
+
+	if (cache->files == NULL) {
+		cache->files = git_hashtable_alloc(
+			8, git_hash__strhash_cb, git_hash__strcmp_cb);
+		if (!cache->files)
+			return git__throw(GIT_ENOMEM, "Could not initialize attribute cache");
+	}
+
+	if (cache->macros == NULL) {
+		cache->macros = git_hashtable_alloc(
+			8, git_hash__strhash_cb, git_hash__strcmp_cb);
+		if (!cache->macros)
+			return git__throw(GIT_ENOMEM, "Could not initialize attribute cache");
+	}
+
+	cache->initialized = 1;
+
+	/* insert default macros */
+	error = git_attr_add_macro(repo, "binary", "-diff -crlf");
+
+	return error;
+}
+
+
+void git_attr_cache_flush(
+	git_repository *repo)
+{
+	if (!repo)
+		return;
+
+	if (repo->attrcache.files) {
+		const void *GIT_UNUSED(name);
+		git_attr_file *file;
+
+		GIT_HASHTABLE_FOREACH(repo->attrcache.files, name, file,
+			git_attr_file__free(file));
+
+		git_hashtable_free(repo->attrcache.files);
+		repo->attrcache.files = NULL;
+	}
+
+	if (repo->attrcache.macros) {
+		const void *GIT_UNUSED(name);
+		git_attr_rule *rule;
+
+		GIT_HASHTABLE_FOREACH(repo->attrcache.macros, name, rule,
+			git_attr_rule__free(rule));
+
+		git_hashtable_free(repo->attrcache.macros);
+		repo->attrcache.macros = NULL;
+	}
+
+	repo->attrcache.initialized = 0;
+}

src/attr_file.h

+/*
+ * Copyright (C) 2009-2011 the libgit2 contributors
+ *
+ * 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_attr_file_h__
+#define INCLUDE_attr_file_h__
+
+#include "git2/attr.h"
+#include "vector.h"
+#include "hashtable.h"
+
+#define GIT_ATTR_FNMATCH_NEGATIVE	(1U << 0)
+#define GIT_ATTR_FNMATCH_DIRECTORY	(1U << 1)
+#define GIT_ATTR_FNMATCH_FULLPATH	(1U << 2)
+#define GIT_ATTR_FNMATCH_MACRO		(1U << 3)
+
+typedef struct {
+	char *pattern;
+	size_t length;
+	unsigned int flags;
+} git_attr_fnmatch;
+
+typedef struct {
+	git_refcount unused;
+	const char *name;
+    unsigned long name_hash;
+} git_attr_name;
+
+typedef struct {
+	git_refcount rc;			/* for macros */
+	char *name;
+    unsigned long name_hash;
+    const char *value;
+	int is_allocated;
+} git_attr_assignment;
+
+typedef struct {
+	git_attr_fnmatch match;
+	git_vector assigns;			/* vector of <git_attr_assignment*> */
+} git_attr_rule;
+
+typedef struct {
+	char *path;					/* cache the path this was loaded from */
+	git_vector rules;			/* vector of <git_attr_rule*> */
+} git_attr_file;
+
+typedef struct {
+	const char *path;
+	const char *basename;
+	int is_dir;
+} git_attr_path;
+
+typedef struct {
+	int initialized;
+	git_hashtable *files;	  /* hash path to git_attr_file */
+	git_hashtable *macros;	  /* hash name to vector<git_attr_assignment> */
+} git_attr_cache;
+
+/*
+ * git_attr_file API
+ */
+
+extern int git_attr_file__from_buffer(
+	git_repository *repo, const char *buf, git_attr_file **out);
+extern int git_attr_file__from_file(
+	git_repository *repo, const char *path, git_attr_file **out);
+
+extern void git_attr_file__free(git_attr_file *file);
+
+extern int git_attr_file__lookup_one(
+	git_attr_file *file,
+	const git_attr_path *path,
+	const char *attr,
+	const char **value);
+
+/* loop over rules in file from bottom to top */
+#define git_attr_file__foreach_matching_rule(file, path, iter, rule)	\
+	git_vector_rforeach(&(file)->rules, (iter), (rule)) \
+		if (git_attr_rule__match_path((rule), (path)) == GIT_SUCCESS)
+
+extern unsigned long git_attr_file__name_hash(const char *name);
+
+
+/*
+ * other utilities
+ */
+
+extern void git_attr_rule__free(git_attr_rule *rule);
+
+extern int git_attr_rule__match_path(
+	git_attr_rule *rule,
+	const git_attr_path *path);
+
+extern git_attr_assignment *git_attr_rule__lookup_assignment(
+	git_attr_rule *rule, const char *name);
+
+extern int git_attr_path__init(
+	git_attr_path *info, const char *path);
+
+extern int git_attr_assignment__parse(
+	git_repository *repo, /* needed to expand macros */
+	git_vector *assigns,
+	const char **scan);
+
+extern int git_attr_cache__insert_macro(
+	git_repository *repo, git_attr_rule *macro);
+
+#endif

src/config.c

@@ -337,6 +337,11 @@ int git_config_get_string(git_config *cfg, const char *name, const char **out)
 	return git__throw(error, "Config value '%s' not found", name);
 }
 
+int git_config_find_global_r(git_buf *path)
+{
+	return git_futils_find_global_file(path, GIT_CONFIG_FILENAME);
+}
+
 int git_config_find_global(char *global_config_path)
 {
 	git_buf path  = GIT_BUF_INIT;
@@ -354,79 +359,9 @@ int git_config_find_global(char *global_config_path)
 	return error;
 }
 
-int git_config_find_global_r(git_buf *path)
+int git_config_find_system_r(git_buf *path)
 {
-	int error;
-	const char *home = getenv("HOME");
-
-#ifdef GIT_WIN32
-	if (home == NULL)
-		home = getenv("USERPROFILE");
-#endif
-
-	if (home == NULL)
-		return git__throw(GIT_EOSERR, "Failed to open global config file. Cannot locate the user's home directory");
-
-	if ((error = git_buf_joinpath(path, home, GIT_CONFIG_FILENAME)) < GIT_SUCCESS)
-		return error;
-
-	if (git_futils_exists(path->ptr) < GIT_SUCCESS) {
-		git_buf_clear(path);
-		return git__throw(GIT_EOSERR, "Failed to open global config file. The file does not exist");
-	}
-
-	return GIT_SUCCESS;
-}
-
-
-
-#if GIT_WIN32
-static int win32_find_system(git_buf *system_config_path)
-{
-	const wchar_t *query = L"%PROGRAMFILES%\\Git\\etc\\gitconfig";
-	wchar_t *apphome_utf16;
-	char *apphome_utf8;
-	DWORD size, ret;
-
-	size = ExpandEnvironmentStringsW(query, NULL, 0);
-	/* The function gave us the full size of the buffer in chars, including NUL */
-	apphome_utf16 = git__malloc(size * sizeof(wchar_t));
-	if (apphome_utf16 == NULL)
-		return GIT_ENOMEM;
-
-	ret = ExpandEnvironmentStringsW(query, apphome_utf16, size);
-	if (ret != size)
-		return git__throw(GIT_ERROR, "Failed to expand environment strings");
-
-	if (_waccess(apphome_utf16, F_OK) < 0) {
-		git__free(apphome_utf16);
-		return GIT_ENOTFOUND;
-	}
-
-	apphome_utf8 = gitwin_from_utf16(apphome_utf16);
-	git__free(apphome_utf16);
-
-	git_buf_attach(system_config_path, apphome_utf8, 0);
-
-	return GIT_SUCCESS;
-}
-#endif
-
-int git_config_find_system_r(git_buf *system_config_path)
-{
-	if (git_buf_sets(system_config_path, "/etc/gitconfig") < GIT_SUCCESS)
-		return git_buf_lasterror(system_config_path);
-
-	if (git_futils_exists(system_config_path->ptr) == GIT_SUCCESS)
-		return GIT_SUCCESS;
-
-	git_buf_clear(system_config_path);
-
-#if GIT_WIN32
-	return win32_find_system(system_config_path);
-#else
-	return GIT_ENOTFOUND;
-#endif
+	return git_futils_find_system_file(path, GIT_CONFIG_FILENAME_SYSTEM);
 }
 
 int git_config_find_system(char *system_config_path)

src/config.h

@@ -14,6 +14,7 @@
 
 #define GIT_CONFIG_FILENAME ".gitconfig"
 #define GIT_CONFIG_FILENAME_INREPO "config"
+#define GIT_CONFIG_FILENAME_SYSTEM "gitconfig"
 #define GIT_CONFIG_FILE_MODE 0666
 
 struct git_config {

src/fileops.c

@@ -403,3 +403,134 @@ int git_futils_contains_file(git_buf *base, const char *file, int append_if_exis
 	return _check_dir_contents(base, file, append_if_exists, &git_futils_isfile);
 }
 
