Skip to content

Latest commit

 

History

History
137 lines (77 loc) · 8.73 KB

CHANGELOG.md

File metadata and controls

137 lines (77 loc) · 8.73 KB
Raw

JSON P3 Change Log

Version 1.3.4

Fixes

  • Fixed decoding of JSONPath escape sequences (those found in name selectors and string literals). Previously we were relying on JSON.parse() to unescape strings, now we have our own unescapeString() function that rejects invalid codepoints and surrogate pairs. See jsonpath-compliance-test-suite #87.
  • Fixed default minimum integer boundary for JSONPath indexes and slice steps. We were off by one.
  • Fixed parsing of JSONPath integer literals with an exponent and an upper case 'e'. We now allow 'e' to be upper case.
  • Fixed handling of trailing commas in JSONPath bracketed segments. We now raise a syntax error.
  • Fixed handling of invalid JSONPath integer and float literals with extra minus signs, leading zeros or too many zeros. We now raise a syntax error in such cases.

Version 1.3.3

Fixes

  • Fixed handling of JSONPath filter expression literals. We now throw a JSONPathSyntaxError if a filter expression contains a literal (string, int, float, boolean, null) that is not part of a comparison or function expression. See jsonpath-compliance-test-suite #81.

Version 1.3.2

Fixes

  • Fixed more I-Regexp to RegExp pattern mapping. See jsonpath-compliance-test-suite#77.
  • We now check that regular expression patterns passed to match and search are valid according to RFC 9485. The standard behavior is to silently return false from these filter function if the pattern is invalid. The throwErrors option can be passed to Match and/or Search to throw an error instead, and the iRegexpCheck option can be set to false to disable I-Regexp checks.

Version 1.3.1

Fixes

  • Fixed I-Regexp to RegExp pattern mapping with the match and search filter functions. We now correctly match the special . character to everything other than \r and \n.

Version 1.3.0

Fixes

  • Fixed the normalized path representation of the non-standard keys selector. Previously the normalized path of a node produced from a keys selector would always result in an empty node list. Note that this fix only applies when the default key token (~) is used. Custom key tokens are problematic when it comes to other, co-operating key selectors. If you're only interested in values from a node list, this wont affect you.

Features

  • Added the non-standard key selector (docs).
  • Added the non-standard keys filter selector (docs).

Version 1.2.1

Fixes

  • Fixed JSONPath name selectors that operator on array values. Previously, if a name selector were a quoted digit, we would treat that digit as an array index. The name selector now selects nothing if the target value is an array.
  • Fixed the lazy JSONPath slice selector. Previously, the lazy slice selector was not slicing values lazily, but creating an intermediate array.
  • Fixed JSONPath syntax error messages resulting from unbalanced parentheses or empty paren expressions.

Version 1.2.0

Fixes

  • Fixed the error and error message arising from JSONPath queries with filter expressions and a missing closing bracket for the segment. Previously we would get a JSONPathLexerError, stating we "can't backup beyond start", which is meant to be an internal error. We now get a JSONPathSyntaxError with the message "unclosed bracketed selection".

Features

  • Added a non-standard keys selector (~), selecting property names from objects. The keys selector is only enabled when setting JSONPathEnvironment's strict option to false (docs, source).
  • Added a non-standard current key identifier (#). # will be the key or index corresponding to @ in a filter expression. The current key identifier is only enabled when setting JSONPathEnvironment's strict option to false (docs, source).

Version 1.1.1

Fixes

  • Fixed evaluation of JSONPath singular queries when they appear in a logical expression and without a comparison operator. Previously we were evaluating logical expressions with the value held by the single element node list, now we treat such filter queries as existence tests.

Version 1.1.0

Fixes

  • Fixed logical operator precedence in JSONPath filter expressions. Previously, logical or (||) had a higher precedence than logical and (&&). Now && binds more tightly than ||.

Features

  • Added nondeterministic to JSONPathEnvironmentOptions and environment variables to control nondeterminism and the location of cts.json when testing for compliance. See the README for a description of these environment variables.

Version 1.0.0

RFC 9535 has been published, replacing the draft IETF JSONPath base.

Version 0.3.1

Breaking Changes

  • Changed the error thrown due to non-singular queries in comparison expressions from JSONPathSyntaxError to JSONPathTypeError.

Fixes

  • Fixed handling of relative and root queries when used as arguments to filter functions. Previously, when those queries resulted in an empty node list, we were converting it to an empty array before passing it to functions that accept ValueType arguments. Now, in such cases, we convert empty node lists to the special result Nothing, which is required by the spec.
  • Fixed well-typedness checks on JSONPath logical expressions (those that involve && or ||) and non-singular filter queries. Previously we were erroneously applying the checks for comparison expressions to logical expressions too. Now non-singular queries in logical expressions act as an existence test.

Version 0.3.0

Fixes

  • Fixed call stack size issues when querying large datasets with the recursive descent selector. This was mostly due to extending arrays using the spread operator. We now iterate and use Array.push().

Features

  • Added jsonpath.lazyQuery(), a lazy alternative to jsonpath.query(). lazyQuery() can be faster and more memory efficient if querying large datasets, especially when using recursive descent selectors. Conversely, query() is usually the better choice when working with small datasets.
  • jsonpath.match() now uses lazyQuery() internally, potentially avoiding a lot of unnecessary work.

Version 0.2.1

Fixes

  • Fixed JSONPath filter operator precedence. Previously logical and (&&) and or (||) were binding more tightly than logical negation (!).

Version 0.2.0

Breaking Changes

  • Rename JSONPathEnvironment.filterRegister to JSONPathEnvironment.functionRegister.
  • Removed JSONPathEnvironment.options in favour of equivalent environment properties. For example, JSONPathEnvironment.options.maxIntIndex is now JSONPathEnvironment.maxIntIndex.

Fixes

  • Fixed well-typedness checks on function calls. Previously we were throwing a JSONPathTypError for some valid queries.
  • Fixed parsing of function calls with comparison expressions as arguments.
  • Fixed a bug with unbalanced parentheses detection in JSONPath expressions when parsing nested function calls. In some cases we were not throwing a JSONPathSyntaxError when a function was not closed, but an inner function was.

Features

  • Implemented Relative JSON Pointers. Use the to(rel) method of JSONPointer, where rel is a relative JSON pointer string and a new JSONPointer is returned.
  • Guard against recursive data structures by implementing the JSONPathEnvironment.maxRecursionDepth option. When using the recursive descent selector (..), if the maximum recursion depth is reached, a JSONPathRecursionLimitError is thrown.
  • Added JSONPathEnvironment.match() and JSONPath.match(), which returns a JSONPathNode for the first value matching a query, or undefined if there are no matches.

Version 0.1.1

Fixes

  • Fix number literals with an implicit exponent sign. Previously we would see a JSONPathSyntaxError for 1e2, but not for 1e+2 or 1e-2. Now 1e2 is equivalent to 1e+2.

Version 0.1.0

Initial release.