Add "as_path" parameters to blob and buffer diffs

This adds parameters to the four functions that allow for blob-to-
blob and blob-to-buffer differencing (either via callbacks or by
making a git_diff_patch object).  These parameters let you say
that filename we should pretend the blob has while doing the diff.
If you pass NULL, there should be no change from the existing
behavior, which is to skip using attributes for file type checks
and just look at content.  With the parameters, you can plug into
the new diff driver functionality and get binary or non-binary
behavior, plus function context regular expressions, etc.

This commit also fixes things so that the git_diff_delta that is
generated by these functions will actually be populated with the
data that we know about the blobs (or buffers) so you can use it
appropriately.  It also fixes a bug in generating patches from
the git_diff_patch objects created via these functions.

Lastly, there is one other behavior change that may matter.  If
there is no difference between the two blobs, these functions no
longer generate any diff callbacks / patches unless you have
passed in GIT_DIFF_INCLUDE_UNMODIFIED.  This is pretty natural,
but could potentially change the behavior of existing usage.

1 files changed, 83 insertions, 77 deletions

diff --git a/src/diff_file.c b/src/diff_file.c
index 4fd1177ae..9d06daafa 100644
--- a/src/diff_file.c
+++ b/src/diff_file.c
@@ -18,23 +18,23 @@
 static bool diff_file_content_binary_by_size(git_diff_file_content *fc)
 {
 	/* if we have diff opts, check max_size vs file size */
-	if ((fc->file.flags & DIFF_FLAGS_KNOWN_BINARY) == 0 &&
+	if ((fc->file->flags & DIFF_FLAGS_KNOWN_BINARY) == 0 &&
 		fc->opts_max_size > 0 &&
-		fc->file.size > fc->opts_max_size)
-		fc->file.flags |= GIT_DIFF_FLAG_BINARY;
+		fc->file->size > fc->opts_max_size)
+		fc->file->flags |= GIT_DIFF_FLAG_BINARY;
 
-	return ((fc->file.flags & GIT_DIFF_FLAG_BINARY) != 0);
+	return ((fc->file->flags & GIT_DIFF_FLAG_BINARY) != 0);
 }
 
 static void diff_file_content_binary_by_content(git_diff_file_content *fc)
 {
-	if ((fc->file.flags & DIFF_FLAGS_KNOWN_BINARY) != 0)
+	if ((fc->file->flags & DIFF_FLAGS_KNOWN_BINARY) != 0)
 		return;
 
 	switch (git_diff_driver_content_is_binary(
 		fc->driver, fc->map.data, fc->map.len)) {
-	case 0: fc->file.flags |= GIT_DIFF_FLAG_NOT_BINARY; break;
-	case 1: fc->file.flags |= GIT_DIFF_FLAG_BINARY; break;
+	case 0: fc->file->flags |= GIT_DIFF_FLAG_NOT_BINARY; break;
+	case 1: fc->file->flags |= GIT_DIFF_FLAG_BINARY; break;
 	default: break;
 	}
 }
@@ -48,38 +48,39 @@ static int diff_file_content_init_common(
 		fc->opts_max_size = opts->max_size ?
 			opts->max_size : DIFF_MAX_FILESIZE;
 
-	if (!fc->driver) {
-		if (git_diff_driver_lookup(&fc->driver, fc->repo, "") < 0)
-			return -1;
+	if (fc->src == GIT_ITERATOR_TYPE_EMPTY)
 		fc->src = GIT_ITERATOR_TYPE_TREE;
-	}
+
+	if (!fc->driver &&
+		git_diff_driver_lookup(&fc->driver, fc->repo, fc->file->path) < 0)
+		return -1;
 
 	/* give driver a chance to modify options */
 	git_diff_driver_update_options(&fc->opts_flags, fc->driver);
 