+int git_futils_find_global_file(git_buf *path, const char *filename)
+{
+	int error;
+	const char *home = getenv("HOME");
+
+#ifdef GIT_WIN32
+	if (home == NULL)
+		home = getenv("USERPROFILE");
+#endif
+
+	if (home == NULL)
+		return git__throw(GIT_EOSERR, "Failed to open global %s file. "
+			"Cannot locate the user's home directory.", filename);
+
+	if ((error = git_buf_joinpath(path, home, filename)) < GIT_SUCCESS)
+		return error;
+
+	if (git_futils_exists(path->ptr) < GIT_SUCCESS) {
+		git_buf_clear(path);
+		return GIT_ENOTFOUND;
+	}
+
+	return GIT_SUCCESS;
+}
+
+#ifdef GIT_WIN32
+typedef struct {
+	wchar_t *path;
+	DWORD len;
+} win32_path;
+
+static const win32_path *win32_system_root(void)
+{
+	static win32_path s_root = { 0, 0 };
+
+	if (s_root.path == NULL) {
+		const wchar_t *root_tmpl = L"%PROGRAMFILES%\\Git\\etc\\";
+
+		s_root.len = ExpandEnvironmentStringsW(root_tmpl, NULL, 0);
+
+		if (s_root.len <= 0) {
+			git__throw(GIT_EOSERR, "Failed to expand environment strings");
+			return NULL;
+		}
+
+		s_root.path = git__calloc(s_root.len, sizeof(wchar_t));
+		if (s_root.path == NULL)
+			return NULL;
+
+		if (ExpandEnvironmentStringsW(root_tmpl, s_root.path, s_root.len) != s_root.len) {
+			git__throw(GIT_EOSERR, "Failed to expand environment strings");
+			git__free(s_root.path);
+			s_root.path = NULL;
+			return NULL;
+		}
+	}
+
+	return &s_root;
+}
+
+static int win32_find_system_file(git_buf *path, const char *filename)
+{
+	int error = GIT_SUCCESS;
+	const win32_path *root = win32_system_root();
+	size_t len;
+	wchar_t *file_utf16 = NULL, *scan;
+	char *file_utf8 = NULL;
+
+	if (!root || !filename || (len = strlen(filename)) == 0)
+		return GIT_ENOTFOUND;
+
+	/* allocate space for wchar_t path to file */
+	file_utf16 = git__calloc(root->len + len + 2, sizeof(wchar_t));
+	if (!file_utf16)
+		return GIT_ENOMEM;
+
+	/* append root + '\\' + filename as wchar_t */
+	memcpy(file_utf16, root->path, root->len * sizeof(wchar_t));
+
+	if (*filename == '/' || *filename == '\\')
+		filename++;
+
+	if (gitwin_append_utf16(file_utf16 + root->len - 1, filename, len + 1) !=
+		(int)len) {
+		error = git__throw(GIT_EOSERR, "Failed to build file path");
+		goto cleanup;
+	}
+
+	for (scan = file_utf16; *scan; scan++)
+		if (*scan == L'/')
+			*scan = L'\\';
+
+	/* check access */
+	if (_waccess(file_utf16, F_OK) < 0) {
+		error = GIT_ENOTFOUND;
+		goto cleanup;
+	}
+
+	/* convert to utf8 */
+	if ((file_utf8 = gitwin_from_utf16(file_utf16)) == NULL)
+		error = GIT_ENOMEM;
+
+	if (file_utf8) {
+		git_path_mkposix(file_utf8);
+		git_buf_attach(path, file_utf8, 0);
+	}
+
+cleanup:
+	git__free(file_utf16);
+
+	return error;
+}
+#endif
+
+int git_futils_find_system_file(git_buf *path, const char *filename)
+{
+	if (git_buf_joinpath(path, "/etc", filename) < GIT_SUCCESS)
+		return git_buf_lasterror(path);
+
+	if (git_futils_exists(path->ptr) == GIT_SUCCESS)
+		return GIT_SUCCESS;
+
+	git_buf_clear(path);
+
+#ifdef GIT_WIN32
+	return win32_find_system_file(path, filename);
+#else
+	return GIT_ENOTFOUND;
+#endif
+}
+

src/fileops.h

@@ -165,4 +165,28 @@ extern int git_futils_direach(
 extern int git_futils_cmp_path(const char *name1, int len1, int isdir1,
 		const char *name2, int len2, int isdir2);
 
+/**
+ * Find a "global" file (i.e. one in a user's home directory).
+ *
+ * @param pathbuf buffer to write the full path into
+ * @param filename name of file to find in the home directory
+ * @return
+ * - GIT_SUCCESS if found;
+ * - GIT_ENOTFOUND if not found;
+ * - GIT_EOSERR on an unspecified OS related error.
+ */
+extern int git_futils_find_global_file(git_buf *path, const char *filename);
+
+/**
+ * Find a "system" file (i.e. one shared for all users of the system).
+ *
+ * @param pathbuf buffer to write the full path into
+ * @param filename name of file to find in the home directory
+ * @return
+ * - GIT_SUCCESS if found;
+ * - GIT_ENOTFOUND if not found;
+ * - GIT_EOSERR on an unspecified OS related error.
+ */
+extern int git_futils_find_system_file(git_buf *path, const char *filename);
+
 #endif /* INCLUDE_fileops_h__ */

src/hashtable.c

@@ -241,3 +241,17 @@ int git_hashtable_merge(git_hashtable *self, git_hashtable *other)
 	return insert_nodes(self, other->nodes, other->key_count);
 }
 
+
+/**
+ * Standard string
+ */
+uint32_t git_hash__strhash_cb(const void *key, int hash_id)
+{
+	static uint32_t hash_seeds[GIT_HASHTABLE_HASHES] = {
+		2147483647,
+		0x5d20bb23,
+		0x7daaab3c
+	};
+
+	return git__hash(key, strlen((const char *)key), hash_seeds[hash_id]);
+}

src/hashtable.h

@@ -76,5 +76,12 @@ GIT_INLINE(int) git_hashtable_insert(git_hashtable *h, const void *key, void *va
 	_node->key = NULL; _node->value = NULL; _self->key_count--;\
 }
 
+/*
+ * If you want a hashtable with standard string keys, you can
+ * just pass git_hash__strcmp_cb and git_hash__strhash_cb to
+ * git_hashtable_alloc.
+ */
+#define git_hash__strcmp_cb git__strcmp_cb
+extern uint32_t git_hash__strhash_cb(const void *key, int hash_id);
 
 #endif

src/refs.c

@@ -31,17 +31,6 @@ struct packref {
 
 static const int default_table_size = 32;
 
-static uint32_t reftable_hash(const void *key, int hash_id)
-{
-	static uint32_t hash_seeds[GIT_HASHTABLE_HASHES] = {
-		2147483647,
-		0x5d20bb23,
-		0x7daaab3c
-	};
-
-	return git__hash(key, strlen((const char *)key), hash_seeds[hash_id]);
-}
-
 static int reference_read(
 	git_fbuffer *file_content,
 	time_t *mtime,
@@ -445,9 +434,7 @@ static int packed_load(git_repository *repo)
 	/* First we make sure we have allocated the hash table */
 	if (ref_cache->packfile == NULL) {
 		ref_cache->packfile = git_hashtable_alloc(
-			default_table_size,
-			reftable_hash,
-			(git_hash_keyeq_ptr)&git__strcmp_cb);
+			default_table_size, git_hash__strhash_cb, git_hash__strcmp_cb);
 
 		if (ref_cache->packfile == NULL) {
 			error = GIT_ENOMEM;

src/repository.c

@@ -59,6 +59,7 @@ void git_repository_free(git_repository *repo)
 
 	git_cache_free(&repo->objects);
 	git_repository__refcache_free(&repo->references);
+	git_attr_cache_flush(repo);
 
 	git__free(repo->path_repository);
 	git__free(repo->workdir);

src/repository.h

@@ -19,6 +19,7 @@
 #include "refs.h"
 #include "buffer.h"
 #include "odb.h"
+#include "attr_file.h"
 
 #define DOT_GIT ".git"
 #define GIT_DIR DOT_GIT "/"
@@ -38,6 +39,7 @@ struct git_repository {
 
 	git_cache objects;
 	git_refcache references;
+	git_attr_cache attrcache;
 
 	char *path_repository;
 	char *workdir;

src/util.c

@@ -348,22 +348,30 @@ uint32_t git__hash(const void *key, int len, uint32_t seed)
  * Copyright (c) 1990 Regents of the University of California.
  * All rights reserved.
  */
-void **git__bsearch(const void *key, void **base, size_t nmemb, int (*compar)(const void *, const void *))
+int git__bsearch(
+	void **array,
+	size_t array_len,
+	const void *key,
+	int (*compare)(const void *, const void *),
+	size_t *position)
 {
-		int lim, cmp;
-		void **p;
-
-		for (lim = nmemb; lim != 0; lim >>= 1) {
-				p = base + (lim >> 1);
-				cmp = (*compar)(key, *p);
-				if (cmp > 0) { /* key > p: move right */
-						base = p + 1;
-						lim--;
-				} else if (cmp == 0) {
-						return (void **)p;
-				} /* else move left */
-		}
-		return NULL;
+	int lim, cmp;
+	void **part, **base = array;
+
+	for (lim = array_len; lim != 0; lim >>= 1) {
+		part = base + (lim >> 1);
+		cmp = (*compare)(key, *part);
+		if (cmp == 0) {
+			*position = (part - array);
+			return GIT_SUCCESS;
+		} else if (cmp > 0) { /* key > p; take right partition */
+			base = part + 1;
+			lim--;
+		} /* else take left partition */
+	}
+
+	*position = (base - array);
+	return GIT_ENOTFOUND;
 }
 
