Make HTTP header parsing case-insensitive.

[cubemap] / parse.h

diff --git a/parse.h b/parse.h
index de260640f56ebc4a0cbe0f6b06a39ead8d6bcbc0..a2e35802d38d31e804e4b16dcf1f948df6f7cfa7 100644 (file)
--- a/parse.h
+++ b/parse.h
@@ -5,14 +5,52 @@
 
 #include <stddef.h>
 #include <string>
+#include <algorithm>
+#include <string>
+#include <unordered_map>
 #include <vector>
 
+// Locale-unaware tolower(); matches RFC 2616 no matter what the locale is set to.
+static inline char ascii_tolower(const char ch)
+{
+       if (ch >= 'A' && ch <= 'Z') {
+               return ch + 'a' - 'A';
+       } else {
+               return ch;
+       }
+}
+
+// Case-insensitive header comparison and hashing.
+struct HTTPLess {
+       bool operator() (const std::string &a, const std::string &b) const
+       {
+               return std::lexicographical_compare(
+                       begin(a), end(a), begin(b), end(b),
+                       [](char a, char b) {
+                               return ascii_tolower(a) < ascii_tolower(b);
+                       });
+       }
+};
+struct HTTPHash {
+       size_t operator() (const std::string &s) const
+       {
+               std::string s_low = s;
+               for (char &ch : s_low) { ch = ascii_tolower(ch); }
+               return std::hash<std::string>() (s_low);
+       }
+};
+using HTTPHeaderMultimap = std::unordered_multimap<std::string, std::string, HTTPHash, HTTPLess>;
+
 // Split a line on whitespace, e.g. "foo  bar baz" -> {"foo", "bar", "baz"}.
 std::vector<std::string> split_tokens(const std::string &line);
 
 // Split a string on \n or \r\n, e.g. "foo\nbar\r\n\nbaz\r\n\r\n" -> {"foo", "bar", "baz"}.
 std::vector<std::string> split_lines(const std::string &str);
 
+// Extract HTTP headers from a request or response. Ignores the first line,
+// where the verb or the return code is.
+HTTPHeaderMultimap extract_headers(const std::vector<std::string> &lines, const std::string &log_context);
+
 // Add the new data to an existing string, looking for \r\n\r\n
 // (typical of HTTP requests and/or responses). Will return one
 // of the given statuses.