[24.2] Fix timeout handling for planemo / galaxy-tool-util

[mvdbeek]

Planemo calls verify_tool with a maxseconds value, and verify_tool sets tool_test_dict.setdefault("maxseconds", maxseconds), which fails, since we are serializing and already setting the default maxseconds while parsing the test dict.

Fixes the IUC tool test timeouts not working when testing against 24.2. Broken in
9695ccc#diff-06835be1d1943ebce8373e2cd1f8e86c7b7813b720398850c978b734dd29e2daR1390 (maxseconds was previously not serialized at all in ToolTestDescription.to_dict)

How to test the changes?

(Select all options that apply)

• I've included appropriate automated tests.
• This is a refactoring of components with existing test coverage.
• Instructions for manual testing are as follows:
  1. [add testing steps and prerequisites here if you didn't write automated tests covering all your changes]

License

• I agree to license these and all my past contributions to the core galaxy codebase under the MIT license.

[mvdbeek]

Hmm, this does fix it for my local planemo test setup

diff --git a/tests/test_cmd_test.py b/tests/test_cmd_test.py
index 7d0475f33f..d034f34b44 100644
--- a/tests/test_cmd_test.py
+++ b/tests/test_cmd_test.py
@@ -114,7 +114,7 @@ class CmdTestTestCase(CliTestCase):
         """Test if --test_timeout parameter is working."""
         with self._isolate(), NamedTemporaryFile(prefix="timeout_test_json") as json_out:
             test_artifact = os.path.join(TEST_TOOLS_DIR, "timeout.xml")
-            test_command = self._test_command("--test_output_json", json_out.name)
+            test_command = self._test_command("--test_output_json", json_out.name, "--galaxy_root", "/Users/mvandenb/src/galaxy")
             test_command = self.append_profile_argument_if_needed(test_command)
             test_command += [
                 "--test_timeout",

but it seems to not have worked for https://github.com/galaxyproject/tools-iuc/actions/runs/12668894193/job/35305453205 where a single test has been running for more than an hour already.

[mvdbeek]

Looks like it did eventually time out. Maybe we need to combine this with a limits section in the managed instances' job config and/or kill the job in the client