User manual and reference guide version 4.0.4

CodeMirror is a code-editor component that can be embedded in Web pages. The core library provides only the editor component, no accompanying buttons, auto-completion, or other IDE functionality. It does provide a rich API on top of which such functionality can be straightforwardly implemented. See the addons included in the distribution, and the list of externally hosted addons, for reusable implementations of extra features.

CodeMirror works with language-specific modes. Modes are JavaScript programs that help color (and optionally indent) text written in a given language. The distribution comes with a number of modes (see the mode/ directory), and it isn't hard to write new ones for other languages.

Basic Usage

The easiest way to use CodeMirror is to simply load the script and style sheet found under lib/ in the distribution, plus a mode script from one of the mode/ directories. (See the compression helper for an easy way to combine scripts.) For example:

<script src="lib/codemirror.js"></script>
<link rel="stylesheet" href="../lib/codemirror.css">
<script src="mode/javascript/javascript.js"></script>

(Alternatively, use a module loader. More about that later.)

Having done this, an editor instance can be created like this:

var myCodeMirror = CodeMirror(document.body);

The editor will be appended to the document body, will start empty, and will use the mode that we loaded. To have more control over the new editor, a configuration object can be passed to CodeMirror as a second argument:

var myCodeMirror = CodeMirror(document.body, {
  value: "function myScript(){return 100;}\n",
  mode:  "javascript"
});

This will initialize the editor with a piece of code already in it, and explicitly tell it to use the JavaScript mode (which is useful when multiple modes are loaded). See below for a full discussion of the configuration options that CodeMirror accepts.

In cases where you don't want to append the editor to an element, and need more control over the way it is inserted, the first argument to the CodeMirror function can also be a function that, when given a DOM element, inserts it into the document somewhere. This could be used to, for example, replace a textarea with a real editor:

var myCodeMirror = CodeMirror(function(elt) {
  myTextArea.parentNode.replaceChild(elt, myTextArea);
}, {value: myTextArea.value});

However, for this use case, which is a common way to use CodeMirror, the library provides a much more powerful shortcut:

var myCodeMirror = CodeMirror.fromTextArea(myTextArea);

This will, among other things, ensure that the textarea's value is updated with the editor's contents when the form (if it is part of a form) is submitted. See the API reference for a full description of this method.

Module loaders

The files in the CodeMirror distribution contain shims for loading them (and their dependencies) in AMD or CommonJS environments. If the variables exports and module exist and have type object, CommonJS-style require will be used. If not, but there is a function define with an amd property present, AMD-style (RequireJS) will be used.

It is possible to use Browserify or similar tools to statically build modules using CodeMirror. Alternatively, use RequireJS to dynamically load dependencies at runtime. Both of these approaches have the advantage that they don't use the global namespace and can, thus, do things like load multiple versions of CodeMirror alongside each other.

Here's a simple example of using RequireJS to load CodeMirror:

require([
  "cm/lib/codemirror", "cm/mode/htmlmixed/htmlmixed"
], function(CodeMirror) {
  CodeMirror.fromTextArea(document.getElementById("code"), {
    lineNumbers: true,
    mode: "htmlmixed"
  });
});

It will automatically load the modes that the mixed HTML mode depends on (XML, JavaScript, and CSS).

Configuration

Both the CodeMirror function and its fromTextArea method take as second (optional) argument an object containing configuration options. Any option not supplied like this will be taken from CodeMirror.defaults, an object containing the default options. You can update this object to change the defaults on your page.

Options are not checked in any way, so setting bogus option values is bound to lead to odd errors.

These are the supported options:

value: string|CodeMirror.Doc
The starting value of the editor. Can be a string, or a document object.
mode: string|object
The mode to use. When not given, this will default to the first mode that was loaded. It may be a string, which either simply names the mode or is a MIME type associated with the mode. Alternatively, it may be an object containing configuration options for the mode, with a name property that names the mode (for example {name: "javascript", json: true}). The demo pages for each mode contain information about what configuration parameters the mode supports. You can ask CodeMirror which modes and MIME types have been defined by inspecting the CodeMirror.modes and CodeMirror.mimeModes objects. The first maps mode names to their constructors, and the second maps MIME types to mode specs.
theme: string
The theme to style the editor with. You must make sure the CSS file defining the corresponding .cm-s-[name] styles is loaded (see the theme directory in the distribution). The default is "default", for which colors are included in codemirror.css. It is possible to use multiple theming classes at once—for example "foo bar" will assign both the cm-s-foo and the cm-s-bar classes to the editor.
indentUnit: integer
How many spaces a block (whatever that means in the edited language) should be indented. The default is 2.
smartIndent: boolean
Whether to use the context-sensitive indentation that the mode provides (or just indent the same as the line before). Defaults to true.
tabSize: integer
The width of a tab character. Defaults to 4.
indentWithTabs: boolean
Whether, when indenting, the first N*tabSize spaces should be replaced by N tabs. Default is false.
electricChars: boolean
Configures whether the editor should re-indent the current line when a character is typed that might change its proper indentation (only works if the mode supports indentation). Default is true.
specialChars: RegExp
A regular expression used to determine which characters should be replaced by a special placeholder. Mostly useful for non-printing special characters. The default is /[\u0000-\u0019\u00ad\u200b\u2028\u2029\ufeff]/.
specialCharPlaceholder: function(char) → Element
A function that, given a special character identified by the specialChars option, produces a DOM node that is used to represent the character. By default, a red dot () is shown, with a title tooltip to indicate the character code.
rtlMoveVisually: boolean
Determines whether horizontal cursor movement through right-to-left (Arabic, Hebrew) text is visual (pressing the left arrow moves the cursor left) or logical (pressing the left arrow moves to the next lower index in the string, which is visually right in right-to-left text). The default is false on Windows, and true on other platforms.
keyMap: string
Configures the key map to use. The default is "default", which is the only key map defined in codemirror.js itself. Extra key maps are found in the key map directory. See the section on key maps for more information.
extraKeys: object
Can be used to specify extra key bindings for the editor, alongside the ones defined by keyMap. Should be either null, or a valid key map value.
lineWrapping: boolean
Whether CodeMirror should scroll or wrap for long lines. Defaults to false (scroll).
lineNumbers: boolean
Whether to show line numbers to the left of the editor.
firstLineNumber: integer
At which number to start counting lines. Default is 1.
lineNumberFormatter: function(line: integer) → string
A function used to format line numbers. The function is passed the line number, and should return a string that will be shown in the gutter.
gutters: array<string>
Can be used to add extra gutters (beyond or instead of the line number gutter). Should be an array of CSS class names, each of which defines a width (and optionally a background), and which will be used to draw the background of the gutters. May include the CodeMirror-linenumbers class, in order to explicitly set the position of the line number gutter (it will default to be to the right of all other gutters). These class names are the keys passed to setGutterMarker.
fixedGutter: boolean
Determines whether the gutter scrolls along with the content horizontally (false) or whether it stays fixed during horizontal scrolling (true, the default).
coverGutterNextToScrollbar: boolean
When fixedGutter is on, and there is a horizontal scrollbar, by default the gutter will be visible to the left of this scrollbar. If this option is set to true, it will be covered by an element with class CodeMirror-gutter-filler.
readOnly: boolean|string
This disables editing of the editor content by the user. If the special value "nocursor" is given (instead of simply true), focusing of the editor is also disallowed.
showCursorWhenSelecting: boolean
Whether the cursor should be drawn when a selection is active. Defaults to false.
undoDepth: integer
The maximum number of undo levels that the editor stores. Note that this includes selection change events. Defaults to 200.
historyEventDelay: integer
The period of inactivity (in milliseconds) that will cause a new history event to be started when typing or deleting. Defaults to 1250.
tabindex: integer
The tab index to assign to the editor. If not given, no tab index will be assigned.
autofocus: boolean
Can be used to make CodeMirror focus itself on initialization. Defaults to off. When fromTextArea is used, and no explicit value is given for this option, it will be set to true when either the source textarea is focused, or it has an autofocus attribute and no other element is focused.

Below this a few more specialized, low-level options are listed. These are only useful in very specific situations, you might want to skip them the first time you read this manual.

dragDrop: boolean
Controls whether drag-and-drop is enabled. On by default.
cursorBlinkRate: number
Half-period in milliseconds used for cursor blinking. The default blink rate is 530ms. By setting this to zero, blinking can be disabled.
cursorScrollMargin: number
How much extra space to always keep above and below the cursor when approaching the top or bottom of the visible view in a scrollable document. Default is 0.
cursorHeight: number
Determines the height of the cursor. Default is 1, meaning it spans the whole height of the line. For some fonts (and by some tastes) a smaller height (for example 0.85), which causes the cursor to not reach all the way to the bottom of the line, looks better
resetSelectionOnContextMenu: boolean
Controls whether, when the context menu is opened with a click outside of the current selection, the cursor is moved to the point of the click. Defaults to true.
workTime, workDelay: number
Highlighting is done by a pseudo background-thread that will work for workTime milliseconds, and then use timeout to sleep for workDelay milliseconds. The defaults are 200 and 300, you can change these options to make the highlighting more or less aggressive.
pollInterval: number
Indicates how quickly CodeMirror should poll its input textarea for changes (when focused). Most input is captured by events, but some things, like IME input on some browsers, don't generate events that allow CodeMirror to properly detect it. Thus, it polls. Default is 100 milliseconds.
flattenSpans: boolean
By default, CodeMirror will combine adjacent tokens into a single span if they have the same class. This will result in a simpler DOM tree, and thus perform better. With some kinds of styling (such as rounded corners), this will change the way the document looks. You can set this option to false to disable this behavior.
addModeClass: boolean
When enabled (off by default), an extra CSS class will be added to each token, indicating the (inner) mode that produced it, prefixed with "cm-m-". For example, tokens from the XML mode will get the cm-m-xml class.
maxHighlightLength: number
When highlighting long lines, in order to stay responsive, the editor will give up and simply style the rest of the line as plain text when it reaches a certain position. The default is 10 000. You can set this to Infinity to turn off this behavior.
crudeMeasuringFrom: number
When measuring the character positions in long lines, any line longer than this number (default is 10 000), when line wrapping is off, will simply be assumed to consist of same-sized characters. This means that, on the one hand, measuring will be inaccurate when characters of varying size, right-to-left text, markers, or other irregular elements are present. On the other hand, it means that having such a line won't freeze the user interface because of the expensiveness of the measurements.
viewportMargin: integer
Specifies the amount of lines that are rendered above and below the part of the document that's currently scrolled into view. This affects the amount of updates needed when scrolling, and the amount of work that such an update does. You should usually leave it at its default, 10. Can be set to Infinity to make sure the whole document is always rendered, and thus the browser's text search works on it. This will have bad effects on performance of big documents.

