[ReadingList iOS] Do not store HTTP images when distilling HTTPS pages

ios/chrome/browser/reading_list/url_downloader.cc

Index: ios/chrome/browser/reading_list/url_downloader.cc
diff --git a/ios/chrome/browser/reading_list/url_downloader.cc b/ios/chrome/browser/reading_list/url_downloader.cc
index 93afe9b1f05c3ae1798f94bbfc4740e80127747f..7fb822716a75f62cbef81488a5093699ba60d0ff 100644
--- a/ios/chrome/browser/reading_list/url_downloader.cc
+++ b/ios/chrome/browser/reading_list/url_downloader.cc
@@ -248,10 +248,14 @@ std::string URLDownloader::SaveAndReplaceImagesInHTML(
       // Data URI, the data part of the image is empty, no need to store it.
       continue;
     }
-    base::FilePath local_image_path;
-    std::string local_image_name;
-    if (!SaveImage(url, images[i].url, images[i].data, &local_image_name)) {
-      return std::string();
+    std::string local_image_name = "";

[jif, 2017/01/20 10:45:26]

afaik you don't need to initialize local_image_name

[Olivier, 2017/01/20 10:54:44]

Done.

+    if (!(distilled_url_.SchemeIsCryptographic() &&

[jif, 2017/01/20 10:45:26]

Because it confused me a bit, can you rewrite it as:

 <new line>
// Only save images if it is not mixed content.
// Mixed content is HTTP images on HTTPS pages.
bool image_is_mixed_content = distilled_url_.SchemeIsCryptographic() &&
!images[i].url.SchemeIsCryptographic();
if (!image_is_mixed_content) {

[Olivier, 2017/01/20 10:54:44]

Done.

+          !images[i].url.SchemeIsCryptographic())) {
+      // Only save images if it is not mixed content.
+      // Mixed content is HTTP imaged on HTTPS pages.

[jif, 2017/01/20 10:45:26]

s/imaged/images/

I appreciate the reminder of what is "mixed content".

[Olivier, 2017/01/20 10:54:44]

Done.

+      if (!SaveImage(url, images[i].url, images[i].data, &local_image_name)) {
+        return std::string();

[jif, 2017/01/20 10:45:26]

Optional nit (because not introduced by your CL, and not related to this CL's
change):

Returning an empty updated html seems wrong; it sounds as though the distilled
page will be blank if it contains mixed content images.
Can you add an explanation, maybe in the method's comment in the header?

[Olivier, 2017/01/20 10:54:44]

Returning the html empty will result in a distillation failure, which is an
expected behavior.
Adding a comment in header.

+      }
     }
     std::string image_url = net::EscapeForHTML(images[i].url.spec());
     size_t image_url_size = image_url.size();