Add support for ignoring collapsable white-space

wooorm · wooorm · commit 850a2fe41740 · 2017-02-10T16:59:04.000+01:00
HTML often collapsed multiple white-space into one, or ignores it
completely (such as between block-level elements).  Which means
HTML documents are written with that in mind.  So I think
that superfluous white-space should also be stripped when transforming
to markdown.
diff --git a/handlers/table.js b/handlers/table.js
@@ -50,6 +50,7 @@ function cellsBefore(parent, node) {
       break;
     }
 
+    /* istanbul ignore else - When proper HTML, should always be a cell */
     if (cell(child)) {
       pos++;
     }
diff --git a/handlers/text.js b/handlers/text.js
@@ -2,13 +2,6 @@
 
 module.exports = text;
 
-var u = require('unist-builder');
-var trimLines = require('trim-lines');
-
 function text(h, node) {
-  if (node.value.match(/(\n+)/g)) {
-    return null;
-  }
-
-  return h.augment(node, u('text', trimLines(node.value)));
+  return h.augment(node, {type: 'text', value: node.value});
 }
diff --git a/index.js b/index.js
@@ -2,17 +2,19 @@
 
 module.exports = toMDAST;
 
+var minify = require('rehype-minify-whitespace')();
 var xtend = require('xtend');
 var one = require('./one');
 
 h.augment = augment;
 
 function toMDAST(tree) {
-  return one(h, tree);
+  return one(h, minify(tree));
 }
 
 function h(node, type, props, children) {
   var result;
+
   if (!children && ((typeof props === 'object' && 'length' in props) || typeof props === 'string')) {
     children = props;
     props = {};
diff --git a/one.js b/one.js
@@ -2,7 +2,6 @@
 
 module.exports = one;
 
-var u = require('unist-builder');
 var has = require('has');
 var all = require('./all');
 var handlers = require('./handlers');
@@ -21,7 +20,7 @@ function one(h, node, parent) {
 
 function unknown(h, node) {
   if (node.value) {
-    return h.augment(node, u('text', node.value));
+    return h.augment(node, {type: 'text', value: node.value});
   }
 
   return all(h, node);
diff --git a/package.json b/package.json
@@ -30,8 +30,7 @@
     "hast-util-has-property": "^1.0.0",
     "hast-util-is-element": "^1.0.0",
     "hast-util-to-string": "^1.0.0",
-    "trim-lines": "^1.1.0",
-    "unist-builder": "^1.0.2",
+    "rehype-minify-whitespace": "^1.1.0",
     "unist-util-visit": "^1.1.1",
     "xtend": "^4.0.1"
   },
diff --git a/tests/fixtures/whitespace/index.html b/tests/fixtures/whitespace/index.html
@@ -0,0 +1,23 @@
+<!DOCTYPE html>
+<html>
+  <head>
+    <title>Alpha</title>
+    <meta charset="utf8">
+  </head>
+  <body>
+    <section>
+      <p>
+          Bravo  Charlie
+      </p>
+      <p>
+      </p>
+      <p>Delta
+      </p>
+      <p>
+      Echo</p>
+      <p>
+        <strong>Foxtrot</strong>  <em>and</em> <code>Golf</code>
+      </p>
+    </section>
+  </body>
+</html>
diff --git a/tests/fixtures/whitespace/index.json b/tests/fixtures/whitespace/index.json
@@ -0,0 +1,3 @@
+{
+  "tree": false
+}
diff --git a/tests/fixtures/whitespace/index.md b/tests/fixtures/whitespace/index.md
@@ -0,0 +1,9 @@
+Bravo Charlie
+
+
+
+Delta
+
+Echo
+
+**Foxtrot** _and_ `Golf`
diff --git a/tests/index.js b/tests/index.js
@@ -137,11 +137,13 @@ test('fixtures', function (t) {
         st.deepEqual(remark.stringify(tree), output || '\n', 'should produce the same documents');
       }
 
-      st.deepEqual(
-        tree,
-        remove(remark.run(remark.parse(output)), true),
-        'should produce the same tree as remark'
-      );
+      if (!config || config.tree !== false) {
+        st.deepEqual(
+          tree,
+          remove(remark.run(remark.parse(output)), true),
+          'should produce the same tree as remark'
+        );
+      }
 
       st.end();
     });

Original file line number	Diff line number	Diff line change
`@@ -50,6 +50,7 @@ function cellsBefore(parent, node) {`
`50`	`50`	`break;`
`51`	`51`	`}`
`52`	`52`
	`53`	`+ /* istanbul ignore else - When proper HTML, should always be a cell */`
`53`	`54`	`if (cell(child)) {`
`54`	`55`	`pos++;`
`55`	`56`	`}`
-Original file line number
+Diff line change
@@ @@ -0,0 +1,9 @@ @@
 +Bravo Charlie
++
++
++
 +Delta
++
 +Echo
++
 +**Foxtrot** _and_ `Golf`