Fix "conversion of size_t to unsigned int" warning

net/dns/record_rdata.cc

Index: net/dns/record_rdata.cc
diff --git a/net/dns/record_rdata.cc b/net/dns/record_rdata.cc
index ff9434b489e11a958f44b7d485a00279d28838d8..1180d46973d0ed2dbd5c2601225f743ec33b85ab 100644
--- a/net/dns/record_rdata.cc
+++ b/net/dns/record_rdata.cc
@@ -259,6 +259,11 @@ uint16_t NsecRecordRdata::Type() const {
   return NsecRecordRdata::kType;
 }
 
+uint16_t NsecRecordRdata::bitmap_length() const {
+  DCHECK_LE(bitmap_.size(), 32u);

[Ryan Sleevi, 2016/07/06 15:28:58]

It's unclear if you added this DCHECK as documentary purposes (this is
guaranteed) or as a guard against developer mistake ("size should be less than
32")

If the latter, it's a bad guard - because this is handling untrusted data,
there's no way the caller could validate/preprocess the data to make sure that
this was true.

If the former, it's not sufficient documentation - because I still had to try to
figure out which it was, and why it was safe to document this. (For the record,
it's because of line 241)

The cost of the presumed documentation is that it forces you to de-inline this,
and I'm not sure that's worth it.

+  return static_cast<uint16_t>(bitmap_.size() * 8);
+}
+
 bool NsecRecordRdata::IsEqual(const RecordRdata* other) const {
   if (other->Type() != Type())
     return false;