chore(docs): generate header ids for better linking

- generate ids for all headers
- collect defined anchors
- check broken links (even if the page exists, but the anchor/id does not)

5 files changed, 193 insertions, 62 deletions

diff --git a/docs/spec/domSpec.js b/docs/spec/domSpec.js
index 7bc6a7f4..d10db9dc 100644
--- a/docs/spec/domSpec.js
+++ b/docs/spec/domSpec.js
@@ -1,4 +1,5 @@
 var DOM = require('../src/dom.js').DOM;
+var normalizeHeaderToId = require('../src/dom.js').normalizeHeaderToId;
 
 describe('dom', function() {
   var dom;
@@ -7,6 +8,31 @@ describe('dom', function() {
     dom = new DOM();
   });
 
+  describe('html', function() {
+    it('should add ids to all h tags', function() {
+      dom.html('<h1>Some Header</h1>');
+      expect(dom.toString()).toContain('<h1 id="some-header">Some Header</h1>');
+    });
+
+    it('should collect <a name> anchors too', function() {
+      dom.html('<h2>Xxx <a name="foo"></a> and bar <a name="bar"></a>');
+      expect(dom.anchors).toContain('foo');
+      expect(dom.anchors).toContain('bar');
+    })
+  });
+
+  it('should collect h tag ids', function() {
+    dom.h('Page Title', function() {
+      dom.html('<h1>Second</h1>xxx <h2>Third</h2>');
+      dom.h('Another Header', function() {});
+    });
+
+    expect(dom.anchors).toContain('page-title');
+    expect(dom.anchors).toContain('second');
+    expect(dom.anchors).toContain('second_third');
+    expect(dom.anchors).toContain('another-header');
+  });
+
   describe('h', function() {
 
     it('should render using function', function() {
@@ -25,7 +51,7 @@ describe('dom', function() {
         this.html('<h1>sub-heading</h1>');
       });
       expect(dom.toString()).toContain('<h1 id="heading">heading</h1>');
-      expect(dom.toString()).toContain('<h2>sub-heading</h2>');
+      expect(dom.toString()).toContain('<h2 id="sub-heading">sub-heading</h2>');
     });
 
     it('should properly number nested headings', function() {
@@ -40,12 +66,45 @@ describe('dom', function() {
 
       expect(dom.toString()).toContain('<h1 id="heading">heading</h1>');
       expect(dom.toString()).toContain('<h2 id="heading2">heading2</h2>');
-      expect(dom.toString()).toContain('<h3>heading3</h3>');
+      expect(dom.toString()).toContain('<h3 id="heading2_heading3">heading3</h3>');
 
       expect(dom.toString()).toContain('<h1 id="other1">other1</h1>');
-      expect(dom.toString()).toContain('<h2>other2</h2>');
+      expect(dom.toString()).toContain('<h2 id="other2">other2</h2>');
+    });
+
+
+    it('should add nested ids to all h tags', function() {
+      dom.h('Page Title', function() {
+        dom.h('Second', function() {
+          dom.html('some <h1>Third</h1>');
+        });
+      });
+
+      var resultingHtml = dom.toString();
+      expect(resultingHtml).toContain('<h1 id="page-title">Page Title</h1>');
+      expect(resultingHtml).toContain('<h2 id="second">Second</h2>');
+      expect(resultingHtml).toContain('<h3 id="second_third">Third</h3>');
+    });
+
+  });
+
+
+  describe('normalizeHeaderToId', function() {
+    it('should ignore content in the parenthesis', function() {
+      expect(normalizeHeaderToId('One (more)')).toBe('one');
+    });
+
+    it('should ignore html content', function() {
+      expect(normalizeHeaderToId('Section <a name="section"></a>')).toBe('section');
     });
 
+    it('should ignore special characters', function() {
+      expect(normalizeHeaderToId('Section \'!?')).toBe('section');
+    });
+
+    it('should ignore html entities', function() {
+      expect(normalizeHeaderToId('angular&#39;s-jqlite')).toBe('angulars-jqlite');
+    });
   });
 
 });
diff --git a/docs/spec/ngdocSpec.js b/docs/spec/ngdocSpec.js
index 607745d8..15fd0ba4 100644
--- a/docs/spec/ngdocSpec.js
+++ b/docs/spec/ngdocSpec.js
@@ -262,33 +262,37 @@ describe('ngdoc', function() {
       expect(docs[0].events).toEqual([eventA, eventB]);
       expect(docs[0].properties).toEqual([propA, propB]);
     });
+  });
 
