Make HasAutocompleteAttributeValue handle multi-valued strings

components/autofill/core/common/autofill_util.cc

 // Copyright 2015 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "components/autofill/core/common/autofill_util.h"
 
 #include <algorithm>
 #include <vector>
 
 #include "base/command_line.h"
@@ skipping 114 matching lines @@
   } else {
     if (isCheckable) {
       form_field_data->check_status =
           FormFieldData::CheckStatus::CHECKABLE_BUT_UNCHECKED;
     } else {
       form_field_data->check_status = FormFieldData::CheckStatus::NOT_CHECKABLE;
     }
   }
 }
 
+std::vector<std::string> LowercaseAndTokenizeAttributeString(
+    const std::string& attribute) {
+  return base::SplitString(base::ToLowerASCII(attribute),
+                           base::kWhitespaceASCII, base::KEEP_WHITESPACE,

[vabr (Chromium), 2016/10/13 12:49:31]

I'm curious about the KEEP_WHITESPACE:
Because whitespaces are token separators, they cannot (by definition) be part of
tokens. So KEEP_WHITESPACE leads to the same result as TRIM_WHITESPACE here.
However, TRIM_WHITESPACE would lead to more work inside SplitString.

If you use KEEP_WHITESPACE on purpose, to save work in SplitString (or for any
other reason), please mention the reason in a comment. It may happen that in the
future that reason might end up being valid, and will make it easier to change
this code to reflect that.

If you did not use KEEP_WHITESPACE on purpose, then I recommend changing it to
TRIM_WHITESPACE, which describes better the result (it will contain no
whitespace). Saving work in SplitString is a nice thought, but likely not
noticeable in its effect here, while the benefit of the code creating the right
expectations is rather important.

[jdoerrie, 2016/10/13 13:45:07]

I agree, changed it to TRIM_WHITESPACE.

+                           base::SPLIT_WANT_NONEMPTY);
+}
+
 }  // namespace autofill