fast-import: let importers retrieve blobs

New objects written by fast-import are not available immediately.
Until a checkpoint has been started and finishes writing the pack
index, any new blobs will not be accessible using standard git tools.

So introduce a new way to access them: a "cat-blob" command in the
command stream requests for fast-import to print a blob to stdout or a
file descriptor specified by the argument to --cat-blob-fd.  The value
for cat-blob-fd cannot be specified in the stream because that would
be a layering violation: the decision of where to direct a stream has
to be made when fast-import is started anyway, so we might as well
make the stream format is independent of that detail.

Output uses the same format as "git cat-file --batch".

Thanks to Sverre Rabbelier and Sam Vilain for guidance in designing
the protocol.

Based-on-patch-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: David Barr <david.barr@cordelta.com>
Acked-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>

fast-import.c

@@ -55,6 +55,8 @@ Format of STDIN stream:
     ('from' sp committish lf)?
     lf?;
 
+  cat_blob ::= 'cat-blob' sp (hexsha1 | idnum) lf;
+
   checkpoint ::= 'checkpoint' lf
     lf?;
 
@@ -361,6 +363,9 @@ static uintmax_t next_mark;
 static struct strbuf new_data = STRBUF_INIT;
 static int seen_data_command;
 
+/* Where to write output of cat-blob commands */
+static int cat_blob_fd = STDOUT_FILENO;
+
 static void parse_argv(void);
 
 static void write_branch_report(FILE *rpt, struct branch *b)
@@ -2689,6 +2694,81 @@ static void parse_reset_branch(void)
 		unread_command_buf = 1;
 }
 
+static void cat_blob_write(const char *buf, unsigned long size)
+{
+	if (write_in_full(cat_blob_fd, buf, size) != size)
+		die_errno("Write to frontend failed");
+}
+
+static void cat_blob(struct object_entry *oe, unsigned char sha1[20])
+{
+	struct strbuf line = STRBUF_INIT;
+	unsigned long size;
+	enum object_type type = 0;
+	char *buf;
+
+	if (!oe || oe->pack_id == MAX_PACK_ID) {
+		buf = read_sha1_file(sha1, &type, &size);
+	} else {
+		type = oe->type;
+		buf = gfi_unpack_entry(oe, &size);
+	}
+
+	/*
+	 * Output based on batch_one_object() from cat-file.c.
+	 */
+	if (type <= 0) {
+		strbuf_reset(&line);
+		strbuf_addf(&line, "%s missing\n", sha1_to_hex(sha1));
+		cat_blob_write(line.buf, line.len);
+		strbuf_release(&line);
+		free(buf);
+		return;
+	}
+	if (!buf)
+		die("Can't read object %s", sha1_to_hex(sha1));
+	if (type != OBJ_BLOB)
+		die("Object %s is a %s but a blob was expected.",
+		    sha1_to_hex(sha1), typename(type));
+	strbuf_reset(&line);
+	strbuf_addf(&line, "%s %s %lu\n", sha1_to_hex(sha1),
+						typename(type), size);
+	cat_blob_write(line.buf, line.len);
+	strbuf_release(&line);
+	cat_blob_write(buf, size);
+	cat_blob_write("\n", 1);
+	free(buf);
+}
+
+static void parse_cat_blob(void)
+{
+	const char *p;
+	struct object_entry *oe = oe;
+	unsigned char sha1[20];
+
+	/* cat-blob SP <object> LF */
+	p = command_buf.buf + strlen("cat-blob ");
+	if (*p == ':') {
+		char *x;
+		oe = find_mark(strtoumax(p + 1, &x, 10));
+		if (x == p + 1)
+			die("Invalid mark: %s", command_buf.buf);
+		if (!oe)
+			die("Unknown mark: %s", command_buf.buf);
+		if (*x)
+			die("Garbage after mark: %s", command_buf.buf);
+		hashcpy(sha1, oe->idx.sha1);
+	} else {
+		if (get_sha1_hex(p, sha1))
+			die("Invalid SHA1: %s", command_buf.buf);
+		if (p[40])
+			die("Garbage after SHA1: %s", command_buf.buf);
+		oe = find_object(sha1);
+	}
+
+	cat_blob(oe, sha1);
+}
+
 static void parse_checkpoint(void)
 {
 	if (object_count) {
@@ -2773,6 +2853,14 @@ static void option_export_marks(const char *marks)
 	safe_create_leading_directories_const(export_marks_file);
 }
 
+static void option_cat_blob_fd(const char *fd)
+{
+	unsigned long n = ulong_arg("--cat-blob-fd", fd);
+	if (n > (unsigned long) INT_MAX)
+		die("--cat-blob-fd cannot exceed %d", INT_MAX);
+	cat_blob_fd = (int) n;
+}
+
 static void option_export_pack_edges(const char *edges)
 {
 	if (pack_edges)
@@ -2826,6 +2914,8 @@ static int parse_one_feature(const char *feature, int from_stream)
 		option_import_marks(feature + 13, from_stream);
 	} else if (!prefixcmp(feature, "export-marks=")) {
 		option_export_marks(feature + 13);
+	} else if (!strcmp(feature, "cat-blob")) {
+		; /* Don't die - this feature is supported */
 	} else if (!prefixcmp(feature, "relative-marks")) {
 		relative_marks_paths = 1;
 	} else if (!prefixcmp(feature, "no-relative-marks")) {
@@ -2920,6 +3010,11 @@ static void parse_argv(void)
 		if (parse_one_feature(a + 2, 0))
 			continue;
 
+		if (!prefixcmp(a + 2, "cat-blob-fd=")) {
+			option_cat_blob_fd(a + 2 + strlen("cat-blob-fd="));
+			continue;
+		}
+
 		die("unknown option %s", a);
 	}
 	if (i != global_argc)
@@ -2971,6 +3066,8 @@ int main(int argc, const char **argv)
 			parse_new_tag();
 		else if (!prefixcmp(command_buf.buf, "reset "))
 			parse_reset_branch();
+		else if (!prefixcmp(command_buf.buf, "cat-blob "))
+			parse_cat_blob();
 		else if (!strcmp("checkpoint", command_buf.buf))
 			parse_checkpoint();
 		else if (!prefixcmp(command_buf.buf, "progress "))