Thanks. > The server is MarkLogic, s maybe they are biased against BaseX ;) I'm not > sure about the configuration, but replacing the User-Agent header with > "curl/7.71.1" didn't help.
Fraudulent competition ;·) I’m particularly confused by the response code: 502 usually indicates a server-side configuration error, which should not be triggerable from a client. I noticed it would be more correct if BaseX had not sent a second request if the response code is unequal to 401. In our case, it might have proven to be helpful, as we can see that a) the response to the second request is 401, and b) it contains the Digest WWW-Authenticate header. In addition, the second request contains a Content-Length header, which is missing in the first request. Presumably, due to the missing Content-Length:0 header, the server tries to retrieve a body, but gives up after the timeout and returns 502. With the latest snapshot [1], I’m now creating an explicit 0-bytes-body for all methods except for GET and TRACE. Does this help? > I noticed that in both curl and Postman the first request says, "[HTTP > request 1/2]," whereas BaseX says, "[HTTP request 1/1]" My hope is that numerous of these irregularities will disappear once we’ll have switched to the new JDK 11 HTTP Client [2] with BaseX 10. [1] https://files.basex.org/releases/latest/ [2] https://openjdk.java.net/groups/net/httpclient/intro.html