Events

Various CodeMirror-related objects emit events, which allow client code to react to various situations. Handlers for such events can be registered with the on and off methods on the objects that the event fires on. To fire your own events, use CodeMirror.signal(target, name, args...), where target is a non-DOM-node object.

An editor instance fires the following events. The instance argument always refers to the editor itself.

"change" (instance: CodeMirror, changeObj: object)
Fires every time the content of the editor is changed. The changeObj is a {from, to, text, removed, origin} object containing information about the changes that occurred as second argument. from and to are the positions (in the pre-change coordinate system) where the change started and ended (for example, it might be {ch:0, line:18} if the position is at the beginning of line #19). text is an array of strings representing the text that replaced the changed range (split by line). removed is the text that used to be between from and to, which is overwritten by this change.
"changes" (instance: CodeMirror, changes: array<object<)
Like the "change" event, but batched per operation, passing an array containing all the changes that happened in the operation.
"beforeChange" (instance: CodeMirror, changeObj: object)
This event is fired before a change is applied, and its handler may choose to modify or cancel the change. The changeObj object has from, to, and text properties, as with the "change" event. It also has a cancel() method, which can be called to cancel the change, and, if the change isn't coming from an undo or redo event, an update(from, to, text) method, which may be used to modify the change. Undo or redo changes can't be modified, because they hold some metainformation for restoring old marked ranges that is only valid for that specific change. All three arguments to update are optional, and can be left off to leave the existing value for that field intact. Note: you may not do anything from a "beforeChange" handler that would cause changes to the document or its visualization. Doing so will, since this handler is called directly from the bowels of the CodeMirror implementation, probably cause the editor to become corrupted.
"cursorActivity" (instance: CodeMirror)
Will be fired when the cursor or selection moves, or any change is made to the editor content.
"keyHandled" (instance: CodeMirror, name: string, event: Event)
Fired after a key is handled through a key map. name is the name of the handled key (for example "Ctrl-X" or "'q'"), and event is the DOM keydown or keypress event.
"inputRead" (instance: CodeMirror, changeObj: object)
Fired whenever new input is read from the hidden textarea (typed or pasted by the user).
"beforeSelectionChange" (instance: CodeMirror, obj: {ranges, update})
This event is fired before the selection is moved. Its handler may inspect the set of selection ranges, present as an array of {anchor, head} objects in the ranges property of the obj argument, and optionally change them by calling the update method on this object, passing an array of ranges in the same format. Handlers for this event have the same restriction as "beforeChange" handlers — they should not do anything to directly update the state of the editor.
"viewportChange" (instance: CodeMirror, from: number, to: number)
Fires whenever the view port of the editor changes (due to scrolling, editing, or any other factor). The from and to arguments give the new start and end of the viewport.
"swapDoc" (instance: CodeMirror, oldDoc: Doc)
This is signalled when the editor's document is replaced using the swapDoc method.
"gutterClick" (instance: CodeMirror, line: integer, gutter: string, clickEvent: Event)
Fires when the editor gutter (the line-number area) is clicked. Will pass the editor instance as first argument, the (zero-based) number of the line that was clicked as second argument, the CSS class of the gutter that was clicked as third argument, and the raw mousedown event object as fourth argument.
"gutterContextMenu" (instance: CodeMirror, line: integer, gutter: string, contextMenu: Event: Event)
Fires when the editor gutter (the line-number area) receives a contextmenu event. Will pass the editor instance as first argument, the (zero-based) number of the line that was clicked as second argument, the CSS class of the gutter that was clicked as third argument, and the raw contextmenu mouse event object as fourth argument. You can preventDefault the event, to signal that CodeMirror should do no further handling.
"focus" (instance: CodeMirror)
Fires whenever the editor is focused.
"blur" (instance: CodeMirror)
Fires whenever the editor is unfocused.
"scroll" (instance: CodeMirror)
Fires when the editor is scrolled.
"update" (instance: CodeMirror)
Will be fired whenever CodeMirror updates its DOM display.
"renderLine" (instance: CodeMirror, line: LineHandle, element: Element)
Fired whenever a line is (re-)rendered to the DOM. Fired right after the DOM element is built, before it is added to the document. The handler may mess with the style of the resulting element, or add event handlers, but should not try to change the state of the editor.
"mousedown", "dblclick", "contextmenu", "keydown", "keypress", "keyup", "dragstart", "dragenter", "dragover", "drop" (instance: CodeMirror, event: Event)
Fired when CodeMirror is handling a DOM event of this type. You can preventDefault the event, or give it a truthy codemirrorIgnore property, to signal that CodeMirror should do no further handling.

Document objects (instances of CodeMirror.Doc) emit the following events:

"change" (doc: CodeMirror.Doc, changeObj: object)
Fired whenever a change occurs to the document. changeObj has a similar type as the object passed to the editor's "change" event.
"beforeChange" (doc: CodeMirror.Doc, change: object)
See the description of the same event on editor instances.
"cursorActivity" (doc: CodeMirror.Doc)
Fired whenever the cursor or selection in this document changes.
"beforeSelectionChange" (doc: CodeMirror.Doc, selection: {head, anchor})
Equivalent to the event by the same name as fired on editor instances.

Line handles (as returned by, for example, getLineHandle) support these events:

"delete" ()
Will be fired when the line object is deleted. A line object is associated with the start of the line. Mostly useful when you need to find out when your gutter markers on a given line are removed.
"change" (line: LineHandle, changeObj: object)
Fires when the line's text content is changed in any way (but the line is not deleted outright). The change object is similar to the one passed to change event on the editor object.

Marked range handles (CodeMirror.TextMarker), as returned by markText and setBookmark, emit the following events:

"beforeCursorEnter" ()
Fired when the cursor enters the marked range. From this event handler, the editor state may be inspected but not modified, with the exception that the range on which the event fires may be cleared.
"clear" (from: {line, ch}, to: {line, ch})
Fired when the range is cleared, either through cursor movement in combination with clearOnEnter or through a call to its clear() method. Will only be fired once per handle. Note that deleting the range through text editing does not fire this event, because an undo action might bring the range back into existence. from and to give the part of the document that the range spanned when it was cleared.
"hide" ()
Fired when the last part of the marker is removed from the document by editing operations.
"unhide" ()
Fired when, after the marker was removed by editing, a undo operation brought the marker back.

Line widgets (CodeMirror.LineWidget), returned by addLineWidget, fire these events:

"redraw" ()
Fired whenever the editor re-adds the widget to the DOM. This will happen once right after the widget is added (if it is scrolled into view), and then again whenever it is scrolled out of view and back in again, or when changes to the editor options or the line the widget is on require the widget to be redrawn.

Key Maps

Key maps are ways to associate keys with functionality. A key map is an object mapping strings that identify the keys to functions that implement their functionality.

The CodeMirror distributions comes with Emacs, Vim, and Sublime Text-style keymaps.

Keys are identified either by name or by character. The CodeMirror.keyNames object defines names for common keys and associates them with their key codes. Examples of names defined here are Enter, F5, and Q. These can be prefixed with Shift-, Cmd-, Ctrl-, and Alt- (in that order!) to specify a modifier. So for example, Shift-Ctrl-Space would be a valid key identifier.

Common example: map the Tab key to insert spaces instead of a tab character.

{
  Tab: function(cm) {
    var spaces = Array(cm.getOption("indentUnit") + 1).join(" ");
    cm.replaceSelection(spaces);
  }
}

Alternatively, a character can be specified directly by surrounding it in single quotes, for example '$' or 'q'. Due to limitations in the way browsers fire key events, these may not be prefixed with modifiers.

The CodeMirror.keyMap object associates key maps with names. User code and key map definitions can assign extra properties to this object. Anywhere where a key map is expected, a string can be given, which will be looked up in this object. It also contains the "default" key map holding the default bindings.

The values of properties in key maps can be either functions of a single argument (the CodeMirror instance), strings, or false. Strings refer to commands, which are described below. If the property is set to false, CodeMirror leaves handling of the key up to the browser. A key handler function may return CodeMirror.Pass to indicate that it has decided not to handle the key, and other handlers (or the default behavior) should be given a turn.

Keys mapped to command names that start with the characters "go" (which should be used for cursor-movement actions) will be fired even when an extra Shift modifier is present (i.e. "Up": "goLineUp" matches both up and shift-up). This is used to easily implement shift-selection.

Key maps can defer to each other by defining a fallthrough property. This indicates that when a key is not found in the map itself, one or more other maps should be searched. It can hold either a single key map or an array of key maps.

When a key map contains a nofallthrough property set to true, keys matched against that map will be ignored if they don't match any of the bindings in the map (no further child maps will be tried). When the disableInput property is set to true, the default effect of inserting a character will be suppressed when the key map is active as the top-level map.

Commands

Commands are parameter-less actions that can be performed on an editor. Their main use is for key bindings. Commands are defined by adding properties to the CodeMirror.commands object. A number of common commands are defined by the library itself, most of them used by the default key bindings. The value of a command property must be a function of one argument (an editor instance).

Some of the commands below are referenced in the default key map, but not defined by the core library. These are intended to be defined by user code or addons.

