x.vweb: add configuration options and handle incomplete requests #20868
Conversation
| // done frees the current incomplete request header buffer and reset the HTTP request. | ||
| // use this function when the connection is kept open, but the request is handled | ||
| @[manualfree] | ||
| pub fn (mut ir IncomingRequest) done() { |
There was a problem hiding this comment.
Why is this public? The structure itself is not, and it is used only for the incoming_requests field of RequestParams, which itself is not public.
There was a problem hiding this comment.
Then I think I forgot to use it, I'll check
|
|
|
It shouldn't be a problem to cache the header until it's completely read. |
No, but the problem is that I can't access the data to cache it. And can't read the data first, because the net library only takes a buffered reader has argument when parsing only the request headers. So I'll probably need to add another function as well |
This pr adds a
Configstruct tox.vwebto change the default behaviour of vweb.When vweb encounters an incomplete request it will now wait until a timeout occurs for new data instead of immediately closing the connection. The
Configstruct contains an option to limit the number of incomplete requests per connetion that are allowed before the connection is closed.The
Configstruct also adds the option to limit the maximum HTTP header size and the maximum body length for a request.