 
+  describe('checkBrokenLinks', function() {
+    var docs;
 
-    describe('links checking', function() {
-      var docs;
-      beforeEach(function() {
-        spyOn(console, 'log');
-        docs = [new Doc({section: 'api', id: 'fake.id1', links: ['non-existing-link']}),
-                new Doc({section: 'api', id: 'fake.id2'}),
-                new Doc({section: 'api', id: 'fake.id3'})];
-      });
-
-      it('should log warning when any link doesn\'t exist', function() {
-        ngdoc.merge(docs);
-        expect(console.log).toHaveBeenCalled();
-        expect(console.log.argsForCall[0][0]).toContain('WARNING:');
-      });
+    beforeEach(function() {
+      spyOn(console, 'log');
+      docs = [new Doc({section: 'api', id: 'fake.id1', anchors: ['one']}),
+              new Doc({section: 'api', id: 'fake.id2'}),
+              new Doc({section: 'api', id: 'fake.id3'})];
+    });
 
-      it('should say which link doesn\'t exist', function() {
-        ngdoc.merge(docs);
-        expect(console.log.argsForCall[0][0]).toContain('non-existing-link');
-      });
+    it('should log warning when a linked page does not exist', function() {
+      docs.push(new Doc({section: 'api', id: 'with-broken.link', links: ['non-existing-link']}))
+      ngdoc.checkBrokenLinks(docs);
+      expect(console.log).toHaveBeenCalled();
+      var warningMsg = console.log.argsForCall[0][0]
+      expect(warningMsg).toContain('WARNING:');
+      expect(warningMsg).toContain('non-existing-link');
+      expect(warningMsg).toContain('api/with-broken.link');
+    });
 
-      it('should say where is the non-existing link', function() {
-        ngdoc.merge(docs);
-        expect(console.log.argsForCall[0][0]).toContain('api/fake.id1');
-      });
+    it('should log warning when a linked anchor does not exist', function() {
+      docs.push(new Doc({section: 'api', id: 'with-broken.link', links: ['api/fake.id1#non-existing']}))
+      ngdoc.checkBrokenLinks(docs);
+      expect(console.log).toHaveBeenCalled();
+      var warningMsg = console.log.argsForCall[0][0]
+      expect(warningMsg).toContain('WARNING:');
+      expect(warningMsg).toContain('non-existing');
+      expect(warningMsg).toContain('api/with-broken.link');
     });
   });
 
@@ -524,7 +528,7 @@ describe('ngdoc', function() {
         doc.ngdoc = 'filter';
         doc.parse();
         expect(doc.html()).toContain(
-            '<h3 id="Animations">Animations</h3>\n' +
+            '<h3 id="usage_animations">Animations</h3>\n' +
             '<div class="animations">' +
               '<ul>' +
                 '<li>enter - Add text</li>' +
@@ -541,7 +545,7 @@ describe('ngdoc', function() {
         var doc = new Doc('@ngdoc overview\n@name angular\n@description\n#heading\ntext');
         doc.parse();
         expect(doc.html()).toContain('text');
-        expect(doc.html()).toContain('<h2>heading</h2>');
+        expect(doc.html()).toContain('<h2 id="heading">heading</h2>');
         expect(doc.html()).not.toContain('Description');
       });
     });
diff --git a/docs/src/dom.js b/docs/src/dom.js
index 897a1831..e696faf4 100644
--- a/docs/src/dom.js
+++ b/docs/src/dom.js
@@ -4,6 +4,7 @@
 
 exports.DOM = DOM;
 exports.htmlEscape = htmlEscape;
+exports.normalizeHeaderToId = normalizeHeaderToId;
 
 //////////////////////////////////////////////////////////
 
@@ -16,10 +17,36 @@ function htmlEscape(text){
           .replace(/\}\}/g, '<span>}}</span>');
 }
 