 	/* make sure file is conceivable mmap-able */
-	if ((git_off_t)((size_t)fc->file.size) != fc->file.size)
-		fc->file.flags |= GIT_DIFF_FLAG_BINARY;
+	if ((git_off_t)((size_t)fc->file->size) != fc->file->size)
+		fc->file->flags |= GIT_DIFF_FLAG_BINARY;
 	/* check if user is forcing text diff the file */
 	else if (fc->opts_flags & GIT_DIFF_FORCE_TEXT) {
-		fc->file.flags &= ~GIT_DIFF_FLAG_BINARY;
-		fc->file.flags |= GIT_DIFF_FLAG_NOT_BINARY;
+		fc->file->flags &= ~GIT_DIFF_FLAG_BINARY;
+		fc->file->flags |= GIT_DIFF_FLAG_NOT_BINARY;
 	}
 	/* check if user is forcing binary diff the file */
 	else if (fc->opts_flags & GIT_DIFF_FORCE_BINARY) {
-		fc->file.flags &= ~GIT_DIFF_FLAG_NOT_BINARY;
-		fc->file.flags |= GIT_DIFF_FLAG_BINARY;
+		fc->file->flags &= ~GIT_DIFF_FLAG_NOT_BINARY;
+		fc->file->flags |= GIT_DIFF_FLAG_BINARY;
 	}
 
 	diff_file_content_binary_by_size(fc);
 
-	if ((fc->file.flags & GIT_DIFF_FLAG__NO_DATA) != 0) {
-		fc->file.flags |= GIT_DIFF_FLAG__LOADED;
+	if ((fc->flags & GIT_DIFF_FLAG__NO_DATA) != 0) {
+		fc->flags |= GIT_DIFF_FLAG__LOADED;
 		fc->map.len  = 0;
 		fc->map.data = "";
 	}
 
-	if ((fc->file.flags & GIT_DIFF_FLAG__LOADED) != 0)
+	if ((fc->flags & GIT_DIFF_FLAG__LOADED) != 0)
 		diff_file_content_binary_by_content(fc);
 
 	return 0;
@@ -92,15 +93,14 @@ int git_diff_file_content__init_from_diff(
 	bool use_old)
 {
 	git_diff_delta *delta = git_vector_get(&diff->deltas, delta_index);
-	git_diff_file *file = use_old ? &delta->old_file : &delta->new_file;
 	bool has_data = true;
 
 	memset(fc, 0, sizeof(*fc));
 	fc->repo = diff->repo;
+	fc->file = use_old ? &delta->old_file : &delta->new_file;
 	fc->src  = use_old ? diff->old_src : diff->new_src;
-	memcpy(&fc->file, file, sizeof(fc->file));
 
-	if (git_diff_driver_lookup(&fc->driver, fc->repo, file->path) < 0)
+	if (git_diff_driver_lookup(&fc->driver, fc->repo, fc->file->path) < 0)
 		return -1;
 
 	switch (delta->status) {
@@ -122,7 +122,7 @@ int git_diff_file_content__init_from_diff(
 	}
 
 	if (!has_data)
-		fc->file.flags |= GIT_DIFF_FLAG__NO_DATA;
+		fc->flags |= GIT_DIFF_FLAG__NO_DATA;
 
 	return diff_file_content_init_common(fc, &diff->opts);
 }
@@ -131,21 +131,24 @@ int git_diff_file_content__init_from_blob(
 	git_diff_file_content *fc,
 	git_repository *repo,
 	const git_diff_options *opts,
-	const git_blob *blob)
+	const git_blob *blob,
+	git_diff_file *as_file)
 {
 	memset(fc, 0, sizeof(*fc));
 	fc->repo = repo;
+	fc->file = as_file;
 	fc->blob = blob;
 
 	if (!blob) {
-		fc->file.flags |= GIT_DIFF_FLAG__NO_DATA;
+		fc->flags |= GIT_DIFF_FLAG__NO_DATA;
 	} else {
-		fc->file.flags |= GIT_DIFF_FLAG__LOADED | GIT_DIFF_FLAG_VALID_OID;
-		fc->file.size = git_blob_rawsize(blob);
-		fc->file.mode = 0644;
-		git_oid_cpy(&fc->file.oid, git_blob_id(blob));
+		fc->flags |= GIT_DIFF_FLAG__LOADED;
+		fc->file->flags |= GIT_DIFF_FLAG_VALID_OID;
+		fc->file->size = git_blob_rawsize(blob);
+		fc->file->mode = GIT_FILEMODE_BLOB;
+		git_oid_cpy(&fc->file->oid, git_blob_id(blob));
 
-		fc->map.len = (size_t)fc->file.size;
+		fc->map.len  = (size_t)fc->file->size;
 		fc->map.data = (char *)git_blob_rawcontent(blob);
 	}
 
