Display differences clearly so people can review changes confidently.
The `diff` named export serializes JavaScript **values**, compares them line-by-line, and returns a string which includes comparison lines.
Two named exports compare **strings** character-by-character:
-`diffStringsUnified` returns a string.
-`diffStringsRaw` returns an array of `Diff` objects.
Three named exports compare **arrays of strings** line-by-line:
-`diffLinesUnified` and `diffLinesUnified2` return a string.
-`diffLinesRaw` returns an array of `Diff` objects.
## Installation
To add this package as a dependency of a project, run either of the following commands:
-`npm install jest-diff`
-`yarn add jest-diff`
## Usage of `diff()`
Given JavaScript **values**, `diff(a, b, options?)` does the following:
1.**serialize** the values as strings using the `pretty-format` package
2.**compare** the strings line-by-line using the `diff-sequences` package
3.**format** the changed or common lines using the `chalk` package
To use this function, write either of the following:
-`const {diff} = require('jest-diff');` in CommonJS modules
-`import {diff} from 'jest-diff';` in ECMAScript modules
### Example of `diff()`
```js
consta=['delete','common','changed from'];
constb=['common','changed to','insert'];
constdifference=diff(a,b);
```
The returned **string** consists of:
- annotation lines: describe the two change indicators with labels, and a blank line
- comparison lines: similar to “unified” view on GitHub, but `Expected` lines are green, `Received` lines are red, and common lines are dim (by default, see Options)
```diff
- Expected
+ Received
Array [
- "delete",
"common",
- "changed from",
+ "changed to",
+ "insert",
]
```
### Edge cases of `diff()`
Here are edge cases for the return value:
-`' Comparing two different types of values. …'` if the arguments have **different types** according to the `jest-get-type` package (instances of different classes have the same `'object'` type)
-`'Compared values have no visual difference.'` if the arguments have either **referential identity** according to `Object.is` method or **same serialization** according to the `pretty-format` package
-`null` if either argument is a so-called **asymmetric matcher** in Jasmine or Jest
## Usage of diffStringsUnified
Given **strings**, `diffStringsUnified(a, b, options?)` does the following:
1.**compare** the strings character-by-character using the `diff-sequences` package
2.**clean up** small (often coincidental) common substrings, also known as chaff
3.**format** the changed or common lines using the `chalk` package
Although the function is mainly for **multiline** strings, it compares any strings.
Write either of the following:
-`const {diffStringsUnified} = require('jest-diff');` in CommonJS modules
-`import {diffStringsUnified} from 'jest-diff';` in ECMAScript modules
### Example of diffStringsUnified
```js
consta='common\nchanged from';
constb='common\nchanged to';
constdifference=diffStringsUnified(a,b);
```
The returned **string** consists of:
- annotation lines: describe the two change indicators with labels, and a blank line
- comparison lines: similar to “unified” view on GitHub, and **changed substrings** have **inverse** foreground and background colors (that is, `from` has white-on-green and `to` has white-on-red, which the following example does not show)
```diff
- Expected
+ Received
common
- changed from
+ changed to
```
### Performance of diffStringsUnified
To get the benefit of **changed substrings** within the comparison lines, a character-by-character comparison has a higher computational cost (in time and space) than a line-by-line comparison.
If the input strings can have **arbitrary length**, we recommend that the calling code set a limit, beyond which splits the strings, and then calls `diffLinesUnified` instead. For example, Jest falls back to line-by-line comparison if either string has length greater than 20K characters.
## Usage of diffLinesUnified
Given **arrays of strings**, `diffLinesUnified(aLines, bLines, options?)` does the following:
1.**compare** the arrays line-by-line using the `diff-sequences` package
2.**format** the changed or common lines using the `chalk` package
You might call this function when strings have been split into lines and you do not need to see changed substrings within lines.
### Example of diffLinesUnified
```js
constaLines=['delete','common','changed from'];
constbLines=['common','changed to','insert'];
constdifference=diffLinesUnified(aLines,bLines);
```
```diff
- Expected
+ Received
- delete
common
- changed from
+ changed to
+ insert
```
### Edge cases of diffLinesUnified or diffStringsUnified
Here are edge cases for arguments and return values:
- both `a` and `b` are empty strings: no comparison lines
- only `a` is empty string: all comparison lines have `bColor` and `bIndicator` (see Options)
- only `b` is empty string: all comparison lines have `aColor` and `aIndicator` (see Options)
-`a` and `b` are equal non-empty strings: all comparison lines have `commonColor` and `commonIndicator` (see Options)
## Usage of diffLinesUnified2
Given two **pairs** of arrays of strings, `diffLinesUnified2(aLinesDisplay, bLinesDisplay, aLinesCompare, bLinesCompare, options?)` does the following:
1.**compare** the pair of `Compare` arrays line-by-line using the `diff-sequences` package
2.**format** the corresponding lines in the pair of `Display` arrays using the `chalk` package
Jest calls this function to consider lines as common instead of changed if the only difference is indentation.
You might call this function for case insensitive or Unicode equivalence comparison of lines.