 /**

src/util.h

@@ -105,8 +105,13 @@ extern void git__strtolower(char *str);
 extern int git__fnmatch(const char *pattern, const char *name, int flags);
 
 extern void git__tsort(void **dst, size_t size, int (*cmp)(const void *, const void *));
-extern void **git__bsearch(const void *key, void **base, size_t nmemb,
-	int (*compar)(const void *, const void *));
+
+extern int git__bsearch(
+	void **array,
+	size_t array_len,
+	const void *key,
+	int (*compare)(const void *, const void *),
+	size_t *position);
 
 extern int git__strcmp_cb(const void *a, const void *b);
 

src/vector.c

@@ -29,7 +29,12 @@ static int resize_vector(git_vector *v)
 void git_vector_free(git_vector *v)
 {
 	assert(v);
+
 	git__free(v->contents);
+	v->contents = NULL;
+
+	v->length = 0;
+	v->_alloc_size = 0;
 }
 
 int git_vector_init(git_vector *v, unsigned int initial_size, git_vector_cmp cmp)
@@ -69,6 +74,45 @@ int git_vector_insert(git_vector *v, void *element)
 	return GIT_SUCCESS;
 }
 
+int git_vector_insert_sorted(git_vector *v, void *element, int (*on_dup)(void **old, void *new))
+{
+	int error = GIT_SUCCESS;
+	size_t pos;
+
+	assert(v && v->_cmp);
+
+	if (!v->sorted)
+		git_vector_sort(v);
+
+	if (v->length >= v->_alloc_size) {
+		if (resize_vector(v) < 0)
+			return GIT_ENOMEM;
+	}
+
+	error = git__bsearch(v->contents, v->length, element, v->_cmp, &pos);
+
+	/* If we found the element and have a duplicate handler callback,
+	 * invoke it.  If it returns an error, then cancel insert, otherwise
+	 * proceed with normal insert.
+	 */
+	if (error == GIT_SUCCESS && on_dup != NULL) {
+		error = on_dup(&v->contents[pos], element);
+		if (error != GIT_SUCCESS)
+			return error;
+	}
+
+	/* shift elements to the right */
+	if (pos < v->length) {
+		memmove(v->contents + pos + 1, v->contents + pos,
+		        (v->length - pos) * sizeof(void *));
+	}
+
+	v->contents[pos] = element;
+	v->length++;
+
+	return GIT_SUCCESS;
+}
+
 void git_vector_sort(git_vector *v)
 {
 	assert(v);
@@ -82,7 +126,7 @@ void git_vector_sort(git_vector *v)
 
 int git_vector_bsearch2(git_vector *v, git_vector_cmp key_lookup, const void *key)
 {
-	void **find;
+	size_t pos;
 
 	assert(v && key && key_lookup);
 
@@ -92,9 +136,9 @@ int git_vector_bsearch2(git_vector *v, git_vector_cmp key_lookup, const void *ke
 
 	git_vector_sort(v);
 
-	find = git__bsearch(key, v->contents, v->length, key_lookup);
-	if (find != NULL)
-		return (int)(find - v->contents);
+	if (git__bsearch(v->contents, v->length, key, key_lookup,
+			&pos) == GIT_SUCCESS)
+		return (int)pos;
 
 	return git__throw(GIT_ENOTFOUND, "Can't find element");
 }

src/vector.h

@@ -19,6 +19,8 @@ typedef struct git_vector {
 	int sorted;
 } git_vector;
 
+#define GIT_VECTOR_INIT {0}
+
 int git_vector_init(git_vector *v, unsigned int initial_size, git_vector_cmp cmp);
 void git_vector_free(git_vector *v);
 void git_vector_clear(git_vector *v);
@@ -39,7 +41,13 @@ GIT_INLINE(void *) git_vector_get(git_vector *v, unsigned int position)
 #define git_vector_foreach(v, iter, elem)	\
 	for ((iter) = 0; (iter) < (v)->length && ((elem) = (v)->contents[(iter)], 1); (iter)++ )
 
+#define git_vector_rforeach(v, iter, elem)	\
+	for ((iter) = (v)->length; (iter) > 0 && ((elem) = (v)->contents[(iter)-1], 1); (iter)-- )
+
 int git_vector_insert(git_vector *v, void *element);
+int git_vector_insert_sorted(git_vector *v, void *element,
+	int (*on_dup)(void **old, void *new));
 int git_vector_remove(git_vector *v, unsigned int idx);
 void git_vector_uniq(git_vector *v);
+
 #endif

src/win32/utf-conv.c

@@ -57,6 +57,11 @@ wchar_t* gitwin_to_utf16(const char* str)
 	return ret;
 }
 
+int gitwin_append_utf16(wchar_t *buffer, const char *str, size_t len)
+{
+	return MultiByteToWideChar(_active_codepage, 0, str, -1, buffer, len);
+}
+
 char* gitwin_from_utf16(const wchar_t* str)
 {
 	char* ret;

src/win32/utf-conv.h

@@ -11,6 +11,7 @@
 #define INCLUDE_git_utfconv_h__
 
 wchar_t* gitwin_to_utf16(const char* str);
+int gitwin_append_utf16(wchar_t *buffer, const char *str, size_t len)
 char* gitwin_from_utf16(const wchar_t* str);
 
