Make trailing LF following fast-import `data` commands optional

A few fast-import frontend developers have found it odd that we
require the LF following a `data` command, especially in the exact
byte count format.  Technically we don't need this LF to parse
the stream properly, but having it here does make the stream more
readable to humans.  We can easily make the LF optional by peeking
at the next byte available from the stream and pushing it back into
the buffer if its not LF.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>

diff --git a/Documentation/git-fast-import.txt b/Documentation/git-fast-import.txt
index a92ae6d1a36f800256a477f23c8280e2717b13c9..eb0c8c48ee257b87e4ae28351e91645d9221079e 100644 (file)
--- a/Documentation/git-fast-import.txt
+++ b/Documentation/git-fast-import.txt
@@ -708,13 +708,18 @@ Exact byte count format::
 +
 ....
        'data' SP <count> LF
 +
 ....
        'data' SP <count> LF

-       <raw> LF
+       <raw> LF?

 ....
 +
 where `<count>` is the exact number of bytes appearing within
 `<raw>`.  The value of `<count>` is expressed as an ASCII decimal
 integer.  The `LF` on either side of `<raw>` is not
 included in `<count>` and will not be included in the imported data.
 ....
 +
 where `<count>` is the exact number of bytes appearing within
 `<raw>`.  The value of `<count>` is expressed as an ASCII decimal
 integer.  The `LF` on either side of `<raw>` is not
 included in `<count>` and will not be included in the imported data.

++
+The `LF` after `<raw>` is optional (it used to be required) but
+recommended.  Always including it makes debugging a fast-import
+stream easier as the next command always starts in column 0
+of the next line, even if `<raw>` did not end with an `LF`.

 
 Delimited format::
        A delimiter string is used to mark the end of the data.
 
 Delimited format::
        A delimiter string is used to mark the end of the data.


@@ -726,6 +731,7 @@ Delimited format::
        'data' SP '<<' <delim> LF
        <raw> LF
        <delim> LF
        'data' SP '<<' <delim> LF
        <raw> LF
        <delim> LF

+       LF?

 ....
 +
 where `<delim>` is the chosen delimiter string.  The string `<delim>`
 ....
 +
 where `<delim>` is the chosen delimiter string.  The string `<delim>`


@@ -734,6 +740,8 @@ fast-import will think the data ends earlier than it really does.  The `LF`
 immediately trailing `<raw>` is part of `<raw>`.  This is one of
 the limitations of the delimited format, it is impossible to supply
 a data chunk which does not have an LF as its last byte.
 immediately trailing `<raw>` is part of `<raw>`.  This is one of
 the limitations of the delimited format, it is impossible to supply
 a data chunk which does not have an LF as its last byte.

++
+The `LF` after `<delim> LF` is optional (it used to be required).

 
 `checkpoint`
 ~~~~~~~~~~~~
 
 `checkpoint`
 ~~~~~~~~~~~~

diff --git a/fast-import.c b/fast-import.c
index 98ebe4770d239eeee26aca339166bd271adc5d8d..f950cff5efb5d2f1a26eedf269626a52690e5178 100644 (file)
--- a/fast-import.c
+++ b/fast-import.c
@@ -61,7 +61,7 @@ Format of STDIN stream:
      #
   mark ::= 'mark' sp idnum lf;
   data ::= (delimited_data | exact_data)
      #
   mark ::= 'mark' sp idnum lf;
   data ::= (delimited_data | exact_data)

-    lf;
+    lf?;

 
     # note: delim may be any string but must not contain lf.
     # data_line may contain any data but must not be exactly
 
     # note: delim may be any string but must not contain lf.
     # data_line may contain any data but must not be exactly


@@ -1470,6 +1470,13 @@ static void read_next_command(void)
        } while (!command_buf.eof && command_buf.buf[0] == '#');
 }
 
        } while (!command_buf.eof && command_buf.buf[0] == '#');
 }
 

+static void skip_optional_lf()
+{
+       int term_char = fgetc(stdin);
+       if (term_char != '\n' && term_char != EOF)
+               ungetc(term_char, stdin);
+}
+

 static void cmd_mark(void)
 {
        if (!prefixcmp(command_buf.buf, "mark :")) {
 static void cmd_mark(void)
 {
        if (!prefixcmp(command_buf.buf, "mark :")) {


@@ -1522,9 +1529,7 @@ static void *cmd_data (size_t *size)
                }
        }
 
                }
        }
 

-       if (fgetc(stdin) != '\n')
-               die("An lf did not trail the binary data as expected.");
-
+       skip_optional_lf();

        *size = length;
        return buffer;
 }
        *size = length;
        return buffer;
 }

diff --git a/t/t9300-fast-import.sh b/t/t9300-fast-import.sh
index 1f6426a49e5c42a865ca7fa2c736223b04105f46..5be6f196bdf8778f8a278107f29df8426d7cb02c 100755 (executable)
--- a/t/t9300-fast-import.sh
+++ b/t/t9300-fast-import.sh
@@ -818,4 +818,25 @@ test_expect_success \
        'git-fast-import <input &&
         test `git-rev-parse N3` = `git-rev-parse O1`'
 
        'git-fast-import <input &&
         test `git-rev-parse N3` = `git-rev-parse O1`'
 

+cat >input <<INPUT_END
+commit refs/heads/O2
+committer $GIT_COMMITTER_NAME <$GIT_COMMITTER_EMAIL> $GIT_COMMITTER_DATE
+data <<COMMIT
+dirty directory copy
+COMMIT
+from refs/heads/branch^0
+M 644 inline file2/file5
+data <<EOF
+$file5_data
+EOF
+C file2 file3
+D file2/file5
+
+INPUT_END
+
+test_expect_success \
+       'O: blank lines not necessary after data commands' \
+       'git-fast-import <input &&
+        test `git-rev-parse N3` = `git-rev-parse O2`'
+

 test_done
 test_done