+function nonEmpty(header) {
+  return !!header;
+}
+
+function idFromCurrentHeaders(headers) {
+  if (headers.length === 1) return headers[0];
+  // Do not include the first level title, as that's the title of the page.
+  return headers.slice(1).filter(nonEmpty).join('_');
+}
+
+function normalizeHeaderToId(header) {
+  if (typeof header !== 'string') {
+    return '';
+  }
+
+  return header.toLowerCase()
+      .replace(/<.*>/g, '')         // html tags
+      .replace(/[\!\?\:\.\']/g, '') // special characters
+      .replace(/&#\d\d;/g, '')      // html entities
+      .replace(/\(.*\)/mg, '')      // stuff in parenthesis
+      .replace(/\s$/, '')           // trailing spaces
+      .replace(/\s+/g, '-');        // replace whitespaces with dashes
+}
+
 
 function DOM() {
   this.out = [];
   this.headingDepth = 0;
+  this.currentHeaders = [];
+  this.anchors = [];
 }
 
 var INLINE_TAGS = {
@@ -44,17 +71,28 @@ DOM.prototype = {
   },
 
   html: function(html) {
-    if (html) {
-      var headingDepth = this.headingDepth;
-      for ( var i = 10; i > 0; --i) {
-        html = html
-          .replace(new RegExp('<h' + i + '(.*?)>([\\s\\S]+)<\/h' + i +'>', 'gm'), function(_, attrs, content){
-            var tag = 'h' + (i + headingDepth);
-            return '<' + tag + attrs + '>' + content + '</' + tag + '>';
-          });
-      }
-      this.out.push(html);
-    }
+    if (!html) return;
+
+    var self = this;
+    // rewrite header levels, add ids and collect the ids
+    html = html.replace(/<h(\d)(.*?)>([\s\S]+?)<\/h\1>/gm, function(_, level, attrs, content) {
+      level = parseInt(level, 10) + self.headingDepth; // change header level based on the context
+
+      self.currentHeaders[level - 1] = normalizeHeaderToId(content);
+      self.currentHeaders.length = level;
+
+      var id = idFromCurrentHeaders(self.currentHeaders);
+      self.anchors.push(id);
+      return '<h' + level + attrs + ' id="' + id + '">' + content + '</h' + level + '>';
+    });
+
+    // collect anchors
+    html = html.replace(/<a name="(\w*)">/g, function(match, anchor) {
+      self.anchors.push(anchor);
+      return match;
+    });
+
+    this.out.push(html);
   },
 
   tag: function(name, attr, text) {
@@ -85,17 +123,18 @@ DOM.prototype = {
 
   h: function(heading, content, fn){
     if (content==undefined || (content instanceof Array && content.length == 0)) return;
+
     this.headingDepth++;
+    this.currentHeaders[this.headingDepth - 1] = normalizeHeaderToId(heading);
+    this.currentHeaders.length = this.headingDepth;
+
     var className = null,
         anchor = null;
     if (typeof heading == 'string') {
-      var id = heading.
-          replace(/\(.*\)/mg, '').
-          replace(/[^\d\w\$]/mg, '.').
-          replace(/-+/gm, '-').
-          replace(/-*$/gm, '');
+      var id = idFromCurrentHeaders(this.currentHeaders);
+      this.anchors.push(id);
       anchor = {'id': id};
-      var classNameValue = id.toLowerCase().replace(/[._]/mg, '-');
+      var classNameValue = this.currentHeaders[this.headingDepth - 1]
       if(classNameValue == 'hide') classNameValue = '';
       className = {'class': classNameValue};
     }
diff --git a/docs/src/gen-docs.js b/docs/src/gen-docs.js
index 61fd3b3a..ffbf0068 100755
--- a/docs/src/gen-docs.js
+++ b/docs/src/gen-docs.js
@@ -55,6 +55,8 @@ writer.makeDir('build/docs/', true).then(function() {
     fileFutures.push(writer.output('partials/' + doc.section + '/' + id + '.html', doc.html()));
   });
 