 #endif

tests-clay/attr/file.c

+#include "clay_libgit2.h"
+#include "attr_file.h"
+
+#define get_rule(X) ((git_attr_rule *)git_vector_get(&file->rules,(X)))
+#define get_assign(R,Y) ((git_attr_assignment *)git_vector_get(&(R)->assigns,(Y)))
+
+void test_attr_file__simple_read(void)
+{
+	git_attr_file *file = NULL;
+
+	cl_git_pass(git_attr_file__from_file(NULL, cl_fixture("attr/attr0"), &file));
+	cl_assert_strequal(cl_fixture("attr/attr0"), file->path);
+	cl_assert(file->rules.length == 1);
+
+	git_attr_rule *rule = get_rule(0);
+	cl_assert(rule != NULL);
+	cl_assert_strequal("*", rule->match.pattern);
+	cl_assert(rule->match.length == 1);
+	cl_assert(rule->match.flags == 0);
+
+	cl_assert(rule->assigns.length == 1);
+	git_attr_assignment *assign = get_assign(rule, 0);
+	cl_assert(assign != NULL);
+	cl_assert_strequal("binary", assign->name);
+	cl_assert(assign->value == GIT_ATTR_TRUE);
+	cl_assert(!assign->is_allocated);
+
+	git_attr_file__free(file);
+}
+
+void test_attr_file__match_variants(void)
+{
+	git_attr_file *file = NULL;
+	git_attr_rule *rule;
+	git_attr_assignment *assign;
+
+	cl_git_pass(git_attr_file__from_file(NULL, cl_fixture("attr/attr1"), &file));
+	cl_assert_strequal(cl_fixture("attr/attr1"), file->path);
+	cl_assert(file->rules.length == 10);
+
+	/* let's do a thorough check of this rule, then just verify
+	 * the things that are unique for the later rules
+	 */
+	rule = get_rule(0);
+	cl_assert(rule);
+	cl_assert_strequal("pat0", rule->match.pattern);
+	cl_assert(rule->match.length == strlen("pat0"));
+	cl_assert(rule->match.flags == 0);
+	cl_assert(rule->assigns.length == 1);
+	assign = get_assign(rule,0);
+	cl_assert_strequal("attr0", assign->name);
+	cl_assert(assign->name_hash == git_attr_file__name_hash(assign->name));
+	cl_assert(assign->value == GIT_ATTR_TRUE);
+	cl_assert(!assign->is_allocated);
+
+	rule = get_rule(1);
+	cl_assert_strequal("pat1", rule->match.pattern);
+	cl_assert(rule->match.length == strlen("pat1"));
+	cl_assert(rule->match.flags == GIT_ATTR_FNMATCH_NEGATIVE);
+
+	rule = get_rule(2);
+	cl_assert_strequal("pat2", rule->match.pattern);
+	cl_assert(rule->match.length == strlen("pat2"));
+	cl_assert(rule->match.flags == GIT_ATTR_FNMATCH_DIRECTORY);
+
+	rule = get_rule(3);
+	cl_assert_strequal("pat3dir/pat3file", rule->match.pattern);
+	cl_assert(rule->match.flags == GIT_ATTR_FNMATCH_FULLPATH);
+
+	rule = get_rule(4);
+	cl_assert_strequal("pat4.*", rule->match.pattern);
+	cl_assert(rule->match.flags == 0);
+
+	rule = get_rule(5);
+	cl_assert_strequal("*.pat5", rule->match.pattern);
+
+	rule = get_rule(7);
+	cl_assert_strequal("pat7[a-e]??[xyz]", rule->match.pattern);
+	cl_assert(rule->assigns.length == 1);
+	assign = get_assign(rule,0);
+	cl_assert_strequal("attr7", assign->name);
+	cl_assert(assign->value == GIT_ATTR_TRUE);
+
+	rule = get_rule(8);
+	cl_assert_strequal("pat8 with spaces", rule->match.pattern);
+	cl_assert(rule->match.length == strlen("pat8 with spaces"));
+	cl_assert(rule->match.flags == 0);
+
+	rule = get_rule(9);
+	cl_assert_strequal("pat9", rule->match.pattern);
+
+	git_attr_file__free(file);
+}
+
+static void check_one_assign(
+	git_attr_file *file,
+	int rule_idx,
+	int assign_idx,
+	const char *pattern,
+	const char *name,
+	const char *value,
+	int is_allocated)
+{
+	git_attr_rule *rule = get_rule(rule_idx);
+	git_attr_assignment *assign = get_assign(rule, assign_idx);
+
+	cl_assert_strequal(pattern, rule->match.pattern);
+	cl_assert(rule->assigns.length == 1);
+	cl_assert_strequal(name, assign->name);
+	cl_assert(assign->name_hash == git_attr_file__name_hash(assign->name));
+	cl_assert(assign->is_allocated == is_allocated);
+	if (is_allocated)
+		cl_assert_strequal(value, assign->value);
+	else
+		cl_assert(assign->value == value);
+}
+
+void test_attr_file__assign_variants(void)
+{
+	git_attr_file *file = NULL;
+	git_attr_rule *rule;
+	git_attr_assignment *assign;
+
+	cl_git_pass(git_attr_file__from_file(NULL, cl_fixture("attr/attr2"), &file));
+	cl_assert_strequal(cl_fixture("attr/attr2"), file->path);
+	cl_assert(file->rules.length == 11);
+
+	check_one_assign(file, 0, 0, "pat0", "simple", GIT_ATTR_TRUE, 0);
+	check_one_assign(file, 1, 0, "pat1", "neg", GIT_ATTR_FALSE, 0);
+	check_one_assign(file, 2, 0, "*", "notundef", GIT_ATTR_TRUE, 0);
+	check_one_assign(file, 3, 0, "pat2", "notundef", NULL, 0);
+	check_one_assign(file, 4, 0, "pat3", "assigned", "test-value", 1);
+	check_one_assign(file, 5, 0, "pat4", "rule-with-more-chars", "value-with-more-chars", 1);
+	check_one_assign(file, 6, 0, "pat5", "empty", GIT_ATTR_TRUE, 0);
+	check_one_assign(file, 7, 0, "pat6", "negempty", GIT_ATTR_FALSE, 0);
+
+	rule = get_rule(8);
+	cl_assert_strequal("pat7", rule->match.pattern);
+	cl_assert(rule->assigns.length == 5);
+	/* assignments will be sorted by hash value, so we have to do
+	 * lookups by search instead of by position
+	 */
+	assign = git_attr_rule__lookup_assignment(rule, "multiple");
+	cl_assert(assign);
+	cl_assert_strequal("multiple", assign->name);
+	cl_assert(assign->value == GIT_ATTR_TRUE);
+	assign = git_attr_rule__lookup_assignment(rule, "single");
+	cl_assert(assign);
+	cl_assert_strequal("single", assign->name);
+	cl_assert(assign->value == GIT_ATTR_FALSE);
+	assign = git_attr_rule__lookup_assignment(rule, "values");
+	cl_assert(assign);
+	cl_assert_strequal("values", assign->name);
+	cl_assert_strequal("1", assign->value);
+	assign = git_attr_rule__lookup_assignment(rule, "also");
+	cl_assert(assign);
+	cl_assert_strequal("also", assign->name);
+	cl_assert_strequal("a-really-long-value/*", assign->value);
+	assign = git_attr_rule__lookup_assignment(rule, "happy");
+	cl_assert(assign);
+	cl_assert_strequal("happy", assign->name);
+	cl_assert_strequal("yes!", assign->value);
+	assign = git_attr_rule__lookup_assignment(rule, "other");
+	cl_assert(!assign);
+
+	rule = get_rule(9);
+	cl_assert_strequal("pat8", rule->match.pattern);
+	cl_assert(rule->assigns.length == 2);
+	assign = git_attr_rule__lookup_assignment(rule, "again");
+	cl_assert(assign);
+	cl_assert_strequal("again", assign->name);
+	cl_assert(assign->value == GIT_ATTR_TRUE);
+	assign = git_attr_rule__lookup_assignment(rule, "another");
+	cl_assert(assign);
+	cl_assert_strequal("another", assign->name);
+	cl_assert_strequal("12321", assign->value);
+
+	check_one_assign(file, 10, 0, "pat9", "at-eof", GIT_ATTR_FALSE, 0);
+
+	git_attr_file__free(file);
+}
+
+void test_attr_file__check_attr_examples(void)
+{
+	git_attr_file *file = NULL;
+	git_attr_rule *rule;
+	git_attr_assignment *assign;
+
+	cl_git_pass(git_attr_file__from_file(NULL, cl_fixture("attr/attr3"), &file));
+	cl_assert_strequal(cl_fixture("attr/attr3"), file->path);
+	cl_assert(file->rules.length == 3);
+
+	rule = get_rule(0);
+	cl_assert_strequal("*.java", rule->match.pattern);
+	cl_assert(rule->assigns.length == 3);
+	assign = git_attr_rule__lookup_assignment(rule, "diff");
+	cl_assert_strequal("diff", assign->name);
+	cl_assert_strequal("java", assign->value);
+	assign = git_attr_rule__lookup_assignment(rule, "crlf");
+	cl_assert_strequal("crlf", assign->name);
+	cl_assert(GIT_ATTR_FALSE == assign->value);
+	assign = git_attr_rule__lookup_assignment(rule, "myAttr");
+	cl_assert_strequal("myAttr", assign->name);
+	cl_assert(GIT_ATTR_TRUE == assign->value);
+	assign = git_attr_rule__lookup_assignment(rule, "missing");
+	cl_assert(assign == NULL);
+
+	rule = get_rule(1);
+	cl_assert_strequal("NoMyAttr.java", rule->match.pattern);
+	cl_assert(rule->assigns.length == 1);
+	assign = get_assign(rule, 0);
+	cl_assert_strequal("myAttr", assign->name);
+	cl_assert(assign->value == NULL);
+
+	rule = get_rule(2);
+	cl_assert_strequal("README", rule->match.pattern);
+	cl_assert(rule->assigns.length == 1);
+	assign = get_assign(rule, 0);
+	cl_assert_strequal("caveat", assign->name);
+	cl_assert_strequal("unspecified", assign->value);
+
+	git_attr_file__free(file);
+}

