Merge remote-tracking branch 'carlosmn/indexer-stream' into new-error-handling

examples/network/index-pack.c

@@ -13,6 +13,61 @@ int index_cb(const git_indexer_stats *stats, void *data)
 
 int index_pack(git_repository *repo, int argc, char **argv)
 {
+  git_indexer_stream *idx;
+  git_indexer_stats stats = {0, 0};
+  int error, fd;
+  char hash[GIT_OID_HEXSZ + 1] = {0};
+  ssize_t read_bytes;
+  char buf[512];
+
+  if (argc < 2) {
+	  fprintf(stderr, "I need a packfile\n");
+	  return EXIT_FAILURE;
+  }
+
+  if (git_indexer_stream_new(&idx, ".git") < 0) {
+	  puts("bad idx");
+	  return -1;
+  }
+
+  if ((fd = open(argv[1], 0)) < 0) {
+	  perror("open");
+	  return -1;
+  }
+
+  do {
+	  read_bytes = read(fd, buf, sizeof(buf));
+	  if (read_bytes < 0)
+		  break;
+
+	  if ((error = git_indexer_stream_add(idx, buf, read_bytes, &stats)) < 0)
+		  goto cleanup;
+
+	  printf("\rIndexing %d of %d", stats.processed, stats.total);
+  } while (read_bytes > 0);
+
+  if (read_bytes < 0) {
+	  error = -1;
+	  perror("failed reading");
+	  goto cleanup;
+  }
+
+  if ((error = git_indexer_stream_finalize(idx, &stats)) < 0)
+	  goto cleanup;
+
+  printf("\rIndexing %d of %d\n", stats.processed, stats.total);
+
+  git_oid_fmt(hash, git_indexer_stream_hash(idx));
+  puts(hash);
+
+cleanup:
+  close(fd);
+  git_indexer_stream_free(idx);
+  return error;
+}
+
+int index_pack_old(git_repository *repo, int argc, char **argv)
+{
   git_indexer *indexer;
   git_indexer_stats stats;
   int error;

include/git2/indexer.h

@@ -23,6 +23,51 @@ typedef struct git_indexer_stats {
 
 
 typedef struct git_indexer git_indexer;
+typedef struct git_indexer_stream git_indexer_stream;
+
+/**
+ * Create a new streaming indexer instance
+ *
+ * @param out where to store the inexer instance
+ * @param path to the gitdir (metadata directory)
+ */
+GIT_EXTERN(int) git_indexer_stream_new(git_indexer_stream **out, const char *gitdir);
+
+/**
+ * Add data to the indexer
+ *
+ * @param idx the indexer
+ * @param data the data to add
+ * @param size the size of the data
+ * @param stats stat storage
+ */
+GIT_EXTERN(int) git_indexer_stream_add(git_indexer_stream *idx, void *data, size_t size, git_indexer_stats *stats);
+
+/**
+ * Finalize the pack and index
+ *
+ * Resolve any pending deltas and write out the index file
+ *
+ * @param idx the indexer
+ */
+GIT_EXTERN(int) git_indexer_stream_finalize(git_indexer_stream *idx, git_indexer_stats *stats);
+
+/**
+ * Get the packfile's hash
+ *
+ * A packfile's name is derived from the sorted hashing of all object
+ * names. This is only correct after the index has been finalized.
+ *
+ * @param idx the indexer instance
+ */
+GIT_EXTERN(const git_oid *) git_indexer_stream_hash(git_indexer_stream *idx);
+
+/**
+ * Free the indexer and its resources
+ *
+ * @param idx the indexer to free
+ */
+GIT_EXTERN(void) git_indexer_stream_free(git_indexer_stream *idx);
 
 /**
  * Create a new indexer instance

src/filebuf.c

@@ -194,14 +194,19 @@ int git_filebuf_open(git_filebuf *file, const char *path, int flags)
 
 	memset(file, 0x0, sizeof(git_filebuf));
 
+	if (flags & GIT_FILEBUF_DO_NOT_BUFFER)
+		file->do_not_buffer = true;
+
 	file->buf_size = WRITE_BUFFER_SIZE;
 	file->buf_pos = 0;
 	file->fd = -1;
 	file->last_error = BUFERR_OK;
 
 	/* Allocate the main cache buffer */
+	if (!file->do_not_buffer) {
 		file->buffer = git__malloc(file->buf_size);
 		GITERR_CHECK_ALLOC(file->buffer);
+	}
 
