Performance issue with Tomcat

[lujiajing1126]

As is documented in the comments of MimeHeaders, each call of nextElement() has a complexity of O(n) with a comparison done with all previous elements to ensure the uniqueness.

Jetty has better performance since it uses HashSet to get a set of header keys.

I wonder if we could treat org.apache.catalina.connector.Request as a special case. A simple index-based iteration does improve the performance.

[pavolloffay]

Could you please point us to the code in this repo that uses this API? Is it in the extractor?

[lujiajing1126]

Could you please point us to the code in this repo that uses this API? Is it in the extractor?

In the TextMap implementation,

java-web-servlet-filter/opentracing-web-servlet-filter/src/main/java/io/opentracing/contrib/web/servlet/filter/HttpServletRequestExtractAdapter.java

Lines 55 to 57 in 558145f

	Enumeration<String> headerNamesIt = httpServletRequest.getHeaderNames();
	while (headerNamesIt.hasMoreElements()) {
	String headerName = headerNamesIt.nextElement();