tests-clay/attr/lookup.c

+#include "clay_libgit2.h"
+#include "attr_file.h"
+
+void test_attr_lookup__simple(void)
+{
+	git_attr_file *file = NULL;
+	git_attr_path path;
+	const char *value = NULL;
+
+	cl_git_pass(git_attr_file__from_file(NULL, cl_fixture("attr/attr0"), &file));
+	cl_assert_strequal(cl_fixture("attr/attr0"), file->path);
+	cl_assert(file->rules.length == 1);
+
+	cl_git_pass(git_attr_path__init(&path, "test"));
+	cl_assert_strequal("test", path.path);
+	cl_assert_strequal("test", path.basename);
+	cl_assert(!path.is_dir);
+
+	cl_git_pass(git_attr_file__lookup_one(file,&path,"binary",&value));
+	cl_assert(value == GIT_ATTR_TRUE);
+
+	cl_git_pass(git_attr_file__lookup_one(file,&path,"missing",&value));
+	cl_assert(!value);
+
+	git_attr_file__free(file);
+}
+
+typedef struct {
+	const char *path;
+	const char *attr;
+	const char *expected;
+	int use_strcmp;
+	int force_dir;
+} test_case;
+
+static void run_test_cases(git_attr_file *file, test_case *cases)
+{
+	git_attr_path path;
+	const char *value = NULL;
+	test_case *c;
+	int error;
+
+	for (c = cases; c->path != NULL; c++) {
+		cl_git_pass(git_attr_path__init(&path, c->path));
+
+		if (c->force_dir)
+			path.is_dir = 1;
+
+		error = git_attr_file__lookup_one(file,&path,c->attr,&value);
+		if (error != GIT_SUCCESS)
+			fprintf(stderr, "failure with %s %s %s\n", c->path, c->attr, c->expected);
+		cl_git_pass(error);
+
+		if (c->use_strcmp)
+			cl_assert_strequal(c->expected, value);
+		else
+			cl_assert(c->expected == value);
+	}
+}
+
+void test_attr_lookup__match_variants(void)
+{
+	git_attr_file *file = NULL;
+	git_attr_path path;
+	test_case cases[] = {
+		/* pat0 -> simple match */
+		{ "pat0", "attr0", GIT_ATTR_TRUE, 0, 0 },
+		{ "/testing/for/pat0", "attr0", GIT_ATTR_TRUE, 0, 0 },
+		{ "relative/to/pat0", "attr0", GIT_ATTR_TRUE, 0, 0 },
+		{ "this-contains-pat0-inside", "attr0", NULL, 0, 0 },
+		{ "this-aint-right", "attr0", NULL, 0, 0 },
+		{ "/this/pat0/dont/match", "attr0", NULL, 0, 0 },
+		/* negative match */
+		{ "pat0", "attr1", GIT_ATTR_TRUE, 0, 0 },
+		{ "pat1", "attr1", NULL, 0, 0 },
+		{ "/testing/for/pat1", "attr1", NULL, 0, 0 },
+		{ "/testing/for/pat0", "attr1", GIT_ATTR_TRUE, 0, 0 },
+		{ "/testing/for/pat1/inside", "attr1", GIT_ATTR_TRUE, 0, 0 },
+		{ "misc", "attr1", GIT_ATTR_TRUE, 0, 0 },
+		/* dir match */
+		{ "pat2", "attr2", NULL, 0, 0 },
+		{ "pat2", "attr2", GIT_ATTR_TRUE, 0, 1 },
+		{ "/testing/for/pat2", "attr2", NULL, 0, 0 },
+		{ "/testing/for/pat2", "attr2", GIT_ATTR_TRUE, 0, 1 },
+		{ "/not/pat2/yousee", "attr2", NULL, 0, 0 },
+		{ "/not/pat2/yousee", "attr2", NULL, 0, 1 },
+		/* path match */
+		{ "pat3file", "attr3", NULL, 0, 0 },
+		{ "/pat3dir/pat3file", "attr3", NULL, 0, 0 },
+		{ "pat3dir/pat3file", "attr3", GIT_ATTR_TRUE, 0, 0 },
+		/* pattern* match */
+		{ "pat4.txt", "attr4", GIT_ATTR_TRUE, 0, 0 },
+		{ "/fun/fun/fun/pat4.c", "attr4", GIT_ATTR_TRUE, 0, 0 },
+		{ "pat4.", "attr4", GIT_ATTR_TRUE, 0, 0 },
+		{ "pat4", "attr4", NULL, 0, 0 },
+		{ "/fun/fun/fun/pat4.dir", "attr4", GIT_ATTR_TRUE, 0, 1 },
+		/* *pattern match */
+		{ "foo.pat5", "attr5", GIT_ATTR_TRUE, 0, 0 },
+		{ "foo.pat5", "attr5", GIT_ATTR_TRUE, 0, 1 },
+		{ "/this/is/ok.pat5", "attr5", GIT_ATTR_TRUE, 0, 0 },
+		{ "/this/is/bad.pat5/yousee.txt", "attr5", NULL, 0, 0 },
+		{ "foo.pat5", "attr100", NULL, 0, 0 },
+		/* glob match with slashes */
+		{ "foo.pat6", "attr6", NULL, 0, 0 },
+		{ "pat6/pat6/foobar.pat6", "attr6", GIT_ATTR_TRUE, 0, 0 },
+		{ "pat6/pat6/.pat6", "attr6", GIT_ATTR_TRUE, 0, 0 },
+		{ "pat6/pat6/extra/foobar.pat6", "attr6", NULL, 0, 0 },
+		{ "/prefix/pat6/pat6/foobar.pat6", "attr6", NULL, 0, 0 },
+		{ "/pat6/pat6/foobar.pat6", "attr6", NULL, 0, 0 },
+		/* complex pattern */
+		{ "pat7a12z", "attr7", GIT_ATTR_TRUE, 0, 0 },
+		{ "pat7e__x", "attr7", GIT_ATTR_TRUE, 0, 0 },
+		{ "pat7b/1y", "attr7", NULL, 0, 0 }, /* ? does not match / */
+		{ "pat7e_x", "attr7", NULL, 0, 0 },
+		{ "pat7aaaa", "attr7", NULL, 0, 0 },
+		{ "pat7zzzz", "attr7", NULL, 0, 0 },
+		{ "/this/can/be/anything/pat7a12z", "attr7", GIT_ATTR_TRUE, 0, 0 },
+		{ "but/it/still/must/match/pat7aaaa", "attr7", NULL, 0, 0 },
+		{ "pat7aaay.fail", "attr7", NULL, 0, 0 },
+		/* pattern with spaces */
+		{ "pat8 with spaces", "attr8", GIT_ATTR_TRUE, 0, 0 },
+		{ "/gotta love/pat8 with spaces", "attr8", GIT_ATTR_TRUE, 0, 0 },
+		{ "failing pat8 with spaces", "attr8", NULL, 0, 0 },
+		{ "spaces", "attr8", NULL, 0, 0 },
+		/* pattern at eof */
+		{ "pat9", "attr9", GIT_ATTR_TRUE, 0, 0 },
+		{ "/eof/pat9", "attr9", GIT_ATTR_TRUE, 0, 0 },
+		{ "pat", "attr9", NULL, 0, 0 },
+		{ "at9", "attr9", NULL, 0, 0 },
+		{ "pat9.fail", "attr9", NULL, 0, 0 },
+		/* sentinel at end */
+		{ NULL, NULL, NULL, 0, 0 }
+	};
+
+	cl_git_pass(git_attr_file__from_file(NULL, cl_fixture("attr/attr1"), &file));
+	cl_assert_strequal(cl_fixture("attr/attr1"), file->path);
+	cl_assert(file->rules.length == 10);
+
+	cl_git_pass(git_attr_path__init(&path, "/testing/for/pat0"));
+	cl_assert_strequal("pat0", path.basename);
+
+	run_test_cases(file, cases);
+
+	git_attr_file__free(file);
+}
+
+void test_attr_lookup__assign_variants(void)
+{
+	git_attr_file *file = NULL;
+	test_case cases[] = {
+		/* pat0 -> simple assign */
+		{ "pat0", "simple", GIT_ATTR_TRUE, 0, 0 },
+		{ "/testing/pat0", "simple", GIT_ATTR_TRUE, 0, 0 },
+		{ "pat0", "fail", NULL, 0, 0 },
+		{ "/testing/pat0", "fail", NULL, 0, 0 },
+		/* negative assign */
+		{ "pat1", "neg", GIT_ATTR_FALSE, 0, 0 },
+		{ "/testing/pat1", "neg", GIT_ATTR_FALSE, 0, 0 },
+		{ "pat1", "fail", NULL, 0, 0 },
+		{ "/testing/pat1", "fail", NULL, 0, 0 },
+		/* forced undef */
+		{ "pat1", "notundef", GIT_ATTR_TRUE, 0, 0 },
+		{ "pat2", "notundef", NULL, 0, 0 },
+		{ "/lead/in/pat1", "notundef", GIT_ATTR_TRUE, 0, 0 },
+		{ "/lead/in/pat2", "notundef", NULL, 0, 0 },
+		/* assign value */
+		{ "pat3", "assigned", "test-value", 1, 0 },
+		{ "pat3", "notassigned", NULL, 0, 0 },
+		/* assign value */
+		{ "pat4", "rule-with-more-chars", "value-with-more-chars", 1, 0 },
+		{ "pat4", "notassigned-rule-with-more-chars", NULL, 0, 0 },
+		/* empty assignments */
+		{ "pat5", "empty", GIT_ATTR_TRUE, 0, 0 },
+		{ "pat6", "negempty", GIT_ATTR_FALSE, 0, 0 },
+		/* multiple assignment */
+		{ "pat7", "multiple", GIT_ATTR_TRUE, 0, 0 },
+		{ "pat7", "single", GIT_ATTR_FALSE, 0, 0 },
+		{ "pat7", "values", "1", 1, 0 },
+		{ "pat7", "also", "a-really-long-value/*", 1, 0 },
+		{ "pat7", "happy", "yes!", 1, 0 },
+		{ "pat8", "again", GIT_ATTR_TRUE, 0, 0 },
+		{ "pat8", "another", "12321", 1, 0 },
+		/* bad assignment */
+		{ "patbad0", "simple", NULL, 0, 0 },
+		{ "patbad0", "notundef", GIT_ATTR_TRUE, 0, 0 },
+		{ "patbad1", "simple", NULL, 0, 0 },
+		/* eof assignment */
+		{ "pat9", "at-eof", GIT_ATTR_FALSE, 0, 0 },
+		/* sentinel at end */
+		{ NULL, NULL, NULL, 0, 0 }
+	};
+
+	cl_git_pass(git_attr_file__from_file(NULL, cl_fixture("attr/attr2"), &file));
+	cl_assert(file->rules.length == 11);
+
+	run_test_cases(file, cases);
+
+	git_attr_file__free(file);
+}
+
+void test_attr_lookup__check_attr_examples(void)
+{
+	git_attr_file *file = NULL;
+	test_case cases[] = {
+		{ "foo.java", "diff", "java", 1, 0 },
+		{ "foo.java", "crlf", GIT_ATTR_FALSE, 0, 0 },
+		{ "foo.java", "myAttr", GIT_ATTR_TRUE, 0, 0 },
+		{ "foo.java", "other", NULL, 0, 0 },
+		{ "/prefix/dir/foo.java", "diff", "java", 1, 0 },
+		{ "/prefix/dir/foo.java", "crlf", GIT_ATTR_FALSE, 0, 0 },
+		{ "/prefix/dir/foo.java", "myAttr", GIT_ATTR_TRUE, 0, 0 },
+		{ "/prefix/dir/foo.java", "other", NULL, 0, 0 },
+		{ "NoMyAttr.java", "crlf", GIT_ATTR_FALSE, 0, 0 },
+		{ "NoMyAttr.java", "myAttr", NULL, 0, 0 },
+		{ "NoMyAttr.java", "other", NULL, 0, 0 },
+		{ "/prefix/dir/NoMyAttr.java", "crlf", GIT_ATTR_FALSE, 0, 0 },
+		{ "/prefix/dir/NoMyAttr.java", "myAttr", NULL, 0, 0 },
+		{ "/prefix/dir/NoMyAttr.java", "other", NULL, 0, 0 },
+		{ "README", "caveat", "unspecified", 1, 0 },
+		{ "/specific/path/README", "caveat", "unspecified", 1, 0 },
+		{ "README", "missing", NULL, 0, 0 },
+		{ "/specific/path/README", "missing", NULL, 0, 0 },
+		/* sentinel at end */
+		{ NULL, NULL, NULL, 0, 0 }
+	};
+
+	cl_git_pass(git_attr_file__from_file(NULL, cl_fixture("attr/attr3"), &file));
+	cl_assert(file->rules.length == 3);
+
+	run_test_cases(file, cases);
+
+	git_attr_file__free(file);
+}
+
+void test_attr_lookup__from_buffer(void)
+{
+	git_attr_file *file = NULL;
+	test_case cases[] = {
+		{ "abc", "foo", GIT_ATTR_TRUE, 0, 0 },
+		{ "abc", "bar", GIT_ATTR_TRUE, 0, 0 },
+		{ "abc", "baz", GIT_ATTR_TRUE, 0, 0 },
+		{ "aaa", "foo", GIT_ATTR_TRUE, 0, 0 },
+		{ "aaa", "bar", NULL, 0, 0 },
+		{ "aaa", "baz", GIT_ATTR_TRUE, 0, 0 },
+		{ "qqq", "foo", NULL, 0, 0 },
+		{ "qqq", "bar", NULL, 0, 0 },
+		{ "qqq", "baz", GIT_ATTR_TRUE, 0, 0 },
+		{ NULL, NULL, NULL, 0, 0 }
+	};
+
+	cl_git_pass(git_attr_file__from_buffer(NULL, "a* foo\nabc bar\n* baz", &file));
+	cl_assert(file->rules.length == 3);
+
+	run_test_cases(file, cases);
+
+	git_attr_file__free(file);
+}

