Matching requests and responses with Pact-JVM

To understand the basic rules of matching, see Matching Gotchas. For example test cases for matching, see the Pact Specification Project, version 2.

By default, Pact-JVM will use string equality matching following Postel's Law. This means that for an actual value to match an expected one, they both must consist of the same sequence of characters. For collections (basically Maps and Lists), they must have the same elements that match in the same sequence, with cases where the additional elements in an actual Map are ignored.

Matching rules can be defined for both request and response elements based on JSON-Path syntax.

Matching Bodies

For the most part, matching involves matching request and response bodies in JSON format. Other formats will either have their own matching rules, or will follow the JSON one.

JSON body matching rules

Bodies consist of Objects (Maps of Key-Value pairs), Arrays (Lists) and values (Strings, Numbers, true, false, null). Body matching rules are prefixed with $.body.

The following method is used to determine if two bodies match:

1. If both the actual body and expected body are empty, the bodies match.
2. If the actual body is non-empty, and the expected body empty, the bodies match.
3. If the actual body is empty, and the expected body non-empty, the bodies don't match.
4. Otherwise do a comparison on the contents of the bodies.

For the body contents comparison:

1. If the actual and expected values are both Objects, compare as Maps.
2. If the actual and expected values are both Arrays, compare as Lists.
3. If the expected value is an Object, and the actual is not, they don't match.
4. If the expected value is an Array, and the actual is not, they don't match.
5. Otherwise, compare the values

For comparing Maps

1. If the actual map is non-empty while the expected is empty, they don't match.
2. If we allow unexpected keys, and the number of expected keys is greater than the actual keys, they don't match.
3. If we don't allow unexpected keys, and the expected and actual maps don't have the same number of keys, they don't match.
4. Otherwise, for each expected key and value pair:
   1. if the actual map contains the key, compare the values
   2. otherwise they don't match

Postel's law govers if we allow unexpected keys or not.

For comparing lists

1. If there is a body matcher defined that matches the path to the list, default to that matcher and then compare the list contents.
2. If the expected list is empty and the actual one is not, the lists don't match.
3. Otherwise
   1. compare the list sizes
   2. compare the list contents

For comparing list contents

1. For each value in the expected list:
   1. If the index of the value is less than the actual list's size, compare the value with the actual value at the same index using the method for comparing values.
   2. Otherwise the value doesn't match

For comparing values

1. If there is a matcher defined that matches the path to the value, default to that matcher
2. Otherwise compare the values using equality.

Matching Paths

Paths are matched by the following:

1. If there is a matcher defined for $.path, default to that matcher.
2. Otherwise paths are compared as Strings

Matching Queries

1. If the actual and expected query strings are empty, they match.
2. If the actual is not empty while the expected is, they don't match.
3. If the actual is empty while the expected is not, they don't match.
4. Otherwise convert both into a Map of keys mapped to a list values, and compare those.

Matching Query Maps

Query strings are parsed into a Map of keys mapped to lists of values. Key value pairs can be in any order, but when the same key appears more than once the values are compared in the order they appear in the query string.

Matching Headers

1. Do a case-insensitive sort of the headers by keys
2. For each expected header in the sorted list:
   1. If the actual headers contain that key, compare the header values
   2. Otherwise the header does not match

For matching header values:

1. If there is a matcher defined for $.header.<HEADER_KEY>, default to that matcher
2. Otherwise strip all whitespace after commas and compare the resulting strings.

Matching Request Headers

Request headers are matched by excluding the cookie header.

Matching Request cookies

If the list of expected cookies contains all the actual cookies, the cookies match.

Matching Status Codes

Status codes are compared as integer values.

Matching HTTP Methods

The actual and expected methods are compared as case-insensitive strings.

Matching Rules

Pact supports extending the matching rules on each type of object (Request, Response or Message) with a matchingRules element in the pact file. This is a map of JSON path strings to a matcher. When an item is being compared, if there is an entry in the matching rules that corresponds to the path to the item, the comparison will be delegated to the defined matcher.

Matcher Path expressions

Pact JVM does not support the full JSON path expressions, only ones that match the following rules:

1. All paths start with a dollar ($), representing the root.
2. All path elements are separated by periods (.), except array indices which use square brackets ([]).
3. The second element of the path is the http type that the matcher is applied to (e.g., $.body or $.header).
4. Path elements represent keys.
5. A star (*) can be used to match all keys of a map or all items of an array (one level only).

So the expression $.body.item1.level[2].id will match the highlighted item in the following body:

{
  "item1": {
    "level": [
      {
        "id": 100
      },
      {
        "id": 101
      },
      {
        "id": 102 // <---- $.body.item1.level[2].id
      },
      {
        "id": 103
      }
    ]
  }
}

while $.body.*.level[*].id will match all the ids of all the levels for all items.

Matcher selection algorithm

Due to the star notation, there can be multiple matcher paths defined that correspond to an item. The first, most specific expression is selected by assigning weightings to each path element and taking the product of the weightings. The matcher with the path with the largest weighting is used.

• The root node ($) is assigned the value 2.
• Any path element that does not match is assigned the value 0.
• Any property name that matches a path element is assigned the value 2.
• Any array index that matches a path element is assigned the value 2.
• Any star (*) that matches a property or array index is assigned the value 1.
• Everything else is assigned the value 0.

So for the body with highlighted item:

{
  "item1": {
    "level": [
      {
        "id": 100
      },
      {
        "id": 101
      },
      {
        "id": 102 // <--- Item under consideration
      },
      {
        "id": 103
      }
    ]
  }
}

The expressions will have the following weightings:

expression	weighting calculation	weighting
$	$(2)	2
$.body	$(2).body(2)	4
$.body.item1	$(2).body(2).item1(2)	8
$.body.item2	$(2).body(2).item2(0)	0
$.header.item1	$(2).header(0).item1(2)	0
$.body.item1.level	$(2).body(2).item1(2).level(2)	16
$.body.item1.level[1]	$(2).body(2).item1(2).level(2)[1(2)]	32
$.body.item1.level[1].id	$(2).body(2).item1(2).level(2)[1(2)].id(2)	64
$.body.item1.level[1].name	$(2).body(2).item1(2).level(2)[1(2)].name(0)	0
$.body.item1.level[2]	$(2).body(2).item1(2).level(2)[2(0)]	0
$.body.item1.level[2].id	$(2).body(2).item1(2).level(2)[2(0)].id(2)	0
$.body.item1.level[*].id	$(2).body(2).item1(2).level(2)[*(1)].id(2)	32
$.body.*.level[*].id	$(2).body(2).(1).level(2)[(1)].id(2)	8

So for the item with id 102, the matcher with path $.body.item1.level[1].id and weighting 64 will be selected.