How to report the cucumber tests after a rerun? - ruby

Question

I have a suite of tests that are launched in continuous integration using parallel test. After the deploys the environment is a little unstable, so when the suite is finished, the failed tests are relaunched (by this moment the environment is stable again).
Basically, there is only one jenkins build, and it will execute this:
task :jenkins1_with_rerun do
selenium_successful = system "bundle exec parallel_cucumber features/web/ -o \"-p jenkins1\" -n 3"
p 'start rerun'
rerun_successful = run_rake_task("features:jenkins_rerun")
unless selenium_successful || rerun_successful
raise 'Cucumber tests failed'
end
end
The first execution launches the suite with parallel tests. Here is the cucumber profile:
jenkins1:
<%= std_opts %>
HEADLESS=true
--tags #part1
--tags ~#wip
--format junit
--out log/junit
--format ParallelTests::Cucumber::FailuresLogger
--out parallel_cucumber_failures.log
After this is finished. The second execution starts, launching the failed tests recorded in a file. The cucumber profile for this execution is this:
jenkins_rerun:
ENVIRONMENT=master
SCREENSHOT=true
HEADLESS=true
--format pretty
--no-source
--format junit
--out log/junit_rerun
#parallel_cucumber_failures.log
Now. Everything works. The only problem is that I use the junit report in jenkins to make graphs of failed and successful tests. I may have reports like this:
-------------------Tests---KO---OK
First execution---75----10----65
Rerun-------------10-------0----10
This is 100% green, because all the problems were caused by instability after the deploy. So I want the junit report in jenkins to said that 75 tests have been launched, 75 OK, 0 KO, or something equivalent.
Right now, the junit report of the first execution says that of 75 tests, I have 10 KOs; and the second junit says that of 10 tests, there are 0 KOs.
What may be a good solution to this? It possible to mix the results of both Junit reports?
I will also accept to be able to display in jenkins both junit reports, each one with a graph. But I think that jenkins only allows to show one junit report graph.

Answer 1

I finally resolved this problem.
The solution was to create a "parser" that searches (with RegEx) through the Junit files generated by the rerun, and then goes to the Jnuit files of the original/first run and rewrites the results of those tests that have passed during the rerun.
In the end, if everything is ok in the rerun, the junit graph is 100% green and there are not alerts of failures.