Commands can also be run with the execCommand method.

selectAllCtrl-A (PC), Cmd-A (Mac)
Select the whole content of the editor.
singleSelectionEsc
When multiple selections are present, this deselects all but the primary selection.
killLineCtrl-K (Mac)
Emacs-style line killing. Deletes the part of the line after the cursor. If that consists only of whitespace, the newline at the end of the line is also deleted.
deleteLineCtrl-D (PC), Cmd-D (Mac)
Deletes the whole line under the cursor, including newline at the end.
delLineLeftCmd-Backspace (Mac)
Delete the part of the line before the cursor.
undoCtrl-Z (PC), Cmd-Z (Mac)
Undo the last change.
redoCtrl-Y (PC), Shift-Cmd-Z (Mac), Cmd-Y (Mac)
Redo the last undone change.
undoSelectionCtrl-U (PC), Cmd-U (Mac)
Undo the last change to the selection, or if there are no selection-only changes at the top of the history, undo the last change.
redoSelectionAlt-U (PC), Shift-Cmd-U (Mac)
Redo the last change to the selection, or the last text change if no selection changes remain.
goDocStartCtrl-Up (PC), Cmd-Up (Mac)
Move the cursor to the start of the document.
goDocEndCtrl-Down (PC), Cmd-End (Mac), Cmd-Down (Mac)
Move the cursor to the end of the document.
goLineStartAlt-Left (PC), Cmd-Left (Mac), Ctrl-A (Mac)
Move the cursor to the start of the line.
goLineStartSmartHome
Move to the start of the text on the line, or if we are already there, to the actual start of the line (including whitespace).
goLineEndAlt-Right (PC), Cmd-Right (Mac), Ctrl-E (Mac)
Move the cursor to the end of the line.
goLineLeft
Move the cursor to the left side of the visual line it is on. If this line is wrapped, that may not be the start of the line.
goLineRight
Move the cursor to the right side of the visual line it is on.
goLineUpUp, Ctrl-P (Mac)
Move the cursor up one line.
goLineDownDown, Ctrl-N (Mac)
Move down one line.
goPageUpPageUp, Shift-Ctrl-V (Mac)
Move the cursor up one screen, and scroll up by the same distance.
goPageDownPageDown, Ctrl-V (Mac)
Move the cursor down one screen, and scroll down by the same distance.
goCharLeftLeft, Ctrl-B (Mac)
Move the cursor one character left, going to the previous line when hitting the start of line.
goCharRightRight, Ctrl-F (Mac)
Move the cursor one character right, going to the next line when hitting the end of line.
goColumnLeft
Move the cursor one character left, but don't cross line boundaries.
goColumnRight
Move the cursor one character right, don't cross line boundaries.
goWordLeftAlt-B (Mac)
Move the cursor to the start of the previous word.
goWordRightAlt-F (Mac)
Move the cursor to the end of the next word.
goGroupLeftCtrl-Left (PC), Alt-Left (Mac)
Move to the left of the group before the cursor. A group is a stretch of word characters, a stretch of punctuation characters, a newline, or a stretch of more than one whitespace character.
goGroupRightCtrl-Right (PC), Alt-Right (Mac)
Move to the right of the group after the cursor (see above).
delCharBeforeShift-Backspace, Ctrl-H (Mac)
Delete the character before the cursor.
delCharAfterDelete, Ctrl-D (Mac)
Delete the character after the cursor.
delWordBeforeAlt-Backspace (Mac)
Delete up to the start of the word before the cursor.
delWordAfterAlt-D (Mac)
Delete up to the end of the word after the cursor.
delGroupBeforeCtrl-Backspace (PC), Alt-Backspace (Mac)
Delete to the left of the group before the cursor.
delGroupAfterCtrl-Delete (PC), Ctrl-Alt-Backspace (Mac), Alt-Delete (Mac)
Delete to the start of the group after the cursor.
indentAutoShift-Tab
Auto-indent the current line or selection.
indentMoreCtrl-] (PC), Cmd-] (Mac)
Indent the current line or selection by one indent unit.
indentLessCtrl-[ (PC), Cmd-[ (Mac)
Dedent the current line or selection by one indent unit.
insertTab
Insert a tab character at the cursor.
insertSoftTab
Insert the amount of spaces that match the width a tab at the cursor position would have.
defaultTabTab
If something is selected, indent it by one indent unit. If nothing is selected, insert a tab character.
transposeCharsCtrl-T (Mac)
Swap the characters before and after the cursor.
newlineAndIndentEnter
Insert a newline and auto-indent the new line.
toggleOverwriteInsert
Flip the overwrite flag.
saveCtrl-S (PC), Cmd-S (Mac)
Not defined by the core library, only referred to in key maps. Intended to provide an easy way for user code to define a save command.
findCtrl-F (PC), Cmd-F (Mac)
findNextCtrl-G (PC), Cmd-G (Mac)
findPrevShift-Ctrl-G (PC), Shift-Cmd-G (Mac)
replaceShift-Ctrl-F (PC), Cmd-Alt-F (Mac)
replaceAllShift-Ctrl-R (PC), Shift-Cmd-Alt-F (Mac)
Not defined by the core library, but defined in the search addon (or custom client addons).

Customized Styling

Up to a certain extent, CodeMirror's look can be changed by modifying style sheet files. The style sheets supplied by modes simply provide the colors for that mode, and can be adapted in a very straightforward way. To style the editor itself, it is possible to alter or override the styles defined in codemirror.css.

Some care must be taken there, since a lot of the rules in this file are necessary to have CodeMirror function properly. Adjusting colors should be safe, of course, and with some care a lot of other things can be changed as well. The CSS classes defined in this file serve the following roles:

CodeMirror
The outer element of the editor. This should be used for the editor width, height, borders and positioning. Can also be used to set styles that should hold for everything inside the editor (such as font and font size), or to set a background.
CodeMirror-scroll
Whether the editor scrolls (overflow: auto + fixed height). By default, it does. Setting the CodeMirror class to have height: auto and giving this class overflow-x: auto; overflow-y: hidden; will cause the editor to resize to fit its content.
CodeMirror-focused
Whenever the editor is focused, the top element gets this class. This is used to hide the cursor and give the selection a different color when the editor is not focused.
CodeMirror-gutters
This is the backdrop for all gutters. Use it to set the default gutter background color, and optionally add a border on the right of the gutters.
CodeMirror-linenumbers
Use this for giving a background or width to the line number gutter.
CodeMirror-linenumber
Used to style the actual individual line numbers. These won't be children of the CodeMirror-linenumbers (plural) element, but rather will be absolutely positioned to overlay it. Use this to set alignment and text properties for the line numbers.
CodeMirror-lines
The visible lines. This is where you specify vertical padding for the editor content.
CodeMirror-cursor
The cursor is a block element that is absolutely positioned. You can make it look whichever way you want.
CodeMirror-selected
The selection is represented by span elements with this class.
CodeMirror-matchingbracket, CodeMirror-nonmatchingbracket
These are used to style matched (or unmatched) brackets.

If your page's style sheets do funky things to all div or pre elements (you probably shouldn't do that), you'll have to define rules to cancel these effects out again for elements under the CodeMirror class.

Themes are also simply CSS files, which define colors for various syntactic elements. See the files in the theme directory.

Programming API

A lot of CodeMirror features are only available through its API. Thus, you need to write code (or use addons) if you want to expose them to your users.

Whenever points in the document are represented, the API uses objects with line and ch properties. Both are zero-based. CodeMirror makes sure to 'clip' any positions passed by client code so that they fit inside the document, so you shouldn't worry too much about sanitizing your coordinates. If you give ch a value of null, or don't specify it, it will be replaced with the length of the specified line.

Methods prefixed with doc. can, unless otherwise specified, be called both on CodeMirror (editor) instances and CodeMirror.Doc instances. Methods prefixed with cm. are only available on CodeMirror instances.

Constructor

Constructing an editor instance is done with the CodeMirror(place: Element|fn(Element), ?option: object) constructor. If the place argument is a DOM element, the editor will be appended to it. If it is a function, it will be called, and is expected to place the editor into the document. options may be an element mapping option names to values. The options that it doesn't explicitly specify (or all options, if it is not passed) will be taken from CodeMirror.defaults.

Note that the options object passed to the constructor will be mutated when the instance's options are changed, so you shouldn't share such objects between instances.

See CodeMirror.fromTextArea for another way to construct an editor instance.

Content manipulation methods

doc.getValue(?separator: string) → string
Get the current editor content. You can pass it an optional argument to specify the string to be used to separate lines (defaults to "\n").
doc.setValue(content: string)
Set the editor content.
doc.getRange(from: {line, ch}, to: {line, ch}, ?separator: string) → string
Get the text between the given points in the editor, which should be {line, ch} objects. An optional third argument can be given to indicate the line separator string to use (defaults to "\n").
doc.replaceRange(replacement: string, from: {line, ch}, to: {line, ch}, ?origin: string)
Replace the part of the document between from and to with the given string. from and to must be {line, ch} objects. to can be left off to simply insert the string at position from. When origin is given, it will be passed on to "change" events, and its first letter will be used to determine whether this change can be merged with previous history events, in the way described for selection origins.
doc.getLine(n: integer) → string
Get the content of line n.
doc.lineCount() → integer
Get the number of lines in the editor.
doc.firstLine() → integer
Get the first line of the editor. This will usually be zero but for linked sub-views, or documents instantiated with a non-zero first line, it might return other values.
doc.lastLine() → integer
Get the last line of the editor. This will usually be doc.lineCount() - 1, but for linked sub-views, it might return other values.
doc.getLineHandle(num: integer) → LineHandle
Fetches the line handle for the given line number.
doc.getLineNumber(handle: LineHandle) → integer
Given a line handle, returns the current position of that line (or null when it is no longer in the document).
doc.eachLine(f: (line: LineHandle))
doc.eachLine(start: integer, end: integer, f: (line: LineHandle))
Iterate over the whole document, or if start and end line numbers are given, the range from start up to (not including) end, and call f for each line, passing the line handle. This is a faster way to visit a range of line handlers than calling getLineHandle for each of them. Note that line handles have a text property containing the line's content (as a string).
doc.markClean()
Set the editor content as 'clean', a flag that it will retain until it is edited, and which will be set again when such an edit is undone again. Useful to track whether the content needs to be saved. This function is deprecated in favor of changeGeneration, which allows multiple subsystems to track different notions of cleanness without interfering.
doc.changeGeneration(?closeEvent: boolean) → integer
Returns a number that can later be passed to isClean to test whether any edits were made (and not undone) in the meantime. If closeEvent is true, the current history event will be ‘closed’, meaning it can't be combined with further changes (rapid typing or deleting events are typically combined).
doc.isClean(?generation: integer) → boolean
Returns whether the document is currently clean — not modified since initialization or the last call to markClean if no argument is passed, or since the matching call to changeGeneration if a generation value is given.