@@ -157,18 +160,21 @@ int git_diff_file_content__init_from_raw(
 	git_repository *repo,
 	const git_diff_options *opts,
 	const char *buf,
-	size_t buflen)
+	size_t buflen,
+	git_diff_file *as_file)
 {
 	memset(fc, 0, sizeof(*fc));
 	fc->repo = repo;
+	fc->file = as_file;
 
 	if (!buf) {
-		fc->file.flags |= GIT_DIFF_FLAG__NO_DATA;
+		fc->flags |= GIT_DIFF_FLAG__NO_DATA;
 	} else {
-		fc->file.flags |= GIT_DIFF_FLAG__LOADED | GIT_DIFF_FLAG_VALID_OID;
-		fc->file.size = buflen;
-		fc->file.mode = 0644;
-		git_odb_hash(&fc->file.oid, buf, buflen, GIT_OBJ_BLOB);
+		fc->flags |= GIT_DIFF_FLAG__LOADED;
+		fc->file->flags |= GIT_DIFF_FLAG_VALID_OID;
+		fc->file->size = buflen;
+		fc->file->mode = GIT_FILEMODE_BLOB;
+		git_odb_hash(&fc->file->oid, buf, buflen, GIT_OBJ_BLOB);
 
 		fc->map.len  = buflen;
 		fc->map.data = (char *)buf;
@@ -190,7 +196,7 @@ static int diff_file_content_commit_to_str(
 		unsigned int sm_status = 0;
 		const git_oid *sm_head;
 
-		if ((error = git_submodule_lookup(&sm, fc->repo, fc->file.path)) < 0 ||
+		if ((error = git_submodule_lookup(&sm, fc->repo, fc->file->path)) < 0 ||
 			(error = git_submodule_status(&sm_status, sm)) < 0) {
 			/* GIT_EEXISTS means a "submodule" that has not been git added */
 			if (error == GIT_EEXISTS)
@@ -199,25 +205,25 @@ static int diff_file_content_commit_to_str(
 		}
 
 		/* update OID if we didn't have it previously */
-		if ((fc->file.flags & GIT_DIFF_FLAG_VALID_OID) == 0 &&
+		if ((fc->file->flags & GIT_DIFF_FLAG_VALID_OID) == 0 &&
 			((sm_head = git_submodule_wd_id(sm)) != NULL ||
 			 (sm_head = git_submodule_head_id(sm)) != NULL))
 		{
-			git_oid_cpy(&fc->file.oid, sm_head);
-			fc->file.flags |= GIT_DIFF_FLAG_VALID_OID;
+			git_oid_cpy(&fc->file->oid, sm_head);
+			fc->file->flags |= GIT_DIFF_FLAG_VALID_OID;
 		}
 
 		if (GIT_SUBMODULE_STATUS_IS_WD_DIRTY(sm_status))
 			status = "-dirty";
 	}
 
-	git_oid_tostr(oid, sizeof(oid), &fc->file.oid);
+	git_oid_tostr(oid, sizeof(oid), &fc->file->oid);
 	if (git_buf_printf(&content, "Subproject commit %s%s\n", oid, status) < 0)
 		return -1;
 
 	fc->map.len  = git_buf_len(&content);
 	fc->map.data = git_buf_detach(&content);
-	fc->file.flags |= GIT_DIFF_FLAG__FREE_DATA;
+	fc->flags |= GIT_DIFF_FLAG__FREE_DATA;
 
 	return 0;
 }
@@ -227,27 +233,27 @@ static int diff_file_content_load_blob(git_diff_file_content *fc)
 	int error = 0;
 	git_odb_object *odb_obj = NULL;
 
-	if (git_oid_iszero(&fc->file.oid))
+	if (git_oid_iszero(&fc->file->oid))
 		return 0;
 
-	if (fc->file.mode == GIT_FILEMODE_COMMIT)
+	if (fc->file->mode == GIT_FILEMODE_COMMIT)
 		return diff_file_content_commit_to_str(fc, false);
 
 	/* if we don't know size, try to peek at object header first */
-	if (!fc->file.size) {
+	if (!fc->file->size) {
 		git_odb *odb;
 		size_t len;
 		git_otype type;
 
 		if (!(error = git_repository_odb__weakptr(&odb, fc->repo))) {
 			error = git_odb__read_header_or_object(
-				&odb_obj, &len, &type, odb, &fc->file.oid);
+				&odb_obj, &len, &type, odb, &fc->file->oid);
 			git_odb_free(odb);
 		}
 		if (error)
 			return error;
 