 	/* If we are hashing on-write, allocate a new hash context */
 	if (flags & GIT_FILEBUF_HASH_CONTENTS) {
@@ -345,6 +350,9 @@ int git_filebuf_write(git_filebuf *file, const void *buff, size_t len)
 
 	ENSURE_BUF_OK(file);
 
+	if (file->do_not_buffer)
+		return file->write(file, (void *)buff, len);
+
 	for (;;) {
 		size_t space_left = file->buf_size - file->buf_pos;
 

src/filebuf.h

@@ -19,7 +19,8 @@
 #define GIT_FILEBUF_APPEND				(1 << 2)
 #define GIT_FILEBUF_FORCE				(1 << 3)
 #define GIT_FILEBUF_TEMPORARY			(1 << 4)
-#define GIT_FILEBUF_DEFLATE_SHIFT		(5)
+#define GIT_FILEBUF_DO_NOT_BUFFER		(1 << 5)
+#define GIT_FILEBUF_DEFLATE_SHIFT		(6)
 
 #define GIT_FILELOCK_EXTENSION ".lock\0"
 #define GIT_FILELOCK_EXTLENGTH 6
@@ -41,6 +42,7 @@ struct git_filebuf {
 	size_t buf_size, buf_pos;
 	git_file fd;
 	bool fd_is_open;
+	bool do_not_buffer;
 	int last_error;
 };
 

src/mwindow.c

@@ -89,6 +89,7 @@ void git_mwindow_scan_lru(
 {
 	git_mwindow *w, *w_l;
 
+	puts("LRU");
 	for (w_l = NULL, w = mwf->windows; w; w = w->next) {
 		if (!w->inuse_cnt) {
 			/*
@@ -210,15 +211,17 @@ unsigned char *git_mwindow_open(
 	git_mwindow_ctl *ctl = &GIT_GLOBAL->mem_ctl;
 	git_mwindow *w = *cursor;
 
+
 	if (!w || !git_mwindow_contains(w, offset + extra)) {
 		if (w) {
 			w->inuse_cnt--;
 		}
 
 		for (w = mwf->windows; w; w = w->next) {
-			if (git_mwindow_contains(w, offset + extra))
+			if (git_mwindow_contains(w, offset + extra)) {
 				break;
 			}
+		}
 
 		/*
 		 * If there isn't a suitable window, we need to create a new
@@ -246,6 +249,7 @@ unsigned char *git_mwindow_open(
 	if (left)
 		*left = (unsigned int)(w->window_map.len - offset);
 
+	fflush(stdout);
 	return (unsigned char *) w->window_map.data + offset;
 }
 

src/pack.c

@@ -200,7 +200,8 @@ static unsigned char *pack_window_open(
 	return git_mwindow_open(&p->mwf, w_cursor, offset, 20, left);
  }
 
-static unsigned long packfile_unpack_header1(
+static int packfile_unpack_header1(
+		unsigned long *usedp,
 		size_t *sizep,
 		git_otype *type,
 		const unsigned char *buf,
@@ -215,8 +216,13 @@ static unsigned long packfile_unpack_header1(
 	size = c & 15;
 	shift = 4;
 	while (c & 0x80) {
-		if (len <= used || bitsizeof(long) <= shift)
-			return 0;
+		if (len <= used)
+			return GIT_ESHORTBUFFER;
+
+		if (bitsizeof(long) <= shift) {
+			*usedp = 0;
+			return -1;
+		}
 
 		c = buf[used++];
 		size += (c & 0x7f) << shift;
@@ -224,7 +230,8 @@ static unsigned long packfile_unpack_header1(
 	}
 
 	*sizep = (size_t)size;
-	return used;
+	*usedp = used;
+	return 0;
 }
 
 int git_packfile_unpack_header(
@@ -237,6 +244,7 @@ int git_packfile_unpack_header(
 	unsigned char *base;
 	unsigned int left;
 	unsigned long used;
+	int ret;
 
 	/* pack_window_open() assures us we have [base, base + 20) available
 	 * as a range that we can look at at. (Its actually the hash
@@ -247,10 +255,13 @@ int git_packfile_unpack_header(
 //	base = pack_window_open(p, w_curs, *curpos, &left);
 	base = git_mwindow_open(mwf, w_curs, *curpos, 20, &left);
 	if (base == NULL)
-		return -1;
+		return GIT_ESHORTBUFFER;
 
