Construct a set of tokens based upon the given delimiter, and an indication of whether said delimiter should be considered part of the left side (effectively the name).
Clone a source set of HttpTokens
Add a token with the given name. The content is provided via the specified delegate. We stuff this name & content into the output buffer, and map a new Token onto the appropriate buffer slice.
Add a simple name/value pair to the output
Add a name/date(long) pair to the output
Add a name/integer pair to the output
Create a filter for iterating over the tokens matching a particular name.
Return a char[] representing the output. An empty array is returned if output was not configured. This perhaps could just return our 'output' buffer content, but that would not reflect deletes, or seperators. Better to do it like this instead, for a small cost.
Return the value of the provided header, or null if the header does not exist
Return the date value of the provided header, or the provided default-value if the header does not exist
Return the integer value of the provided header, or the provided default-vaule if the header does not exist
overridable method to handle the case where a token does not have a separator. Apparently, this can happen in HTTP usage
Have tokens been parsed yet?
Iterate over the set of tokens
Read all tokens. Everything is mapped rather than being allocated & copied
Parse an input string.
Output the token list to the provided consumer
remove a token from our list. Returns false if the named token is not found.
Reset this set of tokens.
Indicate whether tokens have been parsed or not.
Maintains a set of HTTP tokens. These tokens include headers, query- parameters, and anything else vaguely related. Both input and output are supported, though a subclass may choose to expose as read-only.
All tokens are mapped directly onto a buffer, so there is no memory allocation or copying involved.
Note that this class does not support deleting tokens, per se. Instead it marks tokens as being 'unused' by setting content to null, avoiding unwarranted reshaping of the token stack. The token stack is reused as time goes on, so there's only minor runtime overhead.