Properly parse meta tag when fields are reversed

Via: node-fetch/node-fetch#682

Author: isaacs

lib/body.js

@@ -293,6 +293,12 @@ const convertBody = (buffer, headers) => {
   if (!res && str) {
     res = /<meta[\s]+?http-equiv=(['"])content-type\1[\s]+?content=(['"])(.+?)\2/i.exec(str)
 
+    if (!res) {
+      res = /<meta[\s]+?content=(['"])(.+?)\1[\s]+?http-equiv=(['"])content-type\3/i.exec(str)
+      if (res)
+        res.pop() // drop last quote
+    }
+
     if (res)
       res = /charset=(.*)/i.exec(res.pop())
   }

test/body.js

@@ -307,6 +307,13 @@ t.test('convert body', t => {
     t.equal(await b.textConverted(), s)
   })
 
+  t.test('html4 meta tag reversed', async t => {
+    const s = '<meta content="text/html; charset=gbk" http-equiv="Content-Type"><div>中文L</div>'
+    const b = new Body(convert(s, 'gbk'))
+    t.equal(await b.textConverted(), s)
+  })
+
+
   t.test('html5 meta tag', async t => {
     const s = '<meta charset="gbk"><div>中文</div>'
     const b = new Body(convert(s, 'gbk'))

test/fixtures/server.js

@@ -164,6 +164,12 @@ class TestServer {
       res.end(convert('<meta http-equiv="Content-Type" content="text/html; charset=gb2312"><div>中文</div>', 'gb2312'))
     }
 
+    if (p === '/encoding/gb2312-reverse') {
+      res.statusCode = 200;
+      res.setHeader('Content-Type', 'text/html');
+      res.end(convert('<meta content="text/html; charset=gb2312" http-equiv="Content-Type"><div>中文</div>', 'gb2312'));
+    }
+
     if (p === '/encoding/shift-jis') {
       res.statusCode = 200
       res.setHeader('Content-Type', 'text/html; charset=Shift-JIS')

test/index.js

@@ -1767,6 +1767,14 @@ t.test('with optional `encoding`', t => {
       })
     }))
 
+  t.test('encoding decode, html4 detect reverse http-equiv', t =>
+    fetch(`${base}encoding/gb2312-reverse`).then(res => {
+      t.equal(res.status, 200)
+      return res.textConverted().then(result => {
+        t.equal(result, '<meta content="text/html; charset=gb2312" http-equiv="Content-Type"><div>中文</div>')
+      })
+    }))
+
   t.test('default to utf8 encoding', t =>
     fetch(`${base}encoding/utf8`).then(res => {
       t.equal(res.status, 200)