bug: grammar consumes the rest of the source string

[char0n]

Did you check existing issues?

• I have read all the tree-sitter docs if it relates to using the parser
• I have searched the existing issues

Tree-Sitter CLI Version, if relevant (output of tree-sitter --version)

tree-sitter 0.24.3 (bdfe32402e85673bbc693216f0a6ef72c98bb665)

Describe the bug

Hi,

I've been playing with how Errors and tree-sitter recovery mechanism work in this grammar and found some interesting cases that I'm not sure how to handle.

---
openapi: 3.1.0
x-top-level: value
info:
   title: Sample API
  unknownFixedField: value
  description: Optional multiline or single-line description in [CommonMark](http://commonmark.org/help/)
    or HTML.
  summary: example summary
  termsOfService: Terms of service
  version::: 0.1.9
  x-version: 0.1.9-beta
  license:
    name: Apache License 2.0
    x-fullName: Apache License 2.0
    identifier::: Apache License 2.0
    url: https://www.apache.org/licenses/LICENSE-2.0 

produces

[ERROR] [0, 0] - [4, 20]
  [block_mapping_pair] [1, 0] - [1, 14]
    key: [flow_node] [1, 0] - [1, 7]
      [plain_scalar] [1, 0] - [1, 7]
        [string_scalar] [1, 0] - [1, 7]
    value: [flow_node] [1, 9] - [1, 14]
      [plain_scalar] [1, 9] - [1, 14]
        [string_scalar] [1, 9] - [1, 14]
  [block_mapping_pair] [2, 0] - [2, 18]
    key: [flow_node] [2, 0] - [2, 11]
      [plain_scalar] [2, 0] - [2, 11]
        [string_scalar] [2, 0] - [2, 11]
    value: [flow_node] [2, 13] - [2, 18]
      [plain_scalar] [2, 13] - [2, 18]
        [string_scalar] [2, 13] - [2, 18]
  [flow_node] [3, 0] - [3, 4]
    [plain_scalar] [3, 0] - [3, 4]
      [string_scalar] [3, 0] - [3, 4]
  [block_mapping_pair] [4, 3] - [4, 20]
    key: [flow_node] [4, 3] - [4, 8]
      [plain_scalar] [4, 3] - [4, 8]
        [string_scalar] [4, 3] - [4, 8]
    value: [flow_node] [4, 10] - [4, 20]
      [plain_scalar] [4, 10] - [4, 20]
        [string_scalar] [4, 10] - [4, 20]

What is happening here is that parser doesn't parse after title: Sample API. This makes it pretty unusable in for example Editor cases. Only part of the source string is parsed and the rest is either ignored or consumed by parser. I would expect CST to contain at least another Error object with additional children nodes.

Did you ever bumped into this issue?

Thanks for any answer

Steps To Reproduce/Bad Parse Tree

https://tree-sitter.github.io/tree-sitter/playground

Expected Behavior/Parse Tree

Parse the entire source string and don't consume it

Repro

No response