[IDEA] Be able to cancel a search that I started

[anibal2j]

Is your feature request related to a problem? Please describe.
I use search daily. I can be searching for anything, an accession number (which I always put in metadata) or a given string (yes, it's bad, but I usually use some other filters to narrow the depth of the search). Some times a search is taking forever and then I notice that it won't return anything. The GUI is pretty much toast now and any new searches will not work until the first one finishes.

Describe your use case
I need this feature so I can accomplish doing a search that I know will work, compared to the one that I started and it's killing the DB for no reason - since no results will come out of it (potentially after scanning the whole table)

Describe the solution you'd like
There should be a cancel search button that just kills the whole pipeline going to the DB.

Describe alternatives you've considered
My workaround is to kill the GUI completely and re-launch a new one. ugh!

[jonbartels]

I know this was discussed at some time on the old forums or the old Mirth Slack.

Some notes:

• the searches are long running HTTP requests waiting for data
• We could kill this client-side only - However this would still leave a potentially long running or slow query going server-side
• We could add a kill option that operates server-side - This would require some sort of coordination in sending the search so it can be killed and server side orchestration to kill it. I don't remember how to do this in Java, JDBC, mybatis etc but I'm sure there are modern tools to do it
• The slow searches can come from three places:
• a slow DB query, how can this be killed with a kill signal?
• slow marshalling of the XML returned to the client
• slow unmarshalling or rendering of the XML in the client