Revert "Simplify getting pdfjs main page"

This reverts commit 9c731bde85627308fdde4730b0181a014096cb47.

We need to set some PDF.js options, so we can't just use the default viewer
with ?file=...

qutebrowser/browser/pdfjs.py

@@ -42,6 +42,66 @@ class PDFJSNotFound(Exception):
         super().__init__(message)
 
 
+def generate_pdfjs_page(url):
+    """Return the html content of a page that displays url with pdfjs.
+
+    Returns a string.
+
+    Args:
+        url: The url of the pdf as QUrl.
+    """
+    if not is_available():
+        return jinja.render('no_pdfjs.html',
+                            url=url.toDisplayString(),
+                            title="PDF.js not found")
+    viewer = get_pdfjs_res('web/viewer.html').decode('utf-8')
+    script = _generate_pdfjs_script(url)
+    html_page = viewer.replace('</body>',
+                               '</body><script>{}</script>'.format(script))
+    return html_page
+
+
+def _generate_pdfjs_script(url):
+    """Generate the script that shows the pdf with pdf.js.
+
+    Args:
+        url: The url of the pdf page as QUrl.
+    """
+    return (
+        'document.addEventListener("DOMContentLoaded", function() {{\n'
+        '  PDFJS.verbosity = PDFJS.VERBOSITY_LEVELS.info;\n'
+        '  (window.PDFView || window.PDFViewerApplication).open("{url}");\n'
+        '}});\n'
+    ).format(url=javascript.string_escape(url.toString(QUrl.FullyEncoded)))
+
+
+def fix_urls(asset):
+    """Take an html page and replace each relative URL with an absolute.
+
+    This is specialized for pdf.js files and not a general purpose function.
+
+    Args:
+        asset: js file or html page as string.
+    """
+    new_urls = [
+        ('viewer.css', 'qute://pdfjs/web/viewer.css'),
+        ('compatibility.js', 'qute://pdfjs/web/compatibility.js'),
+        ('locale/locale.properties',
+         'qute://pdfjs/web/locale/locale.properties'),
+        ('l10n.js', 'qute://pdfjs/web/l10n.js'),
+        ('../build/pdf.js', 'qute://pdfjs/build/pdf.js'),
+        ('debugger.js', 'qute://pdfjs/web/debugger.js'),
+        ('viewer.js', 'qute://pdfjs/web/viewer.js'),
+        ('compressed.tracemonkey-pldi-09.pdf', ''),
+        ('./images/', 'qute://pdfjs/web/images/'),
+        ('../build/pdf.worker.js', 'qute://pdfjs/build/pdf.worker.js'),
+        ('../web/cmaps/', 'qute://pdfjs/web/cmaps/'),
+    ]
+    for original, new in new_urls:
+        asset = asset.replace(original, new)
+    return asset
+
+
 SYSTEM_PDFJS_PATHS = [
     # Debian pdf.js-common
     # Arch Linux pdfjs (AUR)
@@ -154,7 +214,7 @@ def should_use_pdfjs(mimetype):
 
 def get_main_url(filename):
     """Get the URL to be opened to view a local PDF."""
-    url = QUrl('qute://pdfjs/web/viewer.html')
+    url = QUrl('qute://pdfjs/')
 
     file_url = QUrl('qute://pdfjs/file')
     file_url_query = QUrlQuery()

qutebrowser/browser/qutescheme.py

@@ -533,6 +533,11 @@ def qute_pdfjs(url):
             mimetype = 'application/octet-stream'
         return mimetype, data
 
+    if url.path() == '/':
+        filepath = QUrlQuery(url).queryItemValue("file")
+        data = pdfjs.generate_pdfjs_page(QUrl.fromLocalFile(filepath))
+        return 'text/html', data
+
     try:
         data = pdfjs.get_pdfjs_res(url.path())
     except pdfjs.PDFJSNotFound as e:

tests/unit/browser/test_pdfjs.py

@@ -22,6 +22,45 @@ import pytest
 from qutebrowser.browser import pdfjs
 
 
+# Note that we got double protection, once because we use QUrl.FullyEncoded and
+# because we use qutebrowser.utils.javascript.string_escape.  Characters
+# like " are already replaced by QUrl.
+@pytest.mark.parametrize('url, expected', [
+    ('http://foo.bar', "http://foo.bar"),
+    ('http://"', ''),
+    ('\0', '%00'),
+    ('http://foobar/");alert("attack!");',
+     'http://foobar/%22);alert(%22attack!%22);'),
+])
+def test_generate_pdfjs_script(url, expected):
+    expected_open = 'open("{}");'.format(expected)
+    url = QUrl(url)
+    actual = pdfjs._generate_pdfjs_script(url)
+    assert expected_open in actual
+    assert 'PDFView' in actual
+
+
+def test_fix_urls():
+    page = textwrap.dedent("""
+        <html>
+        <script src="viewer.js"></script>
+        <link href="viewer.css">
+        <script src="unrelated.js"></script>
+        </html>
+    """).strip()
+
+    expected = textwrap.dedent("""
+        <html>
+        <script src="qute://pdfjs/web/viewer.js"></script>
+        <link href="qute://pdfjs/web/viewer.css">
+        <script src="unrelated.js"></script>
+        </html>
+    """).strip()
+
+    actual = pdfjs.fix_urls(page)
+    assert actual == expected
+
+
 @pytest.mark.parametrize('path, expected', [
     ('web/viewer.js', 'viewer.js'),
     ('build/locale/foo.bar', 'locale/foo.bar'),