Cursor and selection methods

doc.getSelection(?lineSep: string) → string
Get the currently selected code. Optionally pass a line separator to put between the lines in the output. When multiple selections are present, they are concatenated with instances of lineSep in between.
doc.getSelections(?lineSep: string) → string
Returns an array containing a string for each selection, representing the content of the selections.
doc.replaceSelection(replacement: string, ?select: string)
Replace the selection(s) with the given string. By default, the new selection ends up after the inserted text. The optional select argument can be used to change this—passing "around" will cause the new text to be selected, passing "start" will collapse the selection to the start of the inserted text.
doc.replaceSelections(replacements: array<string>, ?select: string)
The length of the given array should be the same as the number of active selections. Replaces the content of the selections with the strings in the array. The select argument works the same as in replaceSelection.
doc.getCursor(?start: string) → {line, ch}
Retrieve one end of the primary selection. start is a an optional string indicating which end of the selection to return. It may be "from", "to", "head" (the side of the selection that moves when you press shift+arrow), or "anchor" (the fixed side of the selection). Omitting the argument is the same as passing "head". A {line, ch} object will be returned.
doc.listSelections() → array<{anchor, head}<
Retrieves a list of all current selections. These will always be sorted, and never overlap (overlapping selections are merged). Each object in the array contains anchor and head properties referring to {line, ch} objects.
doc.somethingSelected() → boolean
Return true if any text is selected.
doc.setCursor(pos: {line, ch}|number, ?ch: number, ?options: object)
Set the cursor position. You can either pass a single {line, ch} object, or the line and the character as two separate parameters. Will replace all selections with a single, empty selection at the given position. The supported options are the same as for setSelection.
doc.setSelection(anchor: {line, ch}, ?head: {line, ch}, ?options: object)
Set a single selection range. anchor and head should be {line, ch} objects. head defaults to anchor when not given. These options are supported:
scroll: boolean
Determines whether the selection head should be scrolled into view. Defaults to true.
origin: string
Detemines whether the selection history event may be merged with the previous one. When an origin starts with the character +, and the last recorded selection had the same origin and was similar (close in time, both collapsed or both non-collapsed), the new one will replace the old one. When it starts with *, it will always replace the previous event (if that had the same origin). Built-in motion uses the "+move" origin.
doc.setSelections(ranges: array<{anchor, head}>, ?primary: integer, ?options: object)
Sets a new set of selections. There must be at least one selection in the given array. When primary is a number, it determines which selection is the primary one. When it is not given, the primary index is taken from the previous selection, or set to the last range if the previous selection had less ranges than the new one. Supports the same options as setSelection.
doc.addSelection(anchor: {line, ch}, ?head: {line, ch})
Adds a new selection to the existing set of selections, and makes it the primary selection.
doc.extendSelection(from: {line, ch}, ?to: {line, ch}, ?options: object)
Similar to setSelection, but will, if shift is held or the extending flag is set, move the head of the selection while leaving the anchor at its current place. to is optional, and can be passed to ensure a region (for example a word or paragraph) will end up selected (in addition to whatever lies between that region and the current anchor). When multiple selections are present, all but the primary selection will be dropped by this method. Supports the same options as setSelection.
doc.extendSelections(heads: array<{line, ch}>, ?options: object)
An equivalent of extendSelection that acts on all selections at once.
doc.extendSelectionsBy(f: function(range: {anchor, head}) → {anchor, head}), ?options: object)
Applies the given function to all existing selections, and calls extendSelections on the result.
doc.setExtending(value: boolean)
Sets or clears the 'extending' flag, which acts similar to the shift key, in that it will cause cursor movement and calls to extendSelection to leave the selection anchor in place.
doc.getExtending() → boolean
Get the value of the 'extending' flag.
cm.hasFocus() → boolean
Tells you whether the editor currently has focus.
cm.findPosH(start: {line, ch}, amount: integer, unit: string, visually: boolean) → {line, ch, ?hitSide: boolean}
Used to find the target position for horizontal cursor motion. start is a {line, ch} object, amount an integer (may be negative), and unit one of the string "char", "column", or "word". Will return a position that is produced by moving amount times the distance specified by unit. When visually is true, motion in right-to-left text will be visual rather than logical. When the motion was clipped by hitting the end or start of the document, the returned value will have a hitSide property set to true.
cm.findPosV(start: {line, ch}, amount: integer, unit: string) → {line, ch, ?hitSide: boolean}
Similar to findPosH, but used for vertical motion. unit may be "line" or "page". The other arguments and the returned value have the same interpretation as they have in findPosH.

Configuration methods

cm.setOption(option: string, value: any)
Change the configuration of the editor. option should the name of an option, and value should be a valid value for that option.
cm.getOption(option: string) → any
Retrieves the current value of the given option for this editor instance.
cm.addKeyMap(map: object, bottom: boolean)
Attach an additional key map to the editor. This is mostly useful for addons that need to register some key handlers without trampling on the extraKeys option. Maps added in this way have a higher precedence than the extraKeys and keyMap options, and between them, the maps added earlier have a lower precedence than those added later, unless the bottom argument was passed, in which case they end up below other key maps added with this method.
cm.removeKeyMap(map: object)
Disable a keymap added with addKeyMap. Either pass in the key map object itself, or a string, which will be compared against the name property of the active key maps.
cm.addOverlay(mode: string|object, ?options: object)
Enable a highlighting overlay. This is a stateless mini-mode that can be used to add extra highlighting. For example, the search addon uses it to highlight the term that's currently being searched. mode can be a mode spec or a mode object (an object with a token method). The options parameter is optional. If given, it should be an object. Currently, only the opaque option is recognized. This defaults to off, but can be given to allow the overlay styling, when not null, to override the styling of the base mode entirely, instead of the two being applied together.
cm.removeOverlay(mode: string|object)
Pass this the exact value passed for the mode parameter to addOverlay, or a string that corresponds to the name propery of that value, to remove an overlay again.
cm.on(type: string, func: (...args))
Register an event handler for the given event type (a string) on the editor instance. There is also a CodeMirror.on(object, type, func) version that allows registering of events on any object.
cm.off(type: string, func: (...args))
Remove an event handler on the editor instance. An equivalent CodeMirror.off(object, type, func) also exists.

Document management methods

Each editor is associated with an instance of CodeMirror.Doc, its document. A document represents the editor content, plus a selection, an undo history, and a mode. A document can only be associated with a single editor at a time. You can create new documents by calling the CodeMirror.Doc(text, mode, firstLineNumber) constructor. The last two arguments are optional and can be used to set a mode for the document and make it start at a line number other than 0, respectively.

cm.getDoc() → Doc
Retrieve the currently active document from an editor.
doc.getEditor() → CodeMirror
Retrieve the editor associated with a document. May return null.
cm.swapDoc(doc: CodeMirror.Doc) → Doc
Attach a new document to the editor. Returns the old document, which is now no longer associated with an editor.
doc.copy(copyHistory: boolean) → Doc
Create an identical copy of the given doc. When copyHistory is true, the history will also be copied. Can not be called directly on an editor.
doc.linkedDoc(options: object) → Doc
Create a new document that's linked to the target document. Linked documents will stay in sync (changes to one are also applied to the other) until unlinked. These are the options that are supported:
sharedHist: boolean
When turned on, the linked copy will share an undo history with the original. Thus, something done in one of the two can be undone in the other, and vice versa.
from: integer
to: integer
Can be given to make the new document a subview of the original. Subviews only show a given range of lines. Note that line coordinates inside the subview will be consistent with those of the parent, so that for example a subview starting at line 10 will refer to its first line as line 10, not 0.
mode: string|object
By default, the new document inherits the mode of the parent. This option can be set to a mode spec to give it a different mode.
doc.unlinkDoc(doc: CodeMirror.Doc)
Break the link between two documents. After calling this, changes will no longer propagate between the documents, and, if they had a shared history, the history will become separate.
doc.iterLinkedDocs(function: (doc: CodeMirror.Doc, sharedHist: boolean))
Will call the given function for all documents linked to the target document. It will be passed two arguments, the linked document and a boolean indicating whether that document shares history with the target.

History-related methods

doc.undo()
Undo one edit (if any undo events are stored).
doc.redo()
Redo one undone edit.
doc.undoSelection()
Undo one edit or selection change.
doc.redoSelection()
Redo one undone edit or selection change.
doc.historySize() → {undo: integer, redo: integer}
Returns an object with {undo, redo} properties, both of which hold integers, indicating the amount of stored undo and redo operations.
doc.clearHistory()
Clears the editor's undo history.
doc.getHistory() → object
Get a (JSON-serializeable) representation of the undo history.
doc.setHistory(history: object)
Replace the editor's undo history with the one provided, which must be a value as returned by getHistory. Note that this will have entirely undefined results if the editor content isn't also the same as it was when getHistory was called.

