Simple Mode Demo
The mode/simple
addon allows CodeMirror modes to be specified using a relatively simple
declarative format. This format is not as powerful as writing code
directly against the mode
interface, but is a lot easier to get started with, and
sufficiently expressive for many simple language modes.
This interface is still in flux. It is unlikely to be scrapped or overhauled completely, so do start writing code against it, but details might change as it stabilizes, and you might have to tweak your code when upgrading.
Simple modes (loosely based on the Common JavaScript Syntax Highlighting Specification, which never took off), are state machines, where each state has a number of rules that match tokens. A rule describes a type of token that may occur in the current state, and possibly a transition to another state caused by that token.
The CodeMirror.defineSimpleMode(name, states)
method
takes a mode name and an object that describes the mode's states. The
editor below shows an example of such a mode (and is itself
highlighted by the mode shown in it).
Each state is an array of rules. A rule may have the following properties:
regex: string | RegExp
- The regular expression that matches the token. May be a string
or a regex object. When a regex, the
ignoreCase
flag will be taken into account when matching the token. This regex has to capture groups when thetoken
property is an array. If it captures groups, it must capture all of the string (since JS provides no way to find out where a group matched). Currently negative lookbehind assertion for regex is not supported, regardless of browser support. token
: string | array<string> | null- An optional token style. Multiple styles can be specified by
separating them with dots or spaces. When this property holds an array of token styles,
the
regex
for this rule must capture a group for each array item. sol
: boolean- When true, this token will only match at the start of the line.
(The
^
regexp marker doesn't work as you'd expect in this context because of limitations in JavaScript's RegExp API.) next: string
- When a
next
property is present, the mode will transfer to the state named by the property when the token is encountered. push: string
- Like
next
, but instead replacing the current state by the new state, the current state is kept on a stack, and can be returned to with thepop
directive. pop: bool
- When true, and there is another state on the state stack, will cause the mode to pop that state off the stack and transition to it.
mode: {spec, end, persistent}
- Can be used to embed another mode inside a mode. When present,
must hold an object with a
spec
property that describes the embedded mode, and an optionalend
end property that specifies the regexp that will end the extent of the mode. When apersistent
property is set (and true), the nested mode's state will be preserved between occurrences of the mode. indent: bool
- When true, this token changes the indentation to be one unit more than the current line's indentation.
dedent: bool
- When true, this token will pop one scope off the indentation stack.
dedentIfLineStart: bool
- If a token has its
dedent
property set, it will, by default, cause lines where it appears at the start to be dedented. Set this property to false to prevent that behavior.
The meta
property of the states object is special, and
will not be interpreted as a state. Instead, properties set on it will
be set on the mode, which is useful for properties
like lineComment
,
which sets the comment style for a mode. The simple mode addon also
recognizes a few such properties:
dontIndentStates: array<string>
- An array of states in which the mode's auto-indentation should not take effect. Usually used for multi-line comment and string states.