### Example of diffLinesUnified2
```js
import{format}from'pretty-format';
consta={
text:'Ignore indentation in serialized object',
time:'2019-09-19T12:34:56.000Z',
type:'CREATE_ITEM',
};
constb={
payload:{
text:'Ignore indentation in serialized object',
time:'2019-09-19T12:34:56.000Z',
},
type:'CREATE_ITEM',
};
constdifference=diffLinesUnified2(
// serialize with indentation to display lines
format(a).split('\n'),
format(b).split('\n'),
// serialize without indentation to compare lines
format(a,{indent:0}).split('\n'),
format(b,{indent:0}).split('\n'),
);
```
The `text` and `time` properties are common, because their only difference is indentation:
```diff
- Expected
+ Received
Object {
+ payload: Object {
text: 'Ignore indentation in serialized object',
time: '2019-09-19T12:34:56.000Z',
+ },
type: 'CREATE_ITEM',
}
```
The preceding example illustrates why (at least for indentation) it seems more intuitive that the function returns the common line from the `bLinesDisplay` array instead of from the `aLinesDisplay` array.
## Usage of diffStringsRaw
Given **strings** and a boolean option, `diffStringsRaw(a, b, cleanup)` does the following:
1.**compare** the strings character-by-character using the `diff-sequences` package
2. optionally **clean up** small (often coincidental) common substrings, also known as chaff
Because `diffStringsRaw` returns the difference as **data** instead of a string, you can format it as your application requires (for example, enclosed in HTML markup for browser instead of escape sequences for console).
The returned **array** describes substrings as instances of the `Diff` class, which calling code can access like array tuples:
The value at index `0` is one of the following:
| value | named export | description |
| ----: | :------------ | :-------------------- |
| `0` | `DIFF_EQUAL` | in `a` and in `b` |
| `-1` | `DIFF_DELETE` | in `a` but not in `b` |
| `1` | `DIFF_INSERT` | in `b` but not in `a` |
The value at index `1` is a substring of `a` or `b` or both.
For edge case behavior like the `diffLinesUnified` function, you might define a `splitLines0` function, which given an empty string, returns `[]` an empty array:
In contrast to the `diffLinesRaw` function, the `diffLinesUnified` and `diffLinesUnified2` functions **automatically** convert array arguments computed by string `split` method, so callers do **not** need a `splitLine0` function.
## Options
The default options are for the report when an assertion fails from the `expect` package used by Jest.
For other applications, you can provide an options object as a third argument:
### Example of option for color of changed substrings
Although the default inverse of foreground and background colors is hard to beat for changed substrings **within lines**, especially because it highlights spaces, if you want bold font weight on yellow background color:
Because `diff()` does not display substring differences within lines, formatting can help you see when lines differ by the presence or absence of trailing spaces found by `/\s+$/` regular expression.
- If change lines have a background color, then you can see trailing spaces.
- If common lines have default dim color, then you cannot see trailing spaces. You might want yellowish background color to see them.
To display the number of changed lines at the right of annotation lines:
```js
consta=['common','changed from'];
constb=['common','changed to','insert'];
constoptions={
includeChangeCounts:true,
};
constdifference=diff(a,b,options);
```
```diff
- Expected - 1
+ Received + 2
Array [
"common",
- "changed from",
+ "changed to",
+ "insert",
]
```
### Example of option to omit annotation lines
To display only the comparison lines:
```js
consta='common\nchanged from';
constb='common\nchanged to';
constoptions={
omitAnnotationLines:true,
};
constdifference=diffStringsUnified(a,b,options);
```
```diff
common
- changed from
+ changed to
```
### Example of option for empty first or last lines
If the **first** or **last** comparison line is **empty**, because the content is empty and the indicator is a space, you might not notice it.
The replacement option is a string whose default value is `''` empty string.
Because Jest trims the report when a matcher fails, it deletes an empty last line.
Therefore, Jest uses as placeholder the downwards arrow with corner leftwards:
```js
constoptions={
emptyFirstOrLastLinePlaceholder:'↵',// U+21B5
};
```
If a content line is empty, then the corresponding comparison line is automatically trimmed to remove the margin space (represented as a middle dot below) for the default indicators:
| Indicator | untrimmed | trimmed |
| ----------------: | :-------- | :------ |
| `aIndicator` | `'-·'` | `'-'` |
| `bIndicator` | `'+·'` | `'+'` |
| `commonIndicator` | `' ·'` | `''` |
### Example of option for sorting object keys
When two objects are compared their keys are printed in alphabetical order by default. If this was not the original order of the keys the diff becomes harder to read as the keys are not in their original position.
Use `compareKeys` to pass a function which will be used when sorting the object keys.
```js
consta={c:'c',b:'b1',a:'a'};
constb={c:'c',b:'b2',a:'a'};
constoptions={
// The keys will be in their original order
compareKeys:()=>0,
};
constdifference=diff(a,b,options);
```
```diff
- Expected
+ Received
Object {
"c": "c",
- "b": "b1",
+ "b": "b2",
"a": "a",
}
```
Depending on the implementation of `compareKeys` any sort order can be used.