-		fc->file.size = len;
+		fc->file->size = len;
 	}
 
 	if (diff_file_content_binary_by_size(fc))
@@ -259,11 +265,11 @@ static int diff_file_content_load_blob(git_diff_file_content *fc)
 		git_odb_object_free(odb_obj);
 	} else {
 		error = git_blob_lookup(
-			(git_blob **)&fc->blob, fc->repo, &fc->file.oid);
+			(git_blob **)&fc->blob, fc->repo, &fc->file->oid);
 	}
 
 	if (!error) {
-		fc->file.flags |= GIT_DIFF_FLAG__FREE_BLOB;
+		fc->flags |= GIT_DIFF_FLAG__FREE_BLOB;
 		fc->map.data = (void *)git_blob_rawcontent(fc->blob);
 		fc->map.len  = (size_t)git_blob_rawsize(fc->blob);
 	}
@@ -279,16 +285,16 @@ static int diff_file_content_load_workdir_symlink(
 	/* link path on disk could be UTF-16, so prepare a buffer that is
 	 * big enough to handle some UTF-8 data expansion
 	 */
-	alloc_len = (ssize_t)(fc->file.size * 2) + 1;
+	alloc_len = (ssize_t)(fc->file->size * 2) + 1;
 
 	fc->map.data = git__calloc(alloc_len, sizeof(char));
 	GITERR_CHECK_ALLOC(fc->map.data);
 
-	fc->file.flags |= GIT_DIFF_FLAG__FREE_DATA;
+	fc->flags |= GIT_DIFF_FLAG__FREE_DATA;
 
 	read_len = p_readlink(git_buf_cstr(path), fc->map.data, alloc_len);
 	if (read_len < 0) {
-		giterr_set(GITERR_OS, "Failed to read symlink '%s'", fc->file.path);
+		giterr_set(GITERR_OS, "Failed to read symlink '%s'", fc->file->path);
 		return -1;
 	}
 
@@ -307,28 +313,28 @@ static int diff_file_content_load_workdir_file(
 	if (fd < 0)
 		return fd;
 
-	if (!fc->file.size &&
-		!(fc->file.size = git_futils_filesize(fd)))
+	if (!fc->file->size &&
+		!(fc->file->size = git_futils_filesize(fd)))
 		goto cleanup;
 
 	if (diff_file_content_binary_by_size(fc))
 		goto cleanup;
 
 	if ((error = git_filters_load(
-			&filters, fc->repo, fc->file.path, GIT_FILTER_TO_ODB)) < 0)
+			&filters, fc->repo, fc->file->path, GIT_FILTER_TO_ODB)) < 0)
 		goto cleanup;
 	/* error >= is a filter count */
 
 	if (error == 0) {
 		if (!(error = git_futils_mmap_ro(
-				&fc->map, fd, 0, (size_t)fc->file.size)))
-			fc->file.flags |= GIT_DIFF_FLAG__UNMAP_DATA;
+				&fc->map, fd, 0, (size_t)fc->file->size)))
+			fc->flags |= GIT_DIFF_FLAG__UNMAP_DATA;
 		else /* fall through to try readbuffer below */
 			giterr_clear();
 	}
 
 	if (error != 0) {
-		error = git_futils_readbuffer_fd(&raw, fd, (size_t)fc->file.size);
+		error = git_futils_readbuffer_fd(&raw, fd, (size_t)fc->file->size);
 		if (error < 0)
 			goto cleanup;
 
@@ -340,7 +346,7 @@ static int diff_file_content_load_workdir_file(
 		if (!error) {
 			fc->map.len  = git_buf_len(&filtered);
 			fc->map.data = git_buf_detach(&filtered);
-			fc->file.flags |= GIT_DIFF_FLAG__FREE_DATA;
+			fc->flags |= GIT_DIFF_FLAG__FREE_DATA;
 		}
 
 		git_buf_free(&raw);
@@ -359,26 +365,26 @@ static int diff_file_content_load_workdir(git_diff_file_content *fc)
 	int error = 0;
 	git_buf path = GIT_BUF_INIT;
 