+  ngdoc.checkBrokenLinks(docs);
+
   writeTheRest(fileFutures);
 
   return Q.deep(fileFutures);
diff --git a/docs/src/ngdoc.js b/docs/src/ngdoc.js
index 0f94f6eb..24d1aa26 100644
--- a/docs/src/ngdoc.js
+++ b/docs/src/ngdoc.js
@@ -36,6 +36,7 @@ exports.trim = trim;
 exports.metadata = metadata;
 exports.scenarios = scenarios;
 exports.merge = merge;
+exports.checkBrokenLinks = checkBrokenLinks;
 exports.Doc = Doc;
 
 exports.ngVersions = function() {
@@ -169,6 +170,7 @@ function Doc(text, file, line) {
   this.methods = this.methods || [];
   this.events = this.events || [];
   this.links = this.links || [];
+  this.anchors = this.anchors || [];
 }
 Doc.METADATA_IGNORE = (function() {
   var words = fs.readFileSync(__dirname + '/ignore.words', 'utf8');
@@ -242,6 +244,14 @@ Doc.prototype = {
    * @returns {string} Absolute url
    */
   convertUrlToAbsolute: function(url) {
+    var hashIdx = url.indexOf('#');
+
+    // Lowercase hash parts of the links,
+    // so that we can keep correct API names even when the urls are lowercased.
+    if (hashIdx !== -1) {
+      url = url.substr(0, hashIdx) + url.substr(hashIdx).toLowerCase();
+    }
+
     if (url.substr(-1) == '/') return url + 'index';
     if (url.match(/\//)) return url;
     return this.section + '/' + url;
@@ -569,6 +579,8 @@ Doc.prototype = {
       dom.h('Example', self.example, dom.html);
     });
 
+    self.anchors = dom.anchors;
+
     return dom.toString();
 
     //////////////////////////
@@ -606,7 +618,7 @@ Doc.prototype = {
       dom.html('<a href="api/ngAnimate.$animate">Click here</a> to learn more about the steps involved in the animation.');
     }
     if(params.length > 0) {
-      dom.html('<h2 id="parameters">Parameters</h2>');
+      dom.html('<h2>Parameters</h2>');
       dom.html('<table class="variables-matrix table table-bordered table-striped">');
       dom.html('<thead>');
       dom.html('<tr>');
@@ -660,7 +672,7 @@ Doc.prototype = {
   html_usage_returns: function(dom) {
     var self = this;
     if (self.returns) {
-      dom.html('<h2 id="returns">Returns</h2>');
+      dom.html('<h2>Returns</h2>');
       dom.html('<table class="variables-matrix">');
       dom.html('<tr>');
       dom.html('<td>');
@@ -1211,22 +1223,7 @@ function merge(docs){
   });
 
   for(var i = 0; i < docs.length;) {
-    var doc = docs[i];
-
-    // check links - do they exist ?
-    doc.links.forEach(function(link) {
-      // convert #id to path#id
-      if (link[0] == '#') {
-        link = doc.section + '/' + doc.id.split('#').shift() + link;
-      }
-      link = link.split('#').shift();
-      if (!byFullId[link]) {
-        console.log('WARNING: In ' + doc.section + '/' + doc.id + ', non existing link: "' + link + '"');
-      }
-    });
-
-    // merge into parents
-    if (findParent(doc, 'method') || findParent(doc, 'property') || findParent(doc, 'event')) {
+    if (findParent(docs[i], 'method') || findParent(docs[i], 'property') || findParent(docs[i], 'event')) {
       docs.splice(i, 1);
     } else {
       i++;
@@ -1255,6 +1252,36 @@ function merge(docs){
 }
 //////////////////////////////////////////////////////////
 
+
+function checkBrokenLinks(docs) {
+  var byFullId = Object.create(null);
+
+  docs.forEach(function(doc) {
+    byFullId[doc.section + '/' + doc.id] = doc;
+  });
+
+  docs.forEach(function(doc) {
+    doc.links.forEach(function(link) {
+      // convert #id to path#id
+      if (link[0] == '#') {
+        link = doc.section + '/' + doc.id.split('#').shift() + link;
+      }
+
+      var parts = link.split('#');
+      var pageLink = parts[0];
+      var anchorLink = parts[1];
+      var linkedPage = byFullId[pageLink];
+
+      if (!linkedPage) {
+        console.log('WARNING: ' + doc.section + '/' + doc.id + ' (defined in ' + doc.file + ') points to a non existing page "' + link + '"!');
+      } else if (anchorLink && linkedPage.anchors.indexOf(anchorLink) === -1) {
+        console.log('WARNING: ' + doc.section + '/' + doc.id + ' (defined in ' + doc.file + ') points to a non existing anchor "' + link + '"!');
+      }
+    });
+  });
+}
+
+
 function property(name) {
   return function(value){
     return value[name];