I'll try to explain my use-case.
I have a system that acts remotely. Jobs are requested from a GUI and
the logs produced by the jobs have to be streamed back to the GUI via
websocket after being treated by a service of mine developed as a RestXQ
in BaseX10.
In the current solution all the GUIs connect to the same websocket URL.
Sending something to that websocket makes everything available to
everyone and it's up to the javascript code to filter out what is of
interest for the user.
Instead it would be nice to have isolated channels end-to-end for
instance by using the id of the logged user.
Does this make enough sense?
Danke schön Christian.
M.
On 23/06/23 16:46, Christian Grün wrote:
Hi Marco,
As most web applications open only one or a few websockets at the same
time, we decided against supporting wildcards.
Feel free to share background information on your use case with us,
maybe you can convince us to get active ;·)
Ciao,
Christian
On Fri, Jun 23, 2023 at 11:21 AM Marco Lettere <m.lett...@gmail.com> wrote:
Dear all,
so I just found out that it is not possible to create websockets bound
to dynamic paths.
If I try a websocket connection to
declare
%ws:connect('/notification/subject/{$subject}')
function log:subject() {};
I get an error (without significant logs). Whereas
declare
%ws:connect('/notification/subject/one')
function log:subject() {};
works well.
Does this have a motivation? Is there another way of meeting the
requirement of having dynamic websocket channels?
Thanks in advance,
Marco.