Block scalar texts are parsed incorrectly: The final line break should be part of the scalar.

[mmueller2012]

The YAML spec states that the final line break is part of a block scalar unless strip - is used as a chomping mode. The parser however treats the final line break as whitespace and assigns it to a different node.

Here is an example to reproduce the issue:

a: |
  hello
  world
b: |
  hello
  world

The parser generates the following output:

ROOT@0..42
  DOCUMENT@0..41
    BLOCK@0..41
      BLOCK_MAP@0..41
        BLOCK_MAP_ENTRY@0..20
          BLOCK_MAP_KEY@0..1
            FLOW@0..1
              PLAIN_SCALAR@0..1 "a"
          COLON@1..2 ":"
          WHITESPACE@2..3 " "
          BLOCK_MAP_VALUE@3..20
            BLOCK@3..20
              BLOCK_SCALAR@3..20
                BAR@3..4 "|"
                BLOCK_SCALAR_TEXT@4..20 "\n  hello\n  world"
        WHITESPACE@20..21 "\n" <-----
        BLOCK_MAP_ENTRY@21..41
          BLOCK_MAP_KEY@21..22
            FLOW@21..22
              PLAIN_SCALAR@21..22 "b"
          COLON@22..23 ":"
          WHITESPACE@23..24 " "
          BLOCK_MAP_VALUE@24..41
            BLOCK@24..41
              BLOCK_SCALAR@24..41
                BAR@24..25 "|"
                BLOCK_SCALAR_TEXT@25..41 "\n  hello\n  world"
  WHITESPACE@41..42 "\n" <-----

The incorrectly parsed line breaks are marked with <-----. They should be part of the BLOCK_SCALAR_TEXT instead.

I tried to fix it myself, but it turned out to be quite challenging. Parsing the line break as part of BLOCK_SCALAR_TEXT breaks the parent rules which expect the line break as a separator, e.g. the block_map parsing function.