Replace sax with @rgrove/parse-xml

Closes GH-3.

This commit switches from `sax`, a lax parser, to `parse-xml`, a proper and
fast parser.

This does mean that invalid XML can no longer be passed, so it’s a breaking
change (for example, all documents now require a root element, doctypes need to
be capitalized, and CDATA around the root element is no longer okay).
It also means that whitespace around the root element is no longer present
in the tree: XML requires stripping that.

It does mean that positional info is now much better, processing instructions
are supported, and the size is cut by a lot.

Author: wooorm

lib/index.js

@@ -38,10 +38,9 @@
     "index.js"
   ],
   "dependencies": {
-    "@types/sax": "^1.0.0",
-    "@types/unist": "^2.0.0",
+    "@rgrove/parse-xml": "^4.1.0",
     "@types/xast": "^1.0.0",
-    "sax": "^1.0.0",
+    "vfile-location": "^4.0.0",
     "vfile-message": "^3.0.0"
   },
   "devDependencies": {
@@ -75,26 +74,25 @@
   "xo": {
     "prettier": true,
     "rules": {
-      "unicorn/prefer-code-point": "off"
+      "unicorn/prefer-switch": "off"
     },
     "overrides": [
       {
         "files": "test/**/*.js",
         "rules": {
-          "no-await-in-loop": 0
+          "no-await-in-loop": "off"
         }
       }
     ]
   },
   "remarkConfig": {
     "plugins": [
-      "preset-wooorm"
+      "remark-preset-wooorm"
     ]
   },
   "typeCoverage": {
     "atLeast": 100,
     "detail": true,
-    "strict": true,
-    "ignoreCatch": true
+    "strict": true
   }
 }

package.json

@@ -27,15 +27,14 @@
 
 ## What is this?
 
-This package is a utility that takes XML input and turns it into a [xast][]
-syntax tree.
-It uses [`sax`][sax], which turns XML into events, while it turns those events
-into nodes.
+This package is a utility that takes serialized XML as input and turns it into
+a [xast][] syntax tree.
+It uses [`@rgrove/parse-xml`][parse-xml], which is a good and fast XML parser,
+and turns its results into xast.
 
 ## When should I use this?
 
-If you want to handle syntax trees, use this.
-Use [`sax`][sax] itself instead when you want to do other things.
+If you want to use xast syntax trees, use this.
 
 The utility [`xast-util-to-xml`][xast-util-to-xml] does the inverse of this
 utility.
@@ -84,7 +83,7 @@ import {fromXml} from 'xast-util-from-xml'
 
 const tree = fromXml(await fs.readFile('example.xml'))
 
-console.log(tree)
+console.dir(tree, {depth: null})
 ```
 
 …now running `node example.js` yields (positional info removed for brevity):
@@ -121,15 +120,14 @@ console.log(tree)
         },
         {type: 'text', value: '\n'}
       ]
-    },
-    {type: 'text', value: '\n'}
+    }
   ]
 }
 ```
 
 ## API
 
-This package exports the identifier [`fromXml`][fromxml].
+This package exports the identifier [`fromXml`][api-from-xml].
 There is no default export.
 
 ### `fromXml(value)`
@@ -236,8 +234,8 @@ abide by its terms.
 
 [root]: https://github.com/syntax-tree/xast#root
 
-[sax]: https://github.com/isaacs/sax-js
+[parse-xml]: https://github.com/rgrove/parse-xml
 
 [xast-util-to-xml]: https://github.com/syntax-tree/xast-util-to-xml
 
-[fromxml]: #fromxmlvalue
+[api-from-xml]: #fromxmlvalue

readme.md

@@ -109,22 +109,18 @@
           "offset": 144
         }
       }
+    }
+  ],
+  "position": {
+    "start": {
+      "line": 1,
+      "column": 1,
+      "offset": 0
     },
-    {
-      "type": "text",
-      "value": "\n",
-      "position": {
-        "start": {
-          "line": 10,
-          "column": 8,
-          "offset": 144
-        },
-        "end": {
-          "line": 11,
-          "column": 1,
-          "offset": 145
-        }
-      }
+    "end": {
+      "line": 11,
+      "column": 1,
+      "offset": 145
     }
-  ]
+  }
 }