test/with-pgenvblock: add additional test

[alxndrsn]

Example outputs

Old version

Passing

$ ./test/test-with-pgenvblock.sh 
Checking whether with-pgenvblock.pl works… Yes

Failing

$ ./test/test-with-pgenvblock.sh 
Checking whether with-pgenvblock.pl works… No, it doesn't

New version

Passing

$ ./test/test-with-pgenvblock.sh 
[test-with-pgenvblock] Testing...
[test-with-pgenvblock]   Testing: specific postgres-related variables...
[test-with-pgenvblock]     Passed OK.
[test-with-pgenvblock]   Testing: local env with injected var...
[test-with-pgenvblock]     Passed OK.
[test-with-pgenvblock] All tests passed OK.

Failing first test

$ ./test/test-with-pgenvblock.sh 
[test-with-pgenvblock] Testing...
[test-with-pgenvblock]   Testing: specific postgres-related variables...
2c2
< PGDATBASE=4
---
> PGDATABASE=4
[test-with-pgenvblock] !!!
[test-with-pgenvblock] !!! Test failed; see above for differences between '< actual' and '> expected'
[test-with-pgenvblock] !!!

Failing second test

$ ./test/test-with-pgenvblock.sh 
[test-with-pgenvblock] Testing...
[test-with-pgenvblock]   Testing: specific postgres-related variables...
[test-with-pgenvblock]     Passed OK.
[test-with-pgenvblock]   Testing: local env with injected var...
[test-with-pgenvblock]     !!! Test failed.

---

Extracted from discarded PR #2119

Should help with #1793

What has been done to verify that this works as intended?

It's a new test - the test was run.

Also it allowed identification of #2203.

Why is this the best possible solution? Were any other approaches considered?

The logging adds clarity around execution of the existing test, and adds an additional test which makes specific test cases easier to add.

How does this change impact users? Describe intentional behavior changes from code updates. What are the regression risks?

No change - just tests.

Does this change require updates to documentation? If so, please file an issue here and include the link below.

No.

[matthew-white]

I went ahead and tagged @lognaturel, since IIRC she reviewed the original version of this code.

[alxndrsn]

It looks like the old test relied on being the last-executed line, so no longer fails the script if the test fails.

[alxndrsn]

This structure had to change because it wasn't triggering set -e handling due to its structure:

The shell does not exit if the command that fails is part of ... any command executed in a && or || list except the command following the final && or ||
-https://doc.guix.gnu.org/bash/latest/en/html_node/The-