tests-clay/attr/repo.c

+#include "clay_libgit2.h"
+#include "fileops.h"
+#include "git2/attr.h"
+
+static git_repository *g_repo = NULL;
+
+void test_attr_repo__initialize(void)
+{
+	/* Before each test, instantiate the attr repo from the fixtures and
+	 * rename the .gitted to .git so it is a repo with a working dir.
+	 * Also rename gitattributes to .gitattributes, because it contains
+	 * macro definitions which are only allowed in the root.
+	 */
+	cl_fixture_sandbox("attr");
+	cl_git_pass(p_rename("attr/.gitted", "attr/.git"));
+	cl_git_pass(p_rename("attr/gitattributes", "attr/.gitattributes"));
+	cl_git_pass(git_repository_open(&g_repo, "attr/.git"));
+}
+
+void test_attr_repo__cleanup(void)
+{
+	git_repository_free(g_repo);
+	g_repo = NULL;
+	cl_fixture_cleanup("attr");
+}
+
+void test_attr_repo__get_one(void)
+{
+	const char *value;
+	struct {
+		const char *file;
+		const char *attr;
+		const char *expected;
+	} test_cases[] = {
+		{ "root_test1", "repoattr", GIT_ATTR_TRUE },
+		{ "root_test1", "rootattr", GIT_ATTR_TRUE },
+		{ "root_test1", "missingattr", NULL },
+		{ "root_test1", "subattr", NULL },
+		{ "root_test1", "negattr", NULL },
+		{ "root_test2", "repoattr", GIT_ATTR_TRUE },
+		{ "root_test2", "rootattr", GIT_ATTR_FALSE },
+		{ "root_test2", "missingattr", NULL },
+		{ "root_test2", "multiattr", GIT_ATTR_FALSE },
+		{ "root_test3", "repoattr", GIT_ATTR_TRUE },
+		{ "root_test3", "rootattr", NULL },
+		{ "root_test3", "multiattr", "3" },
+		{ "root_test3", "multi2", NULL },
+		{ "subdir/subdir_test1", "repoattr", GIT_ATTR_TRUE },
+		{ "subdir/subdir_test1", "rootattr", GIT_ATTR_TRUE },
+		{ "subdir/subdir_test1", "missingattr", NULL },
+		{ "subdir/subdir_test1", "subattr", "yes" },
+		{ "subdir/subdir_test1", "negattr", GIT_ATTR_FALSE },
+		{ "subdir/subdir_test1", "another", NULL },
+		{ "subdir/subdir_test2.txt", "repoattr", GIT_ATTR_TRUE },
+		{ "subdir/subdir_test2.txt", "rootattr", GIT_ATTR_TRUE },
+		{ "subdir/subdir_test2.txt", "missingattr", NULL },
+		{ "subdir/subdir_test2.txt", "subattr", "yes" },
+		{ "subdir/subdir_test2.txt", "negattr", GIT_ATTR_FALSE },
+		{ "subdir/subdir_test2.txt", "another", "one" },
+		{ NULL, NULL, NULL }
+	}, *scan;
+
+	for (scan = test_cases; scan->file != NULL; scan++) {
+		git_buf b = GIT_BUF_INIT;
+
+		git_buf_printf(&b, "%s:%s == expect %s",
+					   scan->file, scan->attr, scan->expected);
+
+		cl_must_pass_(
+			git_attr_get(g_repo, scan->file, scan->attr, &value) == GIT_SUCCESS,
+			b.ptr);
+
+		git_buf_printf(&b, ", got %s", value);
+
+		if (scan->expected == NULL ||
+			scan->expected == GIT_ATTR_TRUE ||
+			scan->expected == GIT_ATTR_FALSE)
+		{
+			cl_assert_(scan->expected == value, b.ptr);
+		} else {
+			cl_assert_strequal(scan->expected, value);
+		}
+
+		git_buf_free(&b);
+	}
+}
+
+void test_attr_repo__get_many(void)
+{
+	const char *names[4] = { "repoattr", "rootattr", "missingattr", "subattr" };
+	const char *values[4];
+
+	cl_git_pass(git_attr_get_many(g_repo, "root_test1", 4, names, values));
+
+	cl_assert(values[0] == GIT_ATTR_TRUE);
+	cl_assert(values[1] == GIT_ATTR_TRUE);
+	cl_assert(values[2] == NULL);
+	cl_assert(values[3] == NULL);
+
+	cl_git_pass(git_attr_get_many(g_repo, "root_test2", 4, names, values));
+
+	cl_assert(values[0] == GIT_ATTR_TRUE);
+	cl_assert(values[1] == GIT_ATTR_FALSE);
+	cl_assert(values[2] == NULL);
+	cl_assert(values[3] == NULL);
+
+	cl_git_pass(git_attr_get_many(g_repo, "subdir/subdir_test1", 4, names, values));
+
+	cl_assert(values[0] == GIT_ATTR_TRUE);
+	cl_assert(values[1] == GIT_ATTR_TRUE);
+	cl_assert(values[2] == NULL);
+	cl_assert_strequal("yes", values[3]);
+
+}
+
+static int count_attrs(
+	const char *GIT_UNUSED(name),
+	const char *GIT_UNUSED(value),
+	void *payload)
+{
+	GIT_UNUSED_ARG(name);
+	GIT_UNUSED_ARG(value);
+
+	*((int *)payload) += 1;
+
+	return GIT_SUCCESS;
+}
+
+void test_attr_repo__foreach(void)
+{
+	int count;
+
+	count = 0;
+	cl_git_pass(git_attr_foreach(g_repo, "root_test1", &count_attrs, &count));
+	cl_assert(count == 2);
+
+	count = 0;
+	cl_git_pass(git_attr_foreach(g_repo, "subdir/subdir_test1",
+		&count_attrs, &count));
+	cl_assert(count == 4); /* repoattr, rootattr, subattr, negattr */
+
+	count = 0;
+	cl_git_pass(git_attr_foreach(g_repo, "subdir/subdir_test2.txt",
+		&count_attrs, &count));
+	cl_assert(count == 5); /* repoattr, rootattr, subattr, negattr, another */
+}
+
+void test_attr_repo__manpage_example(void)
+{
+	const char *value;
+
+	cl_git_pass(git_attr_get(g_repo, "subdir/abc", "foo", &value));
+	cl_assert(value == GIT_ATTR_TRUE);
+
+	cl_git_pass(git_attr_get(g_repo, "subdir/abc", "bar", &value));
+	cl_assert(value == NULL);
+
+	cl_git_pass(git_attr_get(g_repo, "subdir/abc", "baz", &value));
+	cl_assert(value == GIT_ATTR_FALSE);
+
+	cl_git_pass(git_attr_get(g_repo, "subdir/abc", "merge", &value));
+	cl_assert_strequal("filfre", value);
+
+	cl_git_pass(git_attr_get(g_repo, "subdir/abc", "frotz", &value));
+	cl_assert(value == NULL);
+}
+
+void test_attr_repo__macros(void)
+{
+	const char *names[5] = { "rootattr", "binary", "diff", "crlf", "frotz" };
+	const char *names2[5] = { "mymacro", "positive", "negative", "rootattr", "another" };
+	const char *names3[3] = { "macro2", "multi2", "multi3" };
+	const char *values[5];
+
+	cl_git_pass(git_attr_get_many(g_repo, "binfile", 5, names, values));
+
+	cl_assert(values[0] == GIT_ATTR_TRUE);
+	cl_assert(values[1] == GIT_ATTR_TRUE);
+	cl_assert(values[2] == GIT_ATTR_FALSE);
+	cl_assert(values[3] == GIT_ATTR_FALSE);
+	cl_assert(values[4] == NULL);
+
+	cl_git_pass(git_attr_get_many(g_repo, "macro_test", 5, names2, values));
+
+	cl_assert(values[0] == GIT_ATTR_TRUE);
+	cl_assert(values[1] == GIT_ATTR_TRUE);
+	cl_assert(values[2] == GIT_ATTR_FALSE);
+	cl_assert(values[3] == NULL);
+	cl_assert_strequal("77", values[4]);
+
+	cl_git_pass(git_attr_get_many(g_repo, "macro_test", 3, names3, values));
+
+	cl_assert(values[0] == GIT_ATTR_TRUE);
+	cl_assert(values[1] == GIT_ATTR_FALSE);
+	cl_assert_strequal("answer", values[2]);
+}
+
+void test_attr_repo__bad_macros(void)
+{
+	const char *names[6] = { "rootattr", "positive", "negative",
+		"firstmacro", "secondmacro", "thirdmacro" };
+	const char *values[6];
+
+	cl_git_pass(git_attr_get_many(g_repo, "macro_bad", 6, names, values));
+
+	/* these three just confirm that the "mymacro" rule ran */
+	cl_assert(values[0] == NULL);
+	cl_assert(values[1] == GIT_ATTR_TRUE);
+	cl_assert(values[2] == GIT_ATTR_FALSE);
+
+	/* file contains:
+	 *     # let's try some malicious macro defs
+	 *     [attr]firstmacro -thirdmacro -secondmacro
+	 *     [attr]secondmacro firstmacro -firstmacro
+	 *     [attr]thirdmacro secondmacro=hahaha -firstmacro
+	 *     macro_bad firstmacro secondmacro thirdmacro
+	 *
+	 * firstmacro assignment list ends up with:
+	 *     -thirdmacro -secondmacro
+	 * secondmacro assignment list expands "firstmacro" and ends up with:
+	 *     -thirdmacro -secondmacro -firstmacro
+	 * thirdmacro assignment don't expand so list ends up with:
+	 *     secondmacro="hahaha"
+	 *
+	 * macro_bad assignment list ends up with:
+	 *     -thirdmacro -secondmacro firstmacro &&
+	 *     -thirdmacro -secondmacro -firstmacro secondmacro &&
+	 *     secondmacro="hahaha" thirdmacro
+	 *
+	 * so summary results should be:
+	 *     -firstmacro secondmacro="hahaha" thirdmacro
+	 */
+	cl_assert(values[3] == GIT_ATTR_FALSE);
+	cl_assert_strequal("hahaha", values[4]);
+	cl_assert(values[5] == GIT_ATTR_TRUE);
+}