Text-marking methods

doc.markText(from: {line, ch}, to: {line, ch}, ?options: object) → TextMarker
Can be used to mark a range of text with a specific CSS class name. from and to should be {line, ch} objects. The options parameter is optional. When given, it should be an object that may contain the following configuration options:
className: string
Assigns a CSS class to the marked stretch of text.
inclusiveLeft: boolean
Determines whether text inserted on the left of the marker will end up inside or outside of it.
inclusiveRight: boolean
Like inclusiveLeft, but for the right side.
atomic: boolean
Atomic ranges act as a single unit when cursor movement is concerned—i.e. it is impossible to place the cursor inside of them. In atomic ranges, inclusiveLeft and inclusiveRight have a different meaning—they will prevent the cursor from being placed respectively directly before and directly after the range.
collapsed: boolean
Collapsed ranges do not show up in the display. Setting a range to be collapsed will automatically make it atomic.
clearOnEnter: boolean
When enabled, will cause the mark to clear itself whenever the cursor enters its range. This is mostly useful for text-replacement widgets that need to 'snap open' when the user tries to edit them. The "clear" event fired on the range handle can be used to be notified when this happens.
clearWhenEmpty: boolean
Determines whether the mark is automatically cleared when it becomes empty. Default is true.
replacedWith: Element
Use a given node to display this range. Implies both collapsed and atomic. The given DOM node must be an inline element (as opposed to a block element).
handleMouseEvents: boolean
When replacedWith is given, this determines whether the editor will capture mouse and drag events occurring in this widget. Default is false—the events will be left alone for the default browser handler, or specific handlers on the widget, to capture.
readOnly: boolean
A read-only span can, as long as it is not cleared, not be modified except by calling setValue to reset the whole document. Note: adding a read-only span currently clears the undo history of the editor, because existing undo events being partially nullified by read-only spans would corrupt the history (in the current implementation).
addToHistory: boolean
When set to true (default is false), adding this marker will create an event in the undo history that can be individually undone (clearing the marker).
startStyle: string
Can be used to specify an extra CSS class to be applied to the leftmost span that is part of the marker.
endStyle: string
Equivalent to startStyle, but for the rightmost span.
title: string
When given, will give the nodes created for this span a HTML title attribute with the given value.
shared: boolean
When the target document is linked to other documents, you can set shared to true to make the marker appear in all documents. By default, a marker appears only in its target document.
The method will return an object that represents the marker (with constructor CodeMirror.TextMarker), which exposes three methods: clear(), to remove the mark, find(), which returns a {from, to} object (both holding document positions), indicating the current position of the marked range, or undefined if the marker is no longer in the document, and finally changed(), which you can call if you've done something that might change the size of the marker (for example changing the content of a replacedWith node), and want to cheaply update the display.
doc.setBookmark(pos: {line, ch}, ?options: object) → TextMarker
Inserts a bookmark, a handle that follows the text around it as it is being edited, at the given position. A bookmark has two methods find() and clear(). The first returns the current position of the bookmark, if it is still in the document, and the second explicitly removes the bookmark. The options argument is optional. If given, the following properties are recognized:
widget: Element
Can be used to display a DOM node at the current location of the bookmark (analogous to the replacedWith option to markText).
insertLeft: boolean
By default, text typed when the cursor is on top of the bookmark will end up to the right of the bookmark. Set this option to true to make it go to the left instead.
shared: boolean
See the corresponding option to markText.
doc.findMarks(from: {line, ch}, to: {line, ch}) → array<TextMarker>
Returns an array of all the bookmarks and marked ranges found between the given positions.
doc.findMarksAt(pos: {line, ch}) → array<TextMarker>
Returns an array of all the bookmarks and marked ranges present at the given position.
doc.getAllMarks() → array<TextMarker>
Returns an array containing all marked ranges in the document.

Widget, gutter, and decoration methods

cm.setGutterMarker(line: integer|LineHandle, gutterID: string, value: Element) → LineHandle
Sets the gutter marker for the given gutter (identified by its CSS class, see the gutters option) to the given value. Value can be either null, to clear the marker, or a DOM element, to set it. The DOM element will be shown in the specified gutter next to the specified line.
cm.clearGutter(gutterID: string)
Remove all gutter markers in the gutter with the given ID.
cm.addLineClass(line: integer|LineHandle, where: string, class: string) → LineHandle
Set a CSS class name for the given line. line can be a number or a line handle. where determines to which element this class should be applied, can can be one of "text" (the text element, which lies in front of the selection), "background" (a background element that will be behind the selection), or "wrap" (the wrapper node that wraps all of the line's elements, including gutter elements). class should be the name of the class to apply.
cm.removeLineClass(line: integer|LineHandle, where: string, class: string) → LineHandle
Remove a CSS class from a line. line can be a line handle or number. where should be one of "text", "background", or "wrap" (see addLineClass). class can be left off to remove all classes for the specified node, or be a string to remove only a specific class.
cm.lineInfo(line: integer|LineHandle) → object
Returns the line number, text content, and marker status of the given line, which can be either a number or a line handle. The returned object has the structure {line, handle, text, gutterMarkers, textClass, bgClass, wrapClass, widgets}, where gutterMarkers is an object mapping gutter IDs to marker elements, and widgets is an array of line widgets attached to this line, and the various class properties refer to classes added with addLineClass.
cm.addWidget(pos: {line, ch}, node: Element, scrollIntoView: boolean)
Puts node, which should be an absolutely positioned DOM node, into the editor, positioned right below the given {line, ch} position. When scrollIntoView is true, the editor will ensure that the entire node is visible (if possible). To remove the widget again, simply use DOM methods (move it somewhere else, or call removeChild on its parent).
cm.addLineWidget(line: integer|LineHandle, node: Element, ?options: object) → LineWidget
Adds a line widget, an element shown below a line, spanning the whole of the editor's width, and moving the lines below it downwards. line should be either an integer or a line handle, and node should be a DOM node, which will be displayed below the given line. options, when given, should be an object that configures the behavior of the widget. The following options are supported (all default to false):
coverGutter: boolean
Whether the widget should cover the gutter.
noHScroll: boolean
Whether the widget should stay fixed in the face of horizontal scrolling.
above: boolean
Causes the widget to be placed above instead of below the text of the line.
handleMouseEvents: boolean
Determines whether the editor will capture mouse and drag events occurring in this widget. Default is false—the events will be left alone for the default browser handler, or specific handlers on the widget, to capture.
insertAt: integer
By default, the widget is added below other widgets for the line. This option can be used to place it at a different position (zero for the top, N to put it after the Nth other widget). Note that this only has effect once, when the widget is created.
Note that the widget node will become a descendant of nodes with CodeMirror-specific CSS classes, and those classes might in some cases affect it. This method returns an object that represents the widget placement. It'll have a line property pointing at the line handle that it is associated with, and the following methods:
clear()
Removes the widget.
changed()
Call this if you made some change to the widget's DOM node that might affect its height. It'll force CodeMirror to update the height of the line that contains the widget.

Sizing, scrolling and positioning methods

cm.setSize(width: number|string, height: number|string)
Programatically set the size of the editor (overriding the applicable CSS rules). width and height can be either numbers (interpreted as pixels) or CSS units ("100%", for example). You can pass null for either of them to indicate that that dimension should not be changed.
cm.scrollTo(x: number, y: number)
Scroll the editor to a given (pixel) position. Both arguments may be left as null or undefined to have no effect.
cm.getScrollInfo() → {left, top, width, height, clientWidth, clientHeight}
Get an {left, top, width, height, clientWidth, clientHeight} object that represents the current scroll position, the size of the scrollable area, and the size of the visible area (minus scrollbars).
cm.scrollIntoView(what: {line, ch}|{left, top, right, bottom}|{from, to}|null, ?margin: number)
Scrolls the given position into view. what may be null to scroll the cursor into view, a {line, ch} position to scroll a character into view, a {left, top, right, bottom} pixel range (in editor-local coordinates), or a range {from, to} containing either two character positions or two pixel squares. The margin parameter is optional. When given, it indicates the amount of vertical pixels around the given area that should be made visible as well.
cm.cursorCoords(where: boolean|{line, ch}, mode: string) → {left, top, bottom}
Returns an {left, top, bottom} object containing the coordinates of the cursor position. If mode is "local", they will be relative to the top-left corner of the editable document. If it is "page" or not given, they are relative to the top-left corner of the page. where can be a boolean indicating whether you want the start (true) or the end (false) of the selection, or, if a {line, ch} object is given, it specifies the precise position at which you want to measure.
cm.charCoords(pos: {line, ch}, ?mode: string) → {left, right, top, bottom}
Returns the position and dimensions of an arbitrary character. pos should be a {line, ch} object. This differs from cursorCoords in that it'll give the size of the whole character, rather than just the position that the cursor would have when it would sit at that position.
cm.coordsChar(object: {left, top}, ?mode: string) → {line, ch}
Given an {left, top} object, returns the {line, ch} position that corresponds to it. The optional mode parameter determines relative to what the coordinates are interpreted. It may be "window", "page" (the default), or "local".
cm.lineAtHeight(height: number, ?mode: string) → number
Computes the line at the given pixel height. mode can be one of the same strings that coordsChar accepts.
cm.heightAtLine(line: number, ?mode: string) → number
Computes the height of the top of a line, in the coordinate system specified by mode (see coordsChar), which defaults to "page". When a line below the bottom of the document is specified, the returned value is the bottom of the last line in the document.
cm.defaultTextHeight() → number
Returns the line height of the default font for the editor.
cm.defaultCharWidth() → number
Returns the pixel width of an 'x' in the default font for the editor. (Note that for non-monospace fonts, this is mostly useless, and even for monospace fonts, non-ascii characters might have a different width).
cm.getViewport() → {from: number, to: number}
Returns a {from, to} object indicating the start (inclusive) and end (exclusive) of the currently rendered part of the document. In big documents, when most content is scrolled out of view, CodeMirror will only render the visible part, and a margin around it. See also the viewportChange event.
cm.refresh()
If your code does something to change the size of the editor element (window resizes are already listened for), or unhides it, you should probably follow up by calling this method to ensure CodeMirror is still looking as intended.

Mode, state, and token-related methods

When writing language-aware functionality, it can often be useful to hook into the knowledge that the CodeMirror language mode has. See the section on modes for a more detailed description of how these work.

doc.getMode() → object
Gets the (outer) mode object for the editor. Note that this is distinct from getOption("mode"), which gives you the mode specification, rather than the resolved, instantiated mode object.
doc.getModeAt(pos: {line, ch}) → object
Gets the inner mode at a given position. This will return the same as getMode for simple modes, but will return an inner mode for nesting modes (such as htmlmixed).
cm.getTokenAt(pos: {line, ch}, ?precise: boolean) → object
Retrieves information about the token the current mode found before the given position (a {line, ch} object). The returned object has the following properties:
start
The character (on the given line) at which the token starts.
end
The character at which the token ends.
string
The token's string.
type
The token type the mode assigned to the token, such as "keyword" or "comment" (may also be null).
state
The mode's state at the end of this token.
If precise is true, the token will be guaranteed to be accurate based on recent edits. If false or not specified, the token will use cached state information, which will be faster but might not be accurate if edits were recently made and highlighting has not yet completed.
cm.getTokenTypeAt(pos: {line, ch}) → string
This is a (much) cheaper version of getTokenAt useful for when you just need the type of the token at a given position, and no other information. Will return null for unstyled tokens, and a string, potentially containing multiple space-separated style names, otherwise.
cm.getHelpers(pos: {line, ch}, type: string) → array<helper>
Fetch the set of applicable helper values for the given position. Helpers provide a way to look up functionality appropriate for a mode. The type argument provides the helper namespace (see registerHelper), in which the values will be looked up. When the mode itself has a property that corresponds to the type, that directly determines the keys that are used to look up the helper values (it may be either a single string, or an array of strings). Failing that, the mode's helperType property and finally the mode's name are used.
For example, the JavaScript mode has a property fold containing "brace". When the brace-fold addon is loaded, that defines a helper named brace in the fold namespace. This is then used by the foldcode addon to figure out that it can use that folding function to fold JavaScript code.
When any 'global' helpers are defined for the given namespace, their predicates are called on the current mode and editor, and all those that declare they are applicable will also be added to the array that is returned.
cm.getHelper(pos: {line, ch}, type: string) → helper
Returns the first applicable helper value. See getHelpers.
cm.getStateAfter(?line: integer, ?precise: boolean) → object
Returns the mode's parser state, if any, at the end of the given line number. If no line number is given, the state at the end of the document is returned. This can be useful for storing parsing errors in the state, or getting other kinds of contextual information for a line. precise is defined as in getTokenAt().

