Add mechanism to set a multiplier for test timeouts

[hololeap]

Is your enhancement request related to a problem? Please describe.

I have been handling packaging haskell-language-server for Gentoo Linux, and as such I try to run the test suite bundled with each plugin. While there have been some other issues with the test suites that I am working on solutions for, one problem that occasionally bites me is that it seems that my computer is too slow and will sometimes trigger a timeout failure:

    :info! reports type, constructors and unfiltered instances for multiple types:                    FAIL
      Exception: Timed out waiting to receive a message from the server.
      Last message received:
      {
          "jsonrpc": "2.0",
          "method": "textDocument/publishDiagnostics",
          "params": {
              "diagnostics": [],
              "uri": "file:///tmp/extra-dir-1657892564956/TInfo.hs"
          }
      }
      Use -p '/:info! reports type, constructors and unfiltered instances for multiple types/' to rerun this test only.

Note that this test passed on a previous run, probably because the system load was lower.

(Full test log)

Describe the solution you'd like

I was told that Meson has a solution for problems like this; namely they added a timeout multiplier for tests that you can set on the command line:

  -t TIMEOUT_MULTIPLIER, --timeout-multiplier TIMEOUT_MULTIPLIER
                                        Define a multiplier for test timeout,
                                        for example when running tests in
                                        particular conditions they might take
                                        more time to execute. (<= 0 to disable
                                        timeout)

(Source)

I think that something like this should be implemented for HLS so that those with slower computers can run the test suites without running into these false positives. It could be implemented so that the multiplier is passed via an environment variable or as a flag passed via Cabal's --test-option flag.

[fendor]

The test is flaky and would timeout even if you waited for 10 hours. We need to fix that!
Tracking ticket #3736.

Closing, as the proposal will not solve the issue.