Use the exact-width integer types defined in <stdint.h> rather than

net/base/ip_pattern.h

Index: net/base/ip_pattern.h
diff --git a/net/base/ip_pattern.h b/net/base/ip_pattern.h
index bf1d45583d3582738b9c637898cbfa014b4c2546..d4619eafca965c171f820f978df929ada35db7af 100644
--- a/net/base/ip_pattern.h
+++ b/net/base/ip_pattern.h
@@ -41,7 +41,7 @@ class NET_EXPORT IPPattern {
   bool ParseComponentPattern(const base::StringPiece& text,
                              ComponentPattern* pattern) const;
   // Convert IP component to an int.  Assume hex vs decimal for IPV6 vs V4.
-  bool ValueTextToInt(const base::StringPiece& input, uint32* output) const;
+  bool ValueTextToInt(const base::StringPiece& input, uint32_t* output) const;
 
   bool is_ipv4_;
   // The |ip_mask_| indicates, for each component, if this pattern requires an
@@ -55,7 +55,7 @@ class NET_EXPORT IPPattern {
   // The class invariant is:
   // ip_mask_.size() == component_patterns_.size()
   //                    + size(our ComponentPattern list)
-  std::vector<uint32> component_values_;
+  std::vector<uint32_t> component_values_;
   // If only one component position was specified using a range, then this
   // list will only have 1 element (i.e., we only have patterns for each element
   // of ip_mask_ that is false.)