在 JAVA HttpUrlConnection 中，请求 Header 设置的主要逻辑代码如下：
public synchronized void set(String k, String v) {
for (int i = nkeys; --i >= 0;)
if (k.equalsIgnoreCase(keys[i])) {
values[i] = v;
return;
}
add(k, v);
}
经验证，key应该是唯一的，key与value必须保持一一对应的映射关系。
相反，在Response模块的HeaderFields中，结构被定义为Entry>。即键与值不保持一对一的映射关系。
为什么是这样？ HTTP协议有相关约定吗？
添加：
在HttpClient4中，请求Header设置的主要逻辑代码如下：
/**
* Replaces the first occurence of the header with the same name. If no header with
* the same name is found the given header is added to the end of the list.
*
* @param header the new header that should replace the first header with the same
* name if present in the list.
*/
public void updateHeader(final Header header) {
if (header == null) {
return;
}
// HTTPCORE-361 : we don't use the for-each syntax, i.e.
// for (Header header : headers)
// as that creates an Iterator that needs to be garbage-collected
for (int i = 0; i < this.headers.size(); i++) {
final Header current = this.headers.get(i);
if (current.getName().equalsIgnoreCase(header.getName())) {
this.headers.set(i, header);
return;
}
}
this.headers.add(header);
}
响应标头
/**
* Gets all of the headers with the given name. The returned array
* maintains the relative order in which the headers were added.
*
* <p>Header name comparison is case insensitive.
*
* @param name the name of the header(s) to get
*
* @return an array of length >= 0
*/
public Header[] getHeaders(final String name) {
final List<Header> headersFound = new ArrayList<Header>();
// HTTPCORE-361 : we don't use the for-each syntax, i.e.
// for (Header header : headers)
// as that creates an Iterator that needs to be garbage-collected
for (int i = 0; i < this.headers.size(); i++) {
final Header header = this.headers.get(i);
if (header.getName().equalsIgnoreCase(name)) {
headersFound.add(header);
}
}
return headersFound.toArray(new Header[headersFound.size()]);
}
它们与 HttpUrlConnection 相同
