The CI runs for swh-graph might hang badly on test_api_client.py, here's the most recent exampe.
This specific issues is solved now, but it really shouldn't happen that test running hangs, see below for an analysis by @olasd on how to avoid that exception aren't forwarded, blocking test abortion.
The .jar file is never installed within the tox environment, so the graph backend process fixture never actually succeeds in launching the server. FWIW, when running tox on my system, the tests hang just the same.
root@thyssen:~# unzip -l /home/jenkins/workspace/DGRPH/tests/.tox/dist/swh.graph-0.1.0.post33.zip
Archive: /home/jenkins/workspace/DGRPH/tests/.tox/dist/swh.graph-0.1.0.post33.zip
Length Date Time Name
--------- ---------- ----- ----
0 2019-11-04 10:23 swh.graph-0.1.0.post33/
0 2019-11-04 10:23 swh.graph-0.1.0.post33/swh.graph.egg-info/
0 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/
140 2019-11-04 10:23 swh.graph-0.1.0.post33/MANIFEST.in
18 2019-11-04 10:23 swh.graph-0.1.0.post33/version.txt
35 2019-11-04 10:23 swh.graph-0.1.0.post33/requirements-swh.txt
286 2019-11-04 10:23 swh.graph-0.1.0.post33/README.md
38 2019-11-04 10:23 swh.graph-0.1.0.post33/setup.cfg
163 2019-11-04 10:23 swh.graph-0.1.0.post33/Makefile
1162 2019-11-04 10:23 swh.graph-0.1.0.post33/PKG-INFO
38 2019-11-04 10:23 swh.graph-0.1.0.post33/requirements.txt
2411 2019-11-04 10:23 swh.graph-0.1.0.post33/setup.py
709 2019-11-04 10:23 swh.graph-0.1.0.post33/swh.graph.egg-info/SOURCES.txt
1 2019-11-04 10:23 swh.graph-0.1.0.post33/swh.graph.egg-info/dependency_links.txt
89 2019-11-04 10:23 swh.graph-0.1.0.post33/swh.graph.egg-info/requires.txt
130 2019-11-04 10:23 swh.graph-0.1.0.post33/swh.graph.egg-info/entry_points.txt
1162 2019-11-04 10:23 swh.graph-0.1.0.post33/swh.graph.egg-info/PKG-INFO
4 2019-11-04 10:23 swh.graph-0.1.0.post33/swh.graph.egg-info/top_level.txt
0 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/
127 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/__init__.py
0 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/server/
0 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/tests/
7032 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/backend.py
0 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/__init__.py
5751 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/cli.py
8863 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/webgraph.py
13080 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/pid.py
1565 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/dot.py
3278 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/client.py
5356 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/graph.py
27 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/py.typed
0 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/server/__init__.py
5175 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/server/app.py
0 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/tests/__init__.py
1361 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/tests/conftest.py
4540 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/tests/test_api_client.py
1526 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/tests/test_cli.py
4329 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/tests/test_graph.py
7219 2019-11-04 10:23 swh.graph-0.1.0.post33/swh/graph/tests/test_pid.py
--------- -------
75615 39 filesI guess you at least need a fix to MANIFEST.in to make sure the jar file is properly included in the sdist that tox uses to generate its venvs. Looks like the test data isn't included either.
Why this results in a hang instead of a proper failure needs to be checked as well. I'm guessing the exception is never actually passed to the queue if the GraphServerProcess.start() fails, and so the fixture's queue.get() hangs. The path manipulations to try and find a .jar file and the dataset also seem quite brittle, and we should consider moving to importlib.resources.
AFAICT this is a more general problem, the Java backend can hang forever in case of unexpected situations (uncaught exceptions? I really don't know…), which will make it not respond to any incoming request with no visible output.
We should make this visible and debuggable.
Another simple way to reproduce is just removing the *.jar file and running pytest on test_api_client.py.
This is not even a Java exception, but chances are fixing that case will fix at least a significant part of the general problem, if not all.