tests: conftest: run slowest tests actually

[blueyed]

Regression in 0025e44.

[blueyed]

(With regard to coverage we apparently do not need all the non-pexpect tests using "testdr", which are not explicitly marked as slow.)

[codecov]

Codecov Report

Merging #5331 into master will increase coverage by 6.63%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #5331      +/-   ##
==========================================
+ Coverage   60.38%   67.01%   +6.63%     
==========================================
  Files         115       59      -56     
  Lines       26264    12691   -13573     
  Branches     2589     2128     -461     
==========================================
- Hits        15859     8505    -7354     
+ Misses       9821     3506    -6315     
- Partials      584      680      +96

Impacted Files	Coverage Δ
testing/conftest.py	91.3% <100%> (-8.7%)	⬇️
src/_pytest/recwarn.py	46.77% <0%> (-51.62%)	⬇️
src/_pytest/python_api.py	55.04% <0%> (-42.44%)	⬇️
src/_pytest/junitxml.py	26.62% <0%> (-41.13%)	⬇️
src/_pytest/_argcomplete.py	35.13% <0%> (-40.55%)	⬇️
src/_pytest/monkeypatch.py	49.04% <0%> (-40.13%)	⬇️
src/_pytest/_code/_py2traceback.py	40% <0%> (-37.78%)	⬇️
src/_pytest/config/findpaths.py	54% <0%> (-32%)	⬇️
src/_pytest/assertion/truncate.py	64.81% <0%> (-27.78%)	⬇️
src/_pytest/pathlib.py	63.73% <0%> (-26.93%)	⬇️
... and 91 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c35544a...61b76c7. Read the comment docs.

[blueyed]

re coverage: apparently it was noticed when merging the regression.