Miscellaneous methods

cm.operation(func: () → any) → any
CodeMirror internally buffers changes and only updates its DOM structure after it has finished performing some operation. If you need to perform a lot of operations on a CodeMirror instance, you can call this method with a function argument. It will call the function, buffering up all changes, and only doing the expensive update after the function returns. This can be a lot faster. The return value from this method will be the return value of your function.
cm.indentLine(line: integer, ?dir: string|integer)
Adjust the indentation of the given line. The second argument (which defaults to "smart") may be one of:
"prev"
Base indentation on the indentation of the previous line.
"smart"
Use the mode's smart indentation if available, behave like "prev" otherwise.
"add"
Increase the indentation of the line by one indent unit.
"subtract"
Reduce the indentation of the line.
<integer>
Add (positive number) or reduce (negative number) the indentation by the given amount of spaces.
cm.toggleOverwrite(?value: bool)
Switches between overwrite and normal insert mode (when not given an argument), or sets the overwrite mode to a specific state (when given an argument).
cm.execCommand(name: string)
Runs the command with the given name on the editor.
doc.posFromIndex(index: integer) → {line, ch}
Calculates and returns a {line, ch} object for a zero-based index who's value is relative to the start of the editor's text. If the index is out of range of the text then the returned object is clipped to start or end of the text respectively.
doc.indexFromPos(object: {line, ch}) → integer
The reverse of posFromIndex.
cm.focus()
Give the editor focus.
cm.getInputField() → TextAreaElement
Returns the hidden textarea used to read input.
cm.getWrapperElement() → Element
Returns the DOM node that represents the editor, and controls its size. Remove this from your tree to delete an editor instance.
cm.getScrollerElement() → Element
Returns the DOM node that is responsible for the scrolling of the editor.
cm.getGutterElement() → Element
Fetches the DOM node that contains the editor gutters.

Static properties

The CodeMirror object itself provides several useful properties.

CodeMirror.version: string
It contains a string that indicates the version of the library. This is a triple of integers "major.minor.patch", where patch is zero for releases, and something else (usually one) for dev snapshots.
CodeMirror.fromTextArea(textArea: TextAreaElement, ?config: object)
The method provides another way to initialize an editor. It takes a textarea DOM node as first argument and an optional configuration object as second. It will replace the textarea with a CodeMirror instance, and wire up the form of that textarea (if any) to make sure the editor contents are put into the textarea when the form is submitted. The text in the textarea will provide the content for the editor. A CodeMirror instance created this way has three additional methods:
cm.save()
Copy the content of the editor into the textarea.
cm.toTextArea()
Remove the editor, and restore the original textarea (with the editor's current content).
cm.getTextArea() → TextAreaElement
Returns the textarea that the instance was based on.
CodeMirror.defaults: object
An object containing default values for all options. You can assign to its properties to modify defaults (though this won't affect editors that have already been created).
CodeMirror.defineExtension(name: string, value: any)
If you want to define extra methods in terms of the CodeMirror API, it is possible to use defineExtension. This will cause the given value (usually a method) to be added to all CodeMirror instances created from then on.
CodeMirror.defineDocExtension(name: string, value: any)
Like defineExtension, but the method will be added to the interface for Doc objects instead.
CodeMirror.defineOption(name: string, default: any, updateFunc: function)
Similarly, defineOption can be used to define new options for CodeMirror. The updateFunc will be called with the editor instance and the new value when an editor is initialized, and whenever the option is modified through setOption.
CodeMirror.defineInitHook(func: function)
If your extention just needs to run some code whenever a CodeMirror instance is initialized, use CodeMirror.defineInitHook. Give it a function as its only argument, and from then on, that function will be called (with the instance as argument) whenever a new CodeMirror instance is initialized.
CodeMirror.registerHelper(type: string, name: string, value: helper)
Registers a helper value with the given name in the given namespace (type). This is used to define functionality that may be looked up by mode. Will create (if it doesn't already exist) a property on the CodeMirror object for the given type, pointing to an object that maps names to values. I.e. after doing CodeMirror.registerHelper("hint", "foo", myFoo), the value CodeMirror.hint.foo will point to myFoo.
CodeMirror.registerGlobalHelper(type: string, name: string, predicate: fn(mode, CodeMirror), value: helper)
Acts like registerHelper, but also registers this helper as 'global', meaning that it will be included by getHelpers whenever the given predicate returns true when called with the local mode and editor.
CodeMirror.Pos(line: integer, ?ch: integer)
A constructor for the {line, ch} objects that are used to represent positions in editor documents.
CodeMirror.changeEnd(change: object) → {line, ch}
Utility function that computes an end position from a change (an object with from, to, and text properties, as passed to various event handlers). The returned position will be the end of the changed range, after the change is applied.

Addons

The addon directory in the distribution contains a number of reusable components that implement extra editor functionality (on top of extension functions like defineOption, defineExtension, and registerHelper). In brief, they are:

