The scenario painted in the rebase-exec exercise is a little counter-intuitive. The commits we're rebasing add a bunch of text files, but we never interact with them. One of the commits breaks the test script, which we then need to fix by basically undoing the changes to the test script.
Perhaps moving to a scenario where the test file actually tests a different script which the commits modify would be a little more intuitive.
I'm not sure how we could achieve this yet. My original suggestion was having a script called zero.sh which does some simple math and exits with the result, e.g. exit (2 - 1 -1), and have test.sh just:
The scenario painted in the rebase-exec exercise is a little counter-intuitive. The commits we're rebasing add a bunch of text files, but we never interact with them. One of the commits breaks the test script, which we then need to fix by basically undoing the changes to the test script.
Perhaps moving to a scenario where the test file actually tests a different script which the commits modify would be a little more intuitive.
I'm not sure how we could achieve this yet. My original suggestion was having a script called
zero.shwhich does some simple math and exits with the result, e.g.exit (2 - 1 -1), and havetest.shjust:The commits being rebased could then modify the
zero.shscript, the breaking commit making the math not end up at 0.It's also an arcane scenario, but perhaps a little more intuitive.