-	if (fc->file.mode == GIT_FILEMODE_COMMIT)
+	if (fc->file->mode == GIT_FILEMODE_COMMIT)
 		return diff_file_content_commit_to_str(fc, true);
 
-	if (fc->file.mode == GIT_FILEMODE_TREE)
+	if (fc->file->mode == GIT_FILEMODE_TREE)
 		return 0;
 
 	if (git_buf_joinpath(
-			&path, git_repository_workdir(fc->repo), fc->file.path) < 0)
+			&path, git_repository_workdir(fc->repo), fc->file->path) < 0)
 		return -1;
 
-	if (S_ISLNK(fc->file.mode))
+	if (S_ISLNK(fc->file->mode))
 		error = diff_file_content_load_workdir_symlink(fc, &path);
 	else
 		error = diff_file_content_load_workdir_file(fc, &path);
 
 	/* once data is loaded, update OID if we didn't have it previously */
-	if (!error && (fc->file.flags & GIT_DIFF_FLAG_VALID_OID) == 0) {
+	if (!error && (fc->file->flags & GIT_DIFF_FLAG_VALID_OID) == 0) {
 		error = git_odb_hash(
-			&fc->file.oid, fc->map.data, fc->map.len, GIT_OBJ_BLOB);
-		fc->file.flags |= GIT_DIFF_FLAG_VALID_OID;
+			&fc->file->oid, fc->map.data, fc->map.len, GIT_OBJ_BLOB);
+		fc->file->flags |= GIT_DIFF_FLAG_VALID_OID;
 	}
 
 	git_buf_free(&path);
@@ -389,10 +395,10 @@ int git_diff_file_content__load(git_diff_file_content *fc)
 {
 	int error = 0;
 
-	if ((fc->file.flags & GIT_DIFF_FLAG__LOADED) != 0)
+	if ((fc->flags & GIT_DIFF_FLAG__LOADED) != 0)
 		return 0;
 
-	if (fc->file.flags & GIT_DIFF_FLAG_BINARY)
+	if ((fc->file->flags & GIT_DIFF_FLAG_BINARY) != 0)
 		return 0;
 
 	if (fc->src == GIT_ITERATOR_TYPE_WORKDIR)
@@ -402,7 +408,7 @@ int git_diff_file_content__load(git_diff_file_content *fc)
 	if (error)
 		return error;
 
-	fc->file.flags |= GIT_DIFF_FLAG__LOADED;
+	fc->flags |= GIT_DIFF_FLAG__LOADED;
 
 	diff_file_content_binary_by_content(fc);
 
@@ -411,26 +417,26 @@ int git_diff_file_content__load(git_diff_file_content *fc)
 
 void git_diff_file_content__unload(git_diff_file_content *fc)
 {
-	if (fc->file.flags & GIT_DIFF_FLAG__FREE_DATA) {
+	if (fc->flags & GIT_DIFF_FLAG__FREE_DATA) {
 		git__free(fc->map.data);
 		fc->map.data = "";
 		fc->map.len  = 0;
-		fc->file.flags &= ~GIT_DIFF_FLAG__FREE_DATA;
+		fc->flags &= ~GIT_DIFF_FLAG__FREE_DATA;
 	}
-	else if (fc->file.flags & GIT_DIFF_FLAG__UNMAP_DATA) {
+	else if (fc->flags & GIT_DIFF_FLAG__UNMAP_DATA) {
 		git_futils_mmap_free(&fc->map);
 		fc->map.data = "";
 		fc->map.len  = 0;
-		fc->file.flags &= ~GIT_DIFF_FLAG__UNMAP_DATA;
+		fc->flags &= ~GIT_DIFF_FLAG__UNMAP_DATA;
 	}
 
-	if (fc->file.flags & GIT_DIFF_FLAG__FREE_BLOB) {
+	if (fc->flags & GIT_DIFF_FLAG__FREE_BLOB) {
 		git_blob_free((git_blob *)fc->blob);
 		fc->blob = NULL;
-		fc->file.flags &= ~GIT_DIFF_FLAG__FREE_BLOB;
+		fc->flags &= ~GIT_DIFF_FLAG__FREE_BLOB;
 	}
 
-	fc->file.flags &= ~GIT_DIFF_FLAG__LOADED;
+	fc->flags &= ~GIT_DIFF_FLAG__LOADED;
 }
 
 void git_diff_file_content__clear(git_diff_file_content *fc)