dialog/dialog.js
Provides a very simple way to query users for text input. Adds an openDialog method to CodeMirror instances, which can be called with an HTML fragment or a detached DOM node that provides the prompt (should include an input tag), and a callback function that is called when text has been entered. Also adds an openNotification function that simply shows an HTML fragment as a notification. Depends on addon/dialog/dialog.css.
search/searchcursor.js
Adds the getSearchCursor(query, start, caseFold) → cursor method to CodeMirror instances, which can be used to implement search/replace functionality. query can be a regular expression or a string (only strings will match across lines—if they contain newlines). start provides the starting position of the search. It can be a {line, ch} object, or can be left off to default to the start of the document. caseFold is only relevant when matching a string. It will cause the search to be case-insensitive. A search cursor has the following methods:
findNext() → boolean
findPrevious() → boolean
Search forward or backward from the current position. The return value indicates whether a match was found. If matching a regular expression, the return value will be the array returned by the match method, in case you want to extract matched groups.
from() → {line, ch}
to() → {line, ch}
These are only valid when the last call to findNext or findPrevious did not return false. They will return {line, ch} objects pointing at the start and end of the match.
replace(text: string)
Replaces the currently found match with the given text and adjusts the cursor position to reflect the replacement.
Implements the search commands. CodeMirror has keys bound to these by default, but will not do anything with them unless an implementation is provided. Depends on searchcursor.js, and will make use of openDialog when available to make prompting for search queries less ugly.
edit/matchbrackets.js
Defines an option matchBrackets which, when set to true, causes matching brackets to be highlighted whenever the cursor is next to them. It also adds a method matchBrackets that forces this to happen once, and a method findMatchingBracket that can be used to run the bracket-finding algorithm that this uses internally.
edit/closebrackets.js
Defines an option autoCloseBrackets that will auto-close brackets and quotes when typed. By default, it'll auto-close ()[]{}''"", but you can pass it a string similar to that (containing pairs of matching characters), or an object with pairs and optionally explode properties to customize it. explode should be a similar string that gives the pairs of characters that, when enter is pressed between them, should have the second character also moved to its own line. Demo here.
edit/matchtags.js
Defines an option matchTags that, when enabled, will cause the tags around the cursor to be highlighted (using the CodeMirror-matchingtag class). Also defines a command toMatchingTag, which you can bind a key to in order to jump to the tag mathing the one under the cursor. Depends on the addon/fold/xml-fold.js addon. Demo here.
edit/trailingspace.js
Adds an option showTrailingSpace which, when enabled, adds the CSS class cm-trailingspace to stretches of whitespace at the end of lines. The demo has a nice squiggly underline style for this class.
edit/closetag.js
Provides utility functions for adding automatic tag closing to XML modes. See the demo.
edit/continuelist.js
Markdown specific. Defines a "newlineAndIndentContinueMarkdownList" command command that can be bound to enter to automatically insert the leading characters for continuing a list. See the Markdown mode demo.
comment/comment.js
Addon for commenting and uncommenting code. Adds three methods to CodeMirror instances:
lineComment(from: {line, ch}, to: {line, ch}, ?options: object)
Set the lines in the given range to be line comments. Will fall back to blockComment when no line comment style is defined for the mode.
blockComment(from: {line, ch}, to: {line, ch}, ?options: object)
Wrap the code in the given range in a block comment. Will fall back to lineComment when no block comment style is defined for the mode.
uncomment(from: {line, ch}, to: {line, ch}, ?options: object) → boolean
Try to uncomment the given range. Returns true if a comment range was found and removed, false otherwise.
The options object accepted by these methods may have the following properties:
blockCommentStart, blockCommentEnd, blockCommentLead, lineComment: string
Override the comment string properties of the mode with custom comment strings.
padding: string
A string that will be inserted after opening and before closing comment markers. Defaults to a single space.
commentBlankLines: boolean
Whether, when adding line comments, to also comment lines that contain only whitespace.
indent: boolean
When adding line comments and this is turned on, it will align the comment block to the current indentation of the first line of the block.
fullLines: boolean
When block commenting, this controls whether the whole lines are indented, or only the precise range that is given. Defaults to true.
The addon also defines a toggleComment command, which will try to uncomment the current selection, and if that fails, line-comments it.
fold/foldcode.js
Helps with code folding. Adds a foldCode method to editor instances, which will try to do a code fold starting at the given line, or unfold the fold that is already present. The method takes as first argument the position that should be folded (may be a line number or a Pos), and as second optional argument either a range-finder function, or an options object, supporting the following properties:
rangeFinder: fn(CodeMirror, Pos)
The function that is used to find foldable ranges. If this is not directly passed, it will default to CodeMirror.fold.auto, which uses getHelpers with a "fold" type to find folding functions appropriate for the local mode. There are files in the addon/fold/ directory providing CodeMirror.fold.brace, which finds blocks in brace languages (JavaScript, C, Java, etc), CodeMirror.fold.indent, for languages where indentation determines block structure (Python, Haskell), and CodeMirror.fold.xml, for XML-style languages, and CodeMirror.fold.comment, for folding comment blocks.
widget: string|Element
The widget to show for folded ranges. Can be either a string, in which case it'll become a span with class CodeMirror-foldmarker, or a DOM node.
scanUp: boolean
When true (default is false), the addon will try to find foldable ranges on the lines above the current one if there isn't an eligible one on the given line.
minFoldSize: integer
The minimum amount of lines that a fold should span to be accepted. Defaults to 0, which also allows single-line folds.
See the demo for an example.
fold/foldgutter.js
Provides an option foldGutter, which can be used to create a gutter with markers indicating the blocks that can be folded. Create a gutter using the gutters option, giving it the class CodeMirror-foldgutter or something else if you configure the addon to use a different class, and this addon will show markers next to folded and foldable blocks, and handle clicks in this gutter. Note that CSS styles should be applied to make the gutter, and the fold markers within it, visible. A default set of CSS styles are available in: addon/fold/foldgutter.css . The option can be either set to true, or an object containing the following optional option fields:
gutter: string
The CSS class of the gutter. Defaults to "CodeMirror-foldgutter". You will have to style this yourself to give it a width (and possibly a background). See the default gutter style rules above.
indicatorOpen: string | Element
A CSS class or DOM element to be used as the marker for open, foldable blocks. Defaults to "CodeMirror-foldgutter-open".
indicatorFolded: string | Element
A CSS class or DOM element to be used as the marker for folded blocks. Defaults to "CodeMirror-foldgutter-folded".
rangeFinder: fn(CodeMirror, Pos)
The range-finder function to use when determining whether something can be folded. When not given, CodeMirror.fold.auto will be used as default.
Demo here.
runmode/runmode.js
Can be used to run a CodeMirror mode over text without actually opening an editor instance. See the demo for an example. There are alternate versions of the file avaible for running stand-alone (without including all of CodeMirror) and for running under node.js.
runmode/colorize.js
Provides a convenient way to syntax-highlight code snippets in a webpage. Depends on the runmode addon (or its standalone variant). Provides a CodeMirror.colorize function that can be called with an array (or other array-ish collection) of DOM nodes that represent the code snippets. By default, it'll get all pre tags. Will read the data-lang attribute of these nodes to figure out their language, and syntax-color their content using the relevant CodeMirror mode (you'll have to load the scripts for the relevant modes yourself). A second argument may be provided to give a default mode, used when no language attribute is found for a node. Used in this manual to highlight example code.
mode/overlay.js
Mode combinator that can be used to extend a mode with an 'overlay' — a secondary mode is run over the stream, along with the base mode, and can color specific pieces of text without interfering with the base mode. Defines CodeMirror.overlayMode, which is used to create such a mode. See this demo for a detailed example.
mode/multiplex.js
Mode combinator that can be used to easily 'multiplex' between several modes. Defines CodeMirror.multiplexingMode which, when given as first argument a mode object, and as other arguments any number of {open, close, mode [, delimStyle, innerStyle]} objects, will return a mode object that starts parsing using the mode passed as first argument, but will switch to another mode as soon as it encounters a string that occurs in one of the open fields of the passed objects. When in a sub-mode, it will go back to the top mode again when the close string is encountered. Pass "\n" for open or close if you want to switch on a blank line.
  • When delimStyle is specified, it will be the token style returned for the delimiter tokens.
  • When innerStyle is specified, it will be the token style added for each inner mode token.
The outer mode will not see the content between the delimiters. See this demo for an example.
hint/show-hint.js
Provides a framework for showing autocompletion hints. Defines CodeMirror.showHint, which takes a CodeMirror instance, a hinting function, and optionally an options object, and pops up a widget that allows the user to select a completion. Hinting functions are function that take an editor instance and an optional options object, and return a {list, from, to} object, where list is an array of strings or objects (the completions), and from and to give the start and end of the token that is being completed as {line, ch} objects.
If no hinting function is given, the addon will use CodeMirror.hint.auto, with calls getHelpers with the "hint" type to find applicable hinting functions, and tries them one by one. If that fails, it looks for a "hintWords" helper to fetch a list of completable words for the mode, and uses CodeMirror.hint.fromList to complete from those.
When completions aren't simple strings, they should be objects with the following properties:
text: string
The completion text. This is the only required property.
displayText: string
The text that should be displayed in the menu.
className: string
A CSS class name to apply to the completion's line in the menu.
render: fn(Element, self, data)
A method used to create the DOM structure for showing the completion by appending it to its first argument.
hint: fn(CodeMirror, self, data)
A method used to actually apply the completion, instead of the default behavior.
from: {line, ch}
Optional from position that will be used by pick() instead of the global one passed with the full list of completions.
to: {line, ch}
Optional to position that will be used by pick() instead of the global one passed with the full list of completions.
The plugin understands the following options (the options object will also be passed along to the hinting function, which may understand additional options):
async: boolean
When set to true, the hinting function's signature should be (cm, callback, ?options), and the completion interface will only be popped up when the hinting function calls the callback, passing it the object holding the completions.
completeSingle: boolean
Determines whether, when only a single completion is available, it is completed without showing the dialog. Defaults to true.
alignWithWord: boolean
Whether the pop-up should be horizontally aligned with the start of the word (true, default), or with the cursor (false).
closeOnUnfocus: boolean
When enabled (which is the default), the pop-up will close when the editor is unfocused.
customKeys: keymap
Allows you to provide a custom key map of keys to be active when the pop-up is active. The handlers will be called with an extra argument, a handle to the completion menu, which has moveFocus(n), setFocus(n), pick(), and close() methods (see the source for details), that can be used to change the focused element, pick the current element or close the menu. Additionnaly menuSize() can give you access to the size of the current dropdown menu, length give you the number of availlable completions, and data give you full access to the completion returned by the hinting function.
extraKeys: keymap
Like customKeys above, but the bindings will be added to the set of default bindings, instead of replacing them.
The following events will be fired on the completions object during completion:
"shown" ()
Fired when the pop-up is shown.
"select" (completion, Element)
Fired when a completion is selected. Passed the completion value (string or object) and the DOM node that represents it in the menu.
"pick" (completion)
Fired when a completion is picked. Passed the completion value (string or object).
"close" ()
Fired when the completion is finished.
This addon depends on styles from addon/hint/show-hint.css. Check out the demo for an example.
hint/javascript-hint.js
Defines a simple hinting function for JavaScript (CodeMirror.hint.javascript) and CoffeeScript (CodeMirror.hint.coffeescript) code. This will simply use the JavaScript environment that the editor runs in as a source of information about objects and their properties.
hint/xml-hint.js
Defines CodeMirror.hint.xml, which produces hints for XML tagnames, attribute names, and attribute values, guided by a schemaInfo option (a property of the second argument passed to the hinting function, or the third argument passed to CodeMirror.showHint).
The schema info should be an object mapping tag names to information about these tags, with optionally a "!top" property containing a list of the names of valid top-level tags. The values of the properties should be objects with optional properties children (an array of valid child element names, omit to simply allow all tags to appear) and attrs (an object mapping attribute names to null for free-form attributes, and an array of valid values for restricted attributes). Demo here.
hint/html-hint.js
Provides schema info to the xml-hint addon for HTML documents. Defines a schema object CodeMirror.htmlSchema that you can pass to as a schemaInfo option, and a CodeMirror.hint.html hinting function that automatically calls CodeMirror.hint.xml with this schema data. See the demo.
hint/css-hint.js
A hinting function for CSS, SCSS, or LESS code. Defines CodeMirror.hint.css.
hint/python-hint.js
A very simple hinting function for Python code. Defines CodeMirror.hint.python.
hint/anyword-hint.js
A very simple hinting function (CodeMirror.hint.anyword) that simply looks for words in the nearby code and completes to those. Takes two optional options, word, a regular expression that matches words (sequences of one or more character), and range, which defines how many lines the addon should scan when completing (defaults to 500).
hint/sql-hint.js
A simple SQL hinter. Defines CodeMirror.hint.sql.
search/match-highlighter.js
Adds a highlightSelectionMatches option that can be enabled to highlight all instances of a currently selected word. Can be set either to true or to an object containing the following options: minChars, for the minimum amount of selected characters that triggers a highlight (default 2), style, for the style to be used to highlight the matches (default "matchhighlight", which will correspond to CSS class cm-matchhighlight), and showToken which can be set to true or to a regexp matching the characters that make up a word. When enabled, it causes the current word to be highlighted when nothing is selected (defaults to off). Demo here.
lint/lint.js
Defines an interface component for showing linting warnings, with pluggable warning sources (see json-lint.js, javascript-lint.js, and css-lint.js in the same directory). Defines a lint option that can be set to a warning source (for example CodeMirror.lint.javascript), or to true, in which case getHelper with type "lint" is used to determined a validator function. Depends on addon/lint/lint.css. A demo can be found here.
selection/mark-selection.js
Causes the selected text to be marked with the CSS class CodeMirror-selectedtext when the styleSelectedText option is enabled. Useful to change the colour of the selection (in addition to the background), like in this demo.
selection/active-line.js
Defines a styleActiveLine option that, when enabled, gives the wrapper of the active line the class CodeMirror-activeline, and adds a background with the class CodeMirror-activeline-background. is enabled. See the demo.
mode/loadmode.js
Defines a CodeMirror.requireMode(modename, callback) function that will try to load a given mode and call the callback when it succeeded. You'll have to set CodeMirror.modeURL to a string that mode paths can be constructed from, for example "mode/%N/%N.js"—the %N's will be replaced with the mode name. Also defines CodeMirror.autoLoadMode(instance, mode), which will ensure the given mode is loaded and cause the given editor instance to refresh its mode when the loading succeeded. See the demo.
comment/continuecomment.js
Adds a continueComments option, which sets whether the editor will make the next line continue a comment when you press Enter inside a comment block. Can be set to a boolean to enable/disable this functionality. Set to a string, it will continue comments using a custom shortcut. Set to an object, it will use the key property for a custom shortcut and the boolean continueLineComment property to determine whether single-line comments should be continued (defaulting to true).
display/placeholder.js
Adds a placeholder option that can be used to make text appear in the editor when it is empty and not focused. Also gives the editor a CodeMirror-empty CSS class whenever it doesn't contain any text. See the demo.
display/fullscreen.js
Defines an option fullScreen that, when set to true, will make the editor full-screen (as in, taking up the whole browser window). Depends on fullscreen.css. Demo here.
display/rulers.js
Adds a rulers option, which can be used to show one or more vertical rulers in the editor. The option, if defined, should be given an array of {column, className} objects or numbers. The ruler will be displayed at the column indicated by the number or the column property. The className property can be used to assign a custom style to a ruler. Demo here.
wrap/hardwrap.js
Addon to perform hard line wrapping/breaking for paragraphs of text. Adds these methods to editor instances:
wrapParagraph(?pos: {line, ch}, ?options: object)
Wraps the paragraph at the given position. If pos is not given, it defaults to the cursor position.
wrapRange(from: {line, ch}, to: {line, ch}, ?options: object)
Wraps the given range as one big paragraph.
wrapParagraphsInRange(from: {line, ch}, to: {line, ch}, ?options: object)
Wrapps the paragraphs in (and overlapping with) the given range individually.
The following options are recognized:
paragraphStart, paragraphEnd: RegExp
Blank lines are always considered paragraph boundaries. These options can be used to specify a pattern that causes lines to be considered the start or end of a paragraph.
column: number
The column to wrap at. Defaults to 80.
wrapOn: RegExp
A regular expression that matches only those two-character strings that allow wrapping. By default, the addon wraps on whitespace and after dash characters.
killTrailingSpace: boolean
Whether trailing space caused by wrapping should be preserved, or deleted. Defaults to true.
A demo of the addon is available here.
merge/merge.js
Implements an interface for merging changes, using either a 2-way or a 3-way view. The CodeMirror.MergeView constructor takes arguments similar to the CodeMirror constructor, first a node to append the interface to, and then an options object. Two extra optional options are recognized, origLeft and origRight, which may be strings that provide original versions of the document, which will be shown to the left and right of the editor in non-editable CodeMirror instances. The merge interface will highlight changes between the editable document and the original(s) (demo).
tern/tern.js
Provides integration with the Tern JavaScript analysis engine, for completion, definition finding, and minor refactoring help. See the demo for a very simple integration. For more involved scenarios, see the comments at the top of the addon and the implementation of the (multi-file) demonstration on the Tern website.