tests-clay/clay.h

@@ -59,6 +59,23 @@ void cl_fixture_cleanup(const char *fixture_name);
  */
 extern void clay_on_init(void);
 extern void clay_on_shutdown(void);
+extern void test_attr_file__assign_variants(void);
+extern void test_attr_file__check_attr_examples(void);
+extern void test_attr_file__match_variants(void);
+extern void test_attr_file__simple_read(void);
+extern void test_attr_lookup__assign_variants(void);
+extern void test_attr_lookup__check_attr_examples(void);
+extern void test_attr_lookup__from_buffer(void);
+extern void test_attr_lookup__match_variants(void);
+extern void test_attr_lookup__simple(void);
+extern void test_attr_repo__bad_macros(void);
+extern void test_attr_repo__cleanup(void);
+extern void test_attr_repo__foreach(void);
+extern void test_attr_repo__get_many(void);
+extern void test_attr_repo__get_one(void);
+extern void test_attr_repo__initialize(void);
+extern void test_attr_repo__macros(void);
+extern void test_attr_repo__manpage_example(void);
 extern void test_buf_basic__printf(void);
 extern void test_buf_basic__resize(void);
 extern void test_config_add__cleanup(void);
@@ -125,6 +142,9 @@ extern void test_core_strtol__int64(void);
 extern void test_core_vector__0(void);
 extern void test_core_vector__1(void);
 extern void test_core_vector__2(void);
+extern void test_core_vector__3(void);
+extern void test_core_vector__4(void);
+extern void test_core_vector__5(void);
 extern void test_index_rename__single_file(void);
 extern void test_network_remotelocal__cleanup(void);
 extern void test_network_remotelocal__initialize(void);

tests-clay/clay_main.c

@@ -108,6 +108,27 @@ static int clay_sandbox(void);
 #define clay_on_suite() /* nop */
 
 /* Autogenerated test data by clay */
+static const struct clay_func _clay_cb_attr_file[] = {
+    {"assign_variants", &test_attr_file__assign_variants},
+	{"check_attr_examples", &test_attr_file__check_attr_examples},
+	{"match_variants", &test_attr_file__match_variants},
+	{"simple_read", &test_attr_file__simple_read}
+};
+static const struct clay_func _clay_cb_attr_lookup[] = {
+    {"assign_variants", &test_attr_lookup__assign_variants},
+	{"check_attr_examples", &test_attr_lookup__check_attr_examples},
+	{"from_buffer", &test_attr_lookup__from_buffer},
+	{"match_variants", &test_attr_lookup__match_variants},
+	{"simple", &test_attr_lookup__simple}
+};
+static const struct clay_func _clay_cb_attr_repo[] = {
+    {"bad_macros", &test_attr_repo__bad_macros},
+	{"foreach", &test_attr_repo__foreach},
+	{"get_many", &test_attr_repo__get_many},
+	{"get_one", &test_attr_repo__get_one},
+	{"macros", &test_attr_repo__macros},
+	{"manpage_example", &test_attr_repo__manpage_example}
+};
 static const struct clay_func _clay_cb_buf_basic[] = {
     {"printf", &test_buf_basic__printf},
 	{"resize", &test_buf_basic__resize}
@@ -194,7 +215,10 @@ static const struct clay_func _clay_cb_core_strtol[] = {
 static const struct clay_func _clay_cb_core_vector[] = {
     {"0", &test_core_vector__0},
 	{"1", &test_core_vector__1},
-	{"2", &test_core_vector__2}
+	{"2", &test_core_vector__2},
+	{"3", &test_core_vector__3},
+	{"4", &test_core_vector__4},
+	{"5", &test_core_vector__5}
 };
 static const struct clay_func _clay_cb_index_rename[] = {
     {"single_file", &test_index_rename__single_file}
@@ -309,6 +333,24 @@ static const struct clay_func _clay_cb_status_worktree[] = {
 
 static const struct clay_suite _clay_suites[] = {
     {
+        "attr::file",
+        {NULL, NULL},
+        {NULL, NULL},
+        _clay_cb_attr_file, 4
+    },
+	{
+        "attr::lookup",
+        {NULL, NULL},
+        {NULL, NULL},
+        _clay_cb_attr_lookup, 5
+    },
+	{
+        "attr::repo",
+        {"initialize", &test_attr_repo__initialize},
+        {"cleanup", &test_attr_repo__cleanup},
+        _clay_cb_attr_repo, 6
+    },
+	{
         "buf::basic",
         {NULL, NULL},
         {NULL, NULL},
@@ -396,7 +438,7 @@ static const struct clay_suite _clay_suites[] = {
         "core::vector",
         {NULL, NULL},
         {NULL, NULL},
-        _clay_cb_core_vector, 3
+        _clay_cb_core_vector, 6
     },
 	{
         "index::rename",
@@ -538,8 +580,8 @@ static const struct clay_suite _clay_suites[] = {
     }
 };
 
-static size_t _clay_suite_count = 38;
-static size_t _clay_callback_count = 122;
+static size_t _clay_suite_count = 41;
+static size_t _clay_callback_count = 140;
 
