Issue #1356 - Improve detection of the issuer of the request

src/shared/Utils.h

 #ifndef UTILS_H
 #define UTILS_H
 
 #include <algorithm>
 #include <locale>
 #include <functional>
 #include <string>
 
 #define WM_ALREADY_UP_TO_DATE WM_APP+1
 #define WM_UPDATE_CHECK_ERROR WM_APP+2
@@ skipping 32 matching lines @@
 T TrimString(T text)
 {
   // Via http://stackoverflow.com/questions/216823/whats-the-best-way-to-trim-stdstring
   T trimmed(text);
   std::function<bool(T::value_type)> isspace = std::bind(&std::isspace<T::value_type>, std::placeholders::_1, std::locale::classic());
   trimmed.erase(trimmed.begin(), std::find_if(trimmed.begin(), trimmed.end(), std::not1(isspace)));
   trimmed.erase(std::find_if(trimmed.rbegin(), trimmed.rend(), std::not1(isspace)).base(), trimmed.end());
   return trimmed;
 }
 
+template <class T>
+T ExtractHTTPHeader(const T& allHeaders, const T& targetHeaderName, const T& delimiter)

[Felix Dahlke, 2014/11/03 04:58:16]

1. I think we generally capitalise it like this: ExtractHttpHeader - but I
wouldn't dare say it's consistent :)
2. Since HTTP headers generally end with \n, I don't think we need to pass in
delimiter here.
3. I find it somewhat unintuitive that targetHeaderName needs to include the
colon here. IMO we should just pass the name here and deal with the colon in
this function.

[Oleksandr, 2014/11/03 11:37:37]

2. Http headers can be seperated by '\0' as well. See HTTP_QUERY_RAW_HEADERS
http://msdn.microsoft.com/en-us/library/windows/desktop/aa385351(v=vs.85).aspx.
I just want to keep this function as flexible as possible for future. Made that
argument optional.

[sergei, 2014/11/03 14:17:46]

If we really want to have it in utils then tests are required as well.
But I would say we are trying to generalize it too early and it leads only to
unnecessary logic and questions (btw, it violates such buzzwords KISS and "the
rule of three [refactoring]"). I would keep it in PluginWbPassThrough.cpp file
without default values and with requirement to include colon in target header
name.

[Oleksandr, 2014/11/04 12:02:42]

Yes, I can see how the changes to make this more generalized turned out
unnecessarily complex for reading. And indeed we are not very likely to use this
code anywhere else (no third use). So I would actually agree with Sergei to keep
it simple and revert this part to just a narrowly specialized function in
PluginWbPassThrough.cpp. 

[Felix Dahlke, 2014/11/04 16:13:25]

Fully agree, good points Sergei.

+{
+  auto targetHeaderBeginsAt = allHeaders.find(targetHeaderName);
+  if (targetHeaderBeginsAt == T::npos)
+  {
+    return T();
+  }
+  targetHeaderBeginsAt += targetHeaderName.length();
+  auto targetHeaderEndsAt = allHeaders.find(delimiter, targetHeaderBeginsAt);
+  if (targetHeaderEndsAt == T::npos)
+  {
+    return T();
+  }
+  return allHeaders.substr(targetHeaderBeginsAt, targetHeaderEndsAt - targetHeaderBeginsAt);
+}
+
 #endif // UTILS_H