-	used = packfile_unpack_header1(size_p, type_p, base, left);
-	if (used == 0)
+	ret = packfile_unpack_header1(&used, size_p, type_p, base, left);
+	git_mwindow_close(w_curs);
+	if (ret == GIT_ESHORTBUFFER)
+		return ret;
+	else if (ret < 0)
 		return packfile_error("header length is zero");
 
 	*curpos += used;
@@ -271,12 +282,12 @@ static int packfile_unpack_delta(
 	int error;
 
 	base_offset = get_delta_base(p, w_curs, curpos, delta_type, obj_offset);
+	git_mwindow_close(w_curs);
 	if (base_offset == 0)
 		return packfile_error("delta offset is zero");
 	if (base_offset < 0) /* must actually be an error code */
 		return (int)base_offset;
 
-	git_mwindow_close(w_curs);
 	error = git_packfile_unpack(&base, p, &base_offset);
 
 	/*
@@ -289,6 +300,7 @@ static int packfile_unpack_delta(
 		return error;
 
 	error = packfile_unpack_compressed(&delta, p, w_curs, curpos, delta_size, delta_type);
+	git_mwindow_close(w_curs);
 	if (error < 0) {
 		git__free(base.data);
 		return error;
@@ -327,6 +339,8 @@ int git_packfile_unpack(
 	obj->type = GIT_OBJ_BAD;
 
 	error = git_packfile_unpack_header(&size, &type, &p->mwf, &w_curs, &curpos);
+	git_mwindow_close(&w_curs);
+
 	if (error < 0)
 		return error;
 
@@ -352,8 +366,6 @@ int git_packfile_unpack(
 		break;
 	}
 
-	git_mwindow_close(&w_curs);
-
 	*obj_offset = curpos;
 	return error;
 }
@@ -381,6 +393,7 @@ int packfile_unpack_compressed(
 	if (st != Z_OK) {
 		git__free(buffer);
 		giterr_set(GITERR_ZLIB, "Failed to inflate packfile");
+
 		return -1;
 	}
 
@@ -388,10 +401,17 @@ int packfile_unpack_compressed(
 		in = pack_window_open(p, w_curs, *curpos, &stream.avail_in);
 		stream.next_in = in;
 		st = inflate(&stream, Z_FINISH);
+		git_mwindow_close(w_curs);
 
 		if (!stream.avail_out)
 			break; /* the payload is larger than it should be */
 
+		if (st == Z_BUF_ERROR && in == NULL) {
+			inflateEnd(&stream);
+			git__free(buffer);
+			return GIT_ESHORTBUFFER;
+		}
+
 		*curpos += stream.next_in - in;
 	} while (st == Z_OK || st == Z_BUF_ERROR);
 
@@ -420,10 +440,15 @@ git_off_t get_delta_base(
 	git_otype type,
 	git_off_t delta_obj_offset)
 {
-	unsigned char *base_info = pack_window_open(p, w_curs, *curpos, NULL);
+	unsigned int left = 0;
+	unsigned char *base_info;
 	git_off_t base_offset;
 	git_oid unused;
 
+	base_info = pack_window_open(p, w_curs, *curpos, &left);
+	/* Assumption: the only reason this would fail is because the file is too small */
+	if (base_info == NULL)
+		return GIT_ESHORTBUFFER;
 	/* pack_window_open() assured us we have [base_info, base_info + 20)
 	 * as a range that we can look at without walking off the
 	 * end of the mapped window. Its actually the hash size
@@ -435,6 +460,8 @@ git_off_t get_delta_base(
 		unsigned char c = base_info[used++];
 		base_offset = c & 127;
 		while (c & 128) {
+			if (left <= used)
+				return GIT_ESHORTBUFFER;
 			base_offset += 1;
 			if (!base_offset || MSB(base_offset, 7))
 				return 0; /* overflow */

src/pack.h

@@ -83,6 +83,13 @@ int git_packfile_unpack_header(
 		git_off_t *curpos);
 
 int git_packfile_unpack(git_rawobj *obj, struct git_pack_file *p, git_off_t *obj_offset);
+int packfile_unpack_compressed(
+	git_rawobj *obj,
+	struct git_pack_file *p,
+	git_mwindow **w_curs,
+	git_off_t *curpos,
+	size_t size,
+	git_otype type);
 
 git_off_t get_delta_base(struct git_pack_file *p, git_mwindow **w_curs,
 		git_off_t *curpos, git_otype type,