 /* Core test functions */
 static void

tests-clay/core/vector.c

@@ -64,3 +64,128 @@ void test_core_vector__2(void)
 }
 
 
+static int compare_them(const void *a, const void *b)
+{
+	return (int)((long)a - (long)b);
+}
+
+/* insert_sorted */
+void test_core_vector__3(void)
+{
+	git_vector x;
+	long i;
+	git_vector_init(&x, 1, &compare_them);
+
+	for (i = 0; i < 10; i += 2) {
+		git_vector_insert_sorted(&x, (void*)(i + 1), NULL);
+	}
+
+	for (i = 9; i > 0; i -= 2) {
+		git_vector_insert_sorted(&x, (void*)(i + 1), NULL);
+	}
+
+	cl_assert(x.length == 10);
+	for (i = 0; i < 10; ++i) {
+		cl_assert(git_vector_get(&x, i) == (void*)(i + 1));
+	}
+
+	git_vector_free(&x);
+}
+
+/* insert_sorted with duplicates */
+void test_core_vector__4(void)
+{
+	git_vector x;
+	long i;
+	git_vector_init(&x, 1, &compare_them);
+
+	for (i = 0; i < 10; i += 2) {
+		git_vector_insert_sorted(&x, (void*)(i + 1), NULL);
+	}
+
+	for (i = 9; i > 0; i -= 2) {
+		git_vector_insert_sorted(&x, (void*)(i + 1), NULL);
+	}
+
+	for (i = 0; i < 10; i += 2) {
+		git_vector_insert_sorted(&x, (void*)(i + 1), NULL);
+	}
+
+	for (i = 9; i > 0; i -= 2) {
+		git_vector_insert_sorted(&x, (void*)(i + 1), NULL);
+	}
+
+	cl_assert(x.length == 20);
+	for (i = 0; i < 20; ++i) {
+		cl_assert(git_vector_get(&x, i) == (void*)(i / 2 + 1));
+	}
+
+	git_vector_free(&x);
+}
+
+typedef struct {
+	int content;
+	int count;
+} my_struct;
+
+static int _struct_count = 0;
+
+static int compare_structs(const void *a, const void *b)
+{
+	return ((const my_struct *)a)->content -
+		((const my_struct *)b)->content;
+}
+
+static int merge_structs(void **old_raw, void *new)
+{
+	my_struct *old = *(my_struct **)old_raw;
+	cl_assert(((my_struct *)old)->content == ((my_struct *)new)->content);
+	((my_struct *)old)->count += 1;
+	git__free(new);
+	_struct_count--;
+	return GIT_EEXISTS;
+}
+
+static my_struct *alloc_struct(int value)
+{
+	my_struct *st = git__malloc(sizeof(my_struct));
+	st->content = value;
+	st->count = 0;
+	_struct_count++;
+	return st;
+}
+
+/* insert_sorted with duplicates and special handling */
+void test_core_vector__5(void)
+{
+	git_vector x;
+	int i;
+
+	git_vector_init(&x, 1, &compare_structs);
+
+	for (i = 0; i < 10; i += 2)
+		git_vector_insert_sorted(&x, alloc_struct(i), &merge_structs);
+
+	for (i = 9; i > 0; i -= 2)
+		git_vector_insert_sorted(&x, alloc_struct(i), &merge_structs);
+
+	cl_assert(x.length == 10);
+	cl_assert(_struct_count == 10);
+
+	for (i = 0; i < 10; i += 2)
+		git_vector_insert_sorted(&x, alloc_struct(i), &merge_structs);
+
+	for (i = 9; i > 0; i -= 2)
+		git_vector_insert_sorted(&x, alloc_struct(i), &merge_structs);
+
+	cl_assert(x.length == 10);
+	cl_assert(_struct_count == 10);
+
+	for (i = 0; i < 10; ++i) {
+		cl_assert(((my_struct *)git_vector_get(&x, i))->content == i);
+		git__free(git_vector_get(&x, i));
+		_struct_count--;
+	}
+
+	git_vector_free(&x);
+}

tests/resources/attr/.gitted/HEAD

+ref: refs/heads/master

tests/resources/attr/.gitted/config

+[core]
+	repositoryformatversion = 0
+	filemode = true
+	bare = false
+	logallrefupdates = true
+	ignorecase = true

tests/resources/attr/.gitted/description

+Unnamed repository; edit this file 'description' to name the repository.

tests/resources/attr/.gitted/info/attributes

+* repoattr
+a*	foo !bar -baz

tests/resources/attr/.gitted/info/exclude

+# git ls-files --others --exclude-from=.git/info/exclude
+# Lines that start with '#' are comments.
+# For a project mostly in C, the following would be a good set of
+# exclude patterns (uncomment them if you want to use them):
+# *.[oa]
+# *~

tests/resources/attr/.gitted/logs/HEAD

+0000000000000000000000000000000000000000 6bab5c79cd5140d0f800917f550eb2a3dc32b0da Russell Belfer <arrbee@arrbee.com> 1324416995 -0800	commit (initial): initial test data
+6bab5c79cd5140d0f800917f550eb2a3dc32b0da 605812ab7fe421fdd325a935d35cb06a9234a7d7 Russell Belfer <arrbee@arrbee.com> 1325143098 -0800	commit: latest test updates
+605812ab7fe421fdd325a935d35cb06a9234a7d7 a5d76cad53f66f1312bd995909a5bab3c0820770 Russell Belfer <arrbee@arrbee.com> 1325281762 -0800	commit: more macro tests

tests/resources/attr/.gitted/logs/refs/heads/master

+0000000000000000000000000000000000000000 6bab5c79cd5140d0f800917f550eb2a3dc32b0da Russell Belfer <arrbee@arrbee.com> 1324416995 -0800	commit (initial): initial test data
+6bab5c79cd5140d0f800917f550eb2a3dc32b0da 605812ab7fe421fdd325a935d35cb06a9234a7d7 Russell Belfer <arrbee@arrbee.com> 1325143098 -0800	commit: latest test updates
+605812ab7fe421fdd325a935d35cb06a9234a7d7 a5d76cad53f66f1312bd995909a5bab3c0820770 Russell Belfer <arrbee@arrbee.com> 1325281762 -0800	commit: more macro tests

tests/resources/attr/.gitted/refs/heads/master

+a5d76cad53f66f1312bd995909a5bab3c0820770

tests/resources/attr/attr0

+* binary

tests/resources/attr/attr1

+# a comment followed by some blank lines
+
+
+
+      # another comment that is indented
+
+# variations on fnmatch
+
+pat0 attr0
+!pat1 attr1
+pat2/ attr2
+pat3dir/pat3file attr3
+pat4.* attr4
+       *.pat5 attr5
+pat6/pat6/*.pat6 attr6
+
+pat7[a-e]??[xyz] attr7 # with a comment on the line
+
+pat8\ with\ spaces attr8
+
+   invalid # attr with no assignments doesn't count
+
+also/invalid              
+
+invalid.again/
+
+# next attr is at eof
+
+   pat9 attr9
\ No newline at end of file

tests/resources/attr/attr2

+
+# variations on assignments
+
+pat0 simple
+pat1 -neg
+* notundef
+pat2 !notundef
+pat3 assigned=test-value
+pat4 rule-with-more-chars=value-with-more-chars
+pat5 empty=
+pat6 -negempty=
+pat7 multiple -single values=1 also=a-really-long-value/* happy=yes!
+# the next line has trailing spaces
+pat8 again= another=12321                        
+patbad0 # empty assignment does not count
+# next line will be another simple empty assign that should not count
+            patbad1    
+
+# BTW I think there are 11 valid rules and two "invalid" empty ones
+
+pat9 -at-eof
\ No newline at end of file

tests/resources/attr/attr3

+# These are examples from the git-check-attr.1 man page
+*.java diff=java -crlf myAttr
+NoMyAttr.java !myAttr
+README caveat=unspecified

tests/resources/attr/binfile

+123
\ No newline at end of file

tests/resources/attr/gitattributes

+* rootattr
+root_test2 -rootattr
+root_test3 !rootattr
+binfile	   		   binary
+abc		foo bar baz
+
+root_test2 multiattr
+root_test3 multi2=foo
+
+root_test3 multiattr=1 multiattr=2 multiattr=3 multi2=abc !multi2
+root_test2 multiattr=string -multiattr
+
+[attr]mymacro positive -negative !rootattr
+macro* mymacro another=77
+
+[attr]macro2 multi2 -multi2 multi3 !multi3 multi3=answer
+macro* macro2 macro2 macro2
+
+# let's try some malicious macro defs
+[attr]firstmacro -thirdmacro -secondmacro
+[attr]secondmacro firstmacro -firstmacro
+[attr]thirdmacro secondmacro=hahaha
+
+macro_bad firstmacro secondmacro thirdmacro

tests/resources/attr/macro_bad

+boo

tests/resources/attr/macro_test

+Yo

tests/resources/attr/root_test1

+Hello from the root

tests/resources/attr/root_test2

+Hello from the root

tests/resources/attr/root_test3

+Hello from the root

tests/resources/attr/root_test4.txt

+Hello again

tests/resources/attr/subdir/.gitattributes

+* subattr=yes -negattr
+subdir/*.txt another=one
+ab*			 merge=filfre
+abc			 -foo -bar
+*.c			 frotz

tests/resources/attr/subdir/abc

+# Test file from gitattributes(5) example:
+
+If you have these three gitattributes file:
+
+   (in $GIT_DIR/info/attributes)
+
+	a*      foo !bar -baz
+
+	(in .gitattributes)
+	abc     foo bar baz
+
+	(in t/.gitattributes)
+	ab*     merge=filfre
+	abc     -foo -bar
+	*.c     frotz
+
+the attributes given to path t/abc are computed as follows:
+
+1. By examining t/.gitattributes (which is in the same directory as the path
+   in question), git finds that the first line matches. merge attribute is
+   set. It also finds that the second line matches, and attributes foo and
+   bar are unset.
+2. Then it examines .gitattributes (which is in the parent directory), and
+   finds that the first line matches, but t/.gitattributes file already
+   decided how merge, foo and bar attributes should be given to this path,
+   so it leaves foo and bar unset. Attribute baz is set.
+3. Finally it examines $GIT_DIR/info/attributes. This file is used to
+   override the in-tree settings. The first line is a match, and foo is set,
+   bar is reverted to unspecified state, and baz is unset.
+
+As the result, the attributes assignment to t/abc becomes:
+
+	foo     set to true
+	bar     unspecified
+	baz     set to false
+	merge   set to string value "filfre"
+	frotz   unspecified

tests/resources/attr/subdir/subdir_test1

+Hello from the subdir
+

tests/resources/attr/subdir/subdir_test2.txt

+Hello again

tests/resources/attr/subdir2/subdir2_test1

+Hello from subdir2