Writing CodeMirror Modes

Modes typically consist of a single JavaScript file. This file defines, in the simplest case, a lexer (tokenizer) for your language—a function that takes a character stream as input, advances it past a token, and returns a style for that token. More advanced modes can also handle indentation for the language.

The mode script should call CodeMirror.defineMode to register itself with CodeMirror. This function takes two arguments. The first should be the name of the mode, for which you should use a lowercase string, preferably one that is also the name of the files that define the mode (i.e. "xml" is defined in xml.js). The second argument should be a function that, given a CodeMirror configuration object (the thing passed to the CodeMirror function) and an optional mode configuration object (as in the mode option), returns a mode object.

Typically, you should use this second argument to defineMode as your module scope function (modes should not leak anything into the global scope!), i.e. write your whole mode inside this function.

The main responsibility of a mode script is parsing the content of the editor. Depending on the language and the amount of functionality desired, this can be done in really easy or extremely complicated ways. Some parsers can be stateless, meaning that they look at one element (token) of the code at a time, with no memory of what came before. Most, however, will need to remember something. This is done by using a state object, which is an object that is always passed when reading a token, and which can be mutated by the tokenizer.

Modes that use a state must define a startState method on their mode object. This is a function of no arguments that produces a state object to be used at the start of a document.

The most important part of a mode object is its token(stream, state) method. All modes must define this method. It should read one token from the stream it is given as an argument, optionally update its state, and return a style string, or null for tokens that do not have to be styled. For your styles, you are encouraged to use the 'standard' names defined in the themes (without the cm- prefix). If that fails, it is also possible to come up with your own and write your own CSS theme file.

A typical token string would be "variable" or "comment". Multiple styles can be returned (separated by spaces), for example "string error" for a thing that looks like a string but is invalid somehow (say, missing its closing quote). When a style is prefixed by "line-" or "line-background-", the style will be applied to the whole line, analogous to what the addLineClass method does—styling the "text" in the simple case, and the "background" element when "line-background-" is prefixed.

The stream object that's passed to token encapsulates a line of code (tokens may never span lines) and our current position in that line. It has the following API:

eol() → boolean
Returns true only if the stream is at the end of the line.
sol() → boolean
Returns true only if the stream is at the start of the line.
peek() → string
Returns the next character in the stream without advancing it. Will return an null at the end of the line.
next() → string
Returns the next character in the stream and advances it. Also returns null when no more characters are available.
eat(match: string|regexp|function(char: string) → boolean) → string
match can be a character, a regular expression, or a function that takes a character and returns a boolean. If the next character in the stream 'matches' the given argument, it is consumed and returned. Otherwise, undefined is returned.
eatWhile(match: string|regexp|function(char: string) → boolean) → boolean
Repeatedly calls eat with the given argument, until it fails. Returns true if any characters were eaten.
eatSpace() → boolean
Shortcut for eatWhile when matching white-space.
skipToEnd()
Moves the position to the end of the line.
skipTo(ch: string) → boolean
Skips to the next occurrence of the given character, if found on the current line (doesn't advance the stream if the character does not occur on the line). Returns true if the character was found.
match(pattern: string, ?consume: boolean, ?caseFold: boolean) → boolean
match(pattern: regexp, ?consume: boolean) → array<string>
Act like a multi-character eat—if consume is true or not given—or a look-ahead that doesn't update the stream position—if it is false. pattern can be either a string or a regular expression starting with ^. When it is a string, caseFold can be set to true to make the match case-insensitive. When successfully matching a regular expression, the returned value will be the array returned by match, in case you need to extract matched groups.
backUp(n: integer)
Backs up the stream n characters. Backing it up further than the start of the current token will cause things to break, so be careful.
column() → integer
Returns the column (taking into account tabs) at which the current token starts.
indentation() → integer
Tells you how far the current line has been indented, in spaces. Corrects for tab characters.
current() → string
Get the string between the start of the current token and the current stream position.

By default, blank lines are simply skipped when tokenizing a document. For languages that have significant blank lines, you can define a blankLine(state) method on your mode that will get called whenever a blank line is passed over, so that it can update the parser state.

Because state object are mutated, and CodeMirror needs to keep valid versions of a state around so that it can restart a parse at any line, copies must be made of state objects. The default algorithm used is that a new state object is created, which gets all the properties of the old object. Any properties which hold arrays get a copy of these arrays (since arrays tend to be used as mutable stacks). When this is not correct, for example because a mode mutates non-array properties of its state object, a mode object should define a copyState method, which is given a state and should return a safe copy of that state.

If you want your mode to provide smart indentation (through the indentLine method and the indentAuto and newlineAndIndent commands, to which keys can be bound), you must define an indent(state, textAfter) method on your mode object.

The indentation method should inspect the given state object, and optionally the textAfter string, which contains the text on the line that is being indented, and return an integer, the amount of spaces to indent. It should usually take the indentUnit option into account. An indentation method may return CodeMirror.Pass to indicate that it could not come up with a precise indentation.

To work well with the commenting addon, a mode may define lineComment (string that starts a line comment), blockCommentStart, blockCommentEnd (strings that start and end block comments), and blockCommentLead (a string to put at the start of continued lines in a block comment). All of these are optional.

Finally, a mode may define either an electricChars or an electricInput property, which are used to automatically reindent the line when certain patterns are typed and the electricChars option is enabled. electricChars may be a string, and will trigger a reindent whenever one of the characters in that string are typed. Often, it is more appropriate to use electricInput, which should hold a regular expression, and will trigger indentation when the part of the line before the cursor matches the expression. It should usually end with a $ character, so that it only matches when the indentation-changing pattern was just typed, not when something was typed after the pattern.

So, to summarize, a mode must provide a token method, and it may provide startState, copyState, and indent methods. For an example of a trivial mode, see the diff mode, for a more involved example, see the C-like mode.

Sometimes, it is useful for modes to nest—to have one mode delegate work to another mode. An example of this kind of mode is the mixed-mode HTML mode. To implement such nesting, it is usually necessary to create mode objects and copy states yourself. To create a mode object, there are CodeMirror.getMode(options, parserConfig), where the first argument is a configuration object as passed to the mode constructor function, and the second argument is a mode specification as in the mode option. To copy a state object, call CodeMirror.copyState(mode, state), where mode is the mode that created the given state.

In a nested mode, it is recommended to add an extra method, innerMode which, given a state object, returns a {state, mode} object with the inner mode and its state for the current position. These are used by utility scripts such as the tag closer to get context information. Use the CodeMirror.innerMode helper function to, starting from a mode and a state, recursively walk down to the innermost mode and state.

To make indentation work properly in a nested parser, it is advisable to give the startState method of modes that are intended to be nested an optional argument that provides the base indentation for the block of code. The JavaScript and CSS parser do this, for example, to allow JavaScript and CSS code inside the mixed-mode HTML mode to be properly indented.

It is possible, and encouraged, to associate your mode, or a certain configuration of your mode, with a MIME type. For example, the JavaScript mode associates itself with text/javascript, and its JSON variant with application/json. To do this, call CodeMirror.defineMIME(mime, modeSpec), where modeSpec can be a string or object specifying a mode, as in the mode option.

If a mode specification wants to add some properties to the resulting mode object, typically for use with getHelpers, it may contain a modeProps property, which holds an object. This object's properties will be copied to the actual mode object.

Sometimes, it is useful to add or override mode object properties from external code. The CodeMirror.extendMode function can be used to add properties to mode objects produced for a specific mode. Its first argument is the name of the mode, its second an object that specifies the properties that should be added. This is mostly useful to add utilities that can later be looked up through getMode.