Merge branch 'en/fast-import-parsing-fix'

"git fast-import" update.

* en/fast-import-parsing-fix:
  fast-import: fix erroneous handling of get-mark with empty orphan commits
  fast-import: only allow cat-blob requests where it makes sense
  fast-import: check most prominent commands first
  git-fast-import.txt: fix wording about where ls command can appear
  t9300: demonstrate bug with get-mark and empty orphan commits

Author: gitster

Documentation/git-fast-import.txt

@@ -422,7 +422,12 @@ However it is recommended that a `filedeleteall` command precede
 all `filemodify`, `filecopy`, `filerename` and `notemodify` commands in
 the same commit, as `filedeleteall` wipes the branch clean (see below).
 
-The `LF` after the command is optional (it used to be required).
+The `LF` after the command is optional (it used to be required).  Note
+that for reasons of backward compatibility, if the commit ends with a
+`data` command (i.e. it has has no `from`, `merge`, `filemodify`,
+`filedelete`, `filecopy`, `filerename`, `filedeleteall` or
+`notemodify` commands) then two `LF` commands may appear at the end of
+the command instead of just one.
 
 `author`
 ^^^^^^^^
@@ -966,10 +971,6 @@ might want to refer to in their commit messages.
 	'get-mark' SP ':' <idnum> LF
 ....
 
-This command can be used anywhere in the stream that comments are
-accepted.  In particular, the `get-mark` command can be used in the
-middle of a commit but not in the middle of a `data` command.
-
 See ``Responses To Commands'' below for details about how to read
 this output safely.
 
@@ -996,9 +997,10 @@ Output uses the same format as `git cat-file --batch`:
 	<contents> LF
 ====
 
-This command can be used anywhere in the stream that comments are
-accepted.  In particular, the `cat-blob` command can be used in the
-middle of a commit but not in the middle of a `data` command.
+This command can be used where a `filemodify` directive can appear,
+allowing it to be used in the middle of a commit.  For a `filemodify`
+using an inline directive, it can also appear right before the `data`
+directive.
 
 See ``Responses To Commands'' below for details about how to read
 this output safely.
@@ -1011,8 +1013,8 @@ printing a blob from the active commit (with `cat-blob`) or copying a
 blob or tree from a previous commit for use in the current one (with
 `filemodify`).
 
-The `ls` command can be used anywhere in the stream that comments are
-accepted, including the middle of a commit.
+The `ls` command can also be used where a `filemodify` directive can
+appear, allowing it to be used in the middle of a commit.
 
 Reading from the active commit::
 	This form can only be used in the middle of a `commit`.

fast-import.c

@@ -1748,8 +1748,6 @@ static int read_next_command(void)
 	}
 
 	for (;;) {
-		const char *p;
-
 		if (unread_command_buf) {
 			unread_command_buf = 0;
 		} else {
@@ -1782,14 +1780,6 @@ static int read_next_command(void)
 			rc->prev->next = rc;
 			cmd_tail = rc;
 		}
-		if (skip_prefix(command_buf.buf, "get-mark ", &p)) {
-			parse_get_mark(p);
-			continue;
-		}
-		if (skip_prefix(command_buf.buf, "cat-blob ", &p)) {
-			parse_cat_blob(p);
-			continue;
-		}
 		if (command_buf.buf[0] == '#')
 			continue;
 		return 0;
@@ -2254,8 +2244,15 @@ static void file_change_m(const char *p, struct branch *b)
 			strbuf_addstr(&uq, p);
 			p = uq.buf;
 		}
-		read_next_command();
-		parse_and_store_blob(&last_blob, &oid, 0);
+		while (read_next_command() != EOF) {
+			const char *v;
+			if (skip_prefix(command_buf.buf, "cat-blob ", &v))
+				parse_cat_blob(v);
+			else {
+				parse_and_store_blob(&last_blob, &oid, 0);
+				break;
+			}
+		}
 	} else {
 		enum object_type expected = S_ISDIR(mode) ?
 						OBJ_TREE: OBJ_BLOB;
@@ -2627,6 +2624,8 @@ static void parse_new_commit(const char *arg)
 			file_change_deleteall(b);
 		else if (skip_prefix(command_buf.buf, "ls ", &v))
 			parse_ls(v, b);
+		else if (skip_prefix(command_buf.buf, "cat-blob ", &v))
+			parse_cat_blob(v);
 		else {
 			unread_command_buf = 1;
 			break;
@@ -3303,14 +3302,18 @@ int cmd_main(int argc, const char **argv)
 		const char *v;
 		if (!strcmp("blob", command_buf.buf))
 			parse_new_blob();
-		else if (skip_prefix(command_buf.buf, "ls ", &v))
-			parse_ls(v, NULL);
 		else if (skip_prefix(command_buf.buf, "commit ", &v))
 			parse_new_commit(v);
 		else if (skip_prefix(command_buf.buf, "tag ", &v))
 			parse_new_tag(v);
 		else if (skip_prefix(command_buf.buf, "reset ", &v))
 			parse_reset_branch(v);
+		else if (skip_prefix(command_buf.buf, "ls ", &v))
+			parse_ls(v, NULL);
+		else if (skip_prefix(command_buf.buf, "cat-blob ", &v))
+			parse_cat_blob(v);
+		else if (skip_prefix(command_buf.buf, "get-mark ", &v))
+			parse_get_mark(v);
 		else if (!strcmp("checkpoint", command_buf.buf))
 			parse_checkpoint();
 		else if (!strcmp("done", command_buf.buf))

t/t9300-fast-import.sh

@@ -3262,4 +3262,41 @@ test_expect_success PIPE 'V: checkpoint updates tags after tag' '
 	background_import_still_running
 '
 
+###
+### series W (get-mark and empty orphan commits)
+###
+
+cat >>W-input <<-W_INPUT_END
+	commit refs/heads/W-branch
+	mark :1
+	author Full Name <[email protected]> 1000000000 +0100
+	committer Full Name <[email protected]> 1000000000 +0100
+	data 27
+	Intentionally empty commit
+	LFsget-mark :1
+	W_INPUT_END
+
+test_expect_success !MINGW 'W: get-mark & empty orphan commit with no newlines' '
+	sed -e s/LFs// W-input | tr L "\n" | git fast-import
+'
+
+test_expect_success !MINGW 'W: get-mark & empty orphan commit with one newline' '
+	sed -e s/LFs/L/ W-input | tr L "\n" | git fast-import
+'
+
+test_expect_success !MINGW 'W: get-mark & empty orphan commit with ugly second newline' '
+	# Technically, this should fail as it has too many linefeeds
+	# according to the grammar in fast-import.txt.  But, for whatever
+	# reason, it works.  Since using the correct number of newlines
+	# does not work with older (pre-2.22) versions of git, allow apps
+	# that used this second-newline workaround to keep working by
+	# checking it with this test...
+	sed -e s/LFs/LL/ W-input | tr L "\n" | git fast-import
+'
+
+test_expect_success !MINGW 'W: get-mark & empty orphan commit with erroneous third newline' '
+	# ...but do NOT allow more empty lines than that (see previous test).
+	sed -e s/LFs/LLL/ W-input | tr L "\n" | test_must_fail git fast-import
+'
+
 test_done