edge-lexer

Edge lexer

Generating high level tokens from Edge whitelisted markup

Edge lexer produces a list of tokens by scanning for Edge whitelisted syntax.

This module is a blend of a lexer and an AST generator, since Edge doesn't need a pure lexer that scans for each character. Edge markup is written within other markup languages like HTML or Markdown and walking over each character is waste of resources.

Instead, this module starts by detecting for the Edge whitelisted syntax and then starts the lexical analysis within the detected markup.

Table of contents

• Highlights
• Performance
• Usage
• Terms used
• Tokens
  • Tag Token
  • Escaped Tag Token
  • Raw Token
  • Comment Token
  • NewLine Token
  • Mustache Token
  • Safe Mustache Token
  • Escaped Mustache Token
  • Escaped Safe Mustache Token
• Properties
  • BlockProp
  • Prop
• Mustache expressions
• Errors
• Example
• API Docs
• Raised exceptions
• Maintainers

Highlights

• Zero dependencies (Actually one dependency that is also to standardize edge errors).
• Just uses one regex statement. That also tested against safe-regex for ReDOS
• Allows multiline expressions
• Collects line and columns for accurate stack traces.
• Detects for unclosed tags.
• Detects for unwrapped expressions and raises appropriate errors.

---

Performance

Following measures are taken to keep the analysis performant.

1. Only analyse markup that is detected as Edge whitelisted syntax.
2. Only analyse tags, that are passed to the tokenizer. Which means even if the syntax for tags is whitelisted, the tokeniser will analyse them if they are used by your app.
3. Do not analyse Javascript expression and leave that for edge-parser.
4. Only uses one Regular expression.

---

Usage

import { Tokenizer } from 'edge-lexer'

const template = `Hello {{ username }}`
const tags = {
  if: {
    block: true,
    seekable: true,
  }
}

// Filename is required to add it to error messages
const options = {
  filename: 'welcome.edge'
}

const tokenizer = new Tokenizer(template, tags, options)

tokenizer.parse()
console.log(tokenizer.tokens)

---

Terms used

This guide makes use of the following terms to identify core pieces of the tokenizer.

Term	Token Type	Description
Tag	tag	Tags are used to define logical blocks in the template engine. For example if tag or include tag.
Escaped Tag	e__tag	Escaped tag, Edge will not evaluate it at runtime.
Mustache	mustache	Javascript expression wrapped in curly braces. {{ }}
Safe Mustache	s__mustache	Safe mustache, that doesn't escape the output {{{ }}}
Escaped Mustache	e__mustache	Mustache tag that is escaped
Escaped Safe Mustache	es__mustache	Safe Mustache tag that is escaped
Raw	raw	A raw string, which has no meaning for the template engine
NewLine	newline	Newline
Comment	comment	Edge specific comment block. This will be ripped off in the output.

---

Tokens

Following is the list of Nodes returned by the tokenizer.

Tag Token

{
  type: 'tag'
  filename: 'eval.edge',
  loc: {
    start: {
      line: 1,
      col: 4
    },
    end: {
      line: 1,
      col: 13
    }
  },
  properties: BlockProp,
  children: []
}

Escaped Tag Token

{
- type: 'tag',
+ type: 'e__tag',
  filename: 'eval.edge',
  loc: {
    start: {
      line: 1,
      col: 4
    },
    end: {
      line: 1,
      col: 13
    }
  },
  properties: BlockProp,
  children: []
}

Raw Token

{
  type: 'raw',
  filename: 'eval.edge',
  line: number,
  value: string
}

Comment Token

{
  type: 'comment',
  filename: 'eval.edge',
  line: number,
  value: string
}

NewLine Token

{
  type: 'newline',
  line: number
}

Mustache Token

{
  type: 'mustache',
  filename: 'eval.edge',
  loc: {
    start: {
      line: 1,
      col: 4
    },
    end: {
      line: 1,
      col: 13
    }
  },
  properties: Prop
}

Safe Mustache Token

{
- type: 'mustache',
+ type: 's__mustache',
  filename: 'eval.edge',
  loc: {
    start: {
      line: 1,
      col: 4
    },
    end: {
      line: 1,
      col: 13
    }
  },
  properties: Prop
}

Escaped Mustache Token

{
- type: 'mustache',
+ type: 'e__mustache',
  filename: 'eval.edge',
  loc: {
    start: {
      line: 1,
      col: 4
    },
    end: {
      line: 1,
      col: 13
    }
  },
  properties: Prop
}

Escaped Safe Mustache Token

{
- type: 'mustache',
+ type: 'es__mustache',
  filename: 'eval.edge',
  loc: {
    start: {
      line: 1,
      col: 4
    },
    end: {
      line: 1,
      col: 13
    }
  },
  properties: Prop
}

Key	Value	Description
type	string	The type of node determines the behavior of node
loc	object	loc is only present for tags and mustache tokens
line	number	line is not present for tags and mustache tokens
properties	Prop	Meta data for the node. See Properties to more info
value	string	If token is a raw or comment token, then value is the string in the source file
children	array	Array of recursive nodes. Only exists, when token is a tag

---

Properties

The properties Prop is used to define meta data for a given Node. Nodes like raw, comment and newline, doesn't need any metadata.

BlockProp

The block prop is used by the Block node. The only difference from the regular Prop is the addition of selfclosed attribute.

{
  name: string
  jsArg: string,
  selfclosed: boolean
}

Prop

{
  jsArg: string,
}

Key	Description
jsArg	The jsArg is the Javascript expression to evaluate. Whitespaces and newlines are preserved inside the jsArg
selfclosed	Whether or not the tag was selfclosed during usage.

---

Mustache expressions

For mustache nodes props, the name is the type of mustache expressions. The lexer supports 4 mustache expressions.

mustache

{{ username }}

e__mustache (Escaped mustache)

The following expression is ignored by edge. Helpful when you want this expression to be parsed by a frontend template engine

@{{ username }}

s__mustache (Safe mustache)

The following expression output is considered HTML safe.

{{{ '<p> Hello world </p>' }}}

es__mustache (Escaped safe mustache)

@{{{ '<p> Not touched </p>' }}}

---

Errors

Errors raised by the lexer are always an instance of edge-error and will contain following properties.

error.message
error.line
error.col
error.filename
error.code

---

Example

@if(username)
  <h2> Hello {{ username }} </h2>
@endif

The output of the above text will be

[
  {
    "type": "tag",
    "filename": "eval.edge",
    "properties": {
      "name": "if",
      "jsArg": "username",
      "selfclosed": false
    },
    "loc": {
      "start": {
        "line": 1,
        "col": 4
      },
      "end": {
        "line": 1,
        "col": 13
      }
    },
    "children": [
      {
        "type": "raw",
        "value": "<h2> Hello ",
        "line": 2
      },
      {
        "type": "mustache",
        "filename": "eval.edge",
        "properties": {
          "jsArg": " username "
        },
        "loc": {
          "start": {
            "line": 2,
            "col": 13
          },
          "end": {
            "line": 2,
            "col": 25
          }
        }
      },
      {
        "type": "raw",
        "value": " </h2>",
        "line": 2
      }
    ]
  }
]

API Docs

Following are the autogenerated files via Typedoc

• API

Raised exceptions

Following the links to documented error codes raised by the lexer.

• E_CANNOT_SEEK_STATEMENT
• E_UNCLOSED_CURLY_BRACE
• E_UNCLOSED_PAREN
• E_UNCLOSED_TAG
• E_UNOPENED_PAREN

Maintainers

Harminder virk