Feature/#26 template expr parser (#87)

* first draft of expression parser

* fix preserve whitespace, fix unstable formatting result when node was broken after first text element

* fix #78

* v0.8.5

* simplify formatter rules

* align first text node

* reset example

* remove unused var

* reimplement formatting rule for condition comments

* remove invalid comments

* v0.8.6

* clean attribute expression

* update readme

* update readme

* Merge branch 'master' into feature/#26_temlate_expr_parser

# Conflicts:
#	packages/prettyhtml-formatter/package-lock.json

* fix typo

* fix merge, cleanup comments

* dont enable attribute clean up by default

Author: StarpTech

README.md

@@ -35,6 +35,7 @@ Opinionated general formatter for your Angular, Vue, Svelte or pure HTML5 templa
 - [prettyhtml-sort-attributes](/packages/prettyhtml-sort-attributes) Sort attributes alphabetically.
 - [prettyhtml-quick](/packages/prettyhtml-quick) Formats your changed files based on Git.
 - [webparser](/packages/webparser) Optimized HTML parser for formatters
+- [expression-parser](/packages/expression-parser) Framework agnostic template expression parser.
 - [rehype-webparser](/packages/rehype-webparser) Adapter between HTML parser and rehype.
 - [rehype-minify-whitespace](/packages/rehype-minify-whitespace) Collapse whitespace.
 - [hast-util-from-parse](/packages/hast-util-from-webparser) Transform [webparser](/packages/webparser) AST to HAST.
@@ -50,7 +51,7 @@ Adding this flag before a tag will preserve from whitespace and/or attribute wra
 <div></div>
 ```
 
-2. Preserve only from whitespace processing. This excludes no indentation.
+2. Preserve only from whitespace processing. This excludes indentation.
 
 ```html
 <!--prettyhtml-preserve-whitespace-->

packages/expression-parser/README.md

@@ -0,0 +1,53 @@
+# @starptech/expression-parser
+
+Framework agnostic template expression parser
+
+## Installation
+
+```
+npm install --save @starptech/expression-parser
+```
+
+## Usage
+
+This example shows how we parse HTML
+
+```js
+const parse = require('@starptech/expression-parser')
+const result = parse(`{ a + /<g></g> b }`, { brackets: ['{', '}'] })
+```
+
+## Representation
+
+- `unescape`: Indentify a template marker as escaped. This information is useful to understand why a marker was skipped
+- `expressions`: A list of template expressions
+  - `start`: Start position
+  - `end`: End position
+  - `text`: The content of the expression
+
+```json
+{
+  "unescape": "",
+  "expressions": [
+    {
+      "start": 1,
+      "end": 19,
+      "text": " a + /<g></g> b "
+    }
+  ]
+}
+```
+
+## Details
+
+There may be more than one expression as part of one attribute value or text node, or only one replacing the entire value or node.
+
+When used as the whole attribute value, there's no need to enclose the expression inside quotes, even if the expression contains whitespace.
+
+Single and double quotes can be nested inside the expression.
+
+To emit opening (left) brackets as literal text wherever an opening bracket is expected, the bracket must be prefixed with a backslash (the JS escape char '\'). This character is preserved in the output, but the parser will add a replace property for the attribute or node containing the escaped bracket, whose value is the bracket itself.
+
+## Credits
+
+The parser is a modificated version of [Riot](https://github.com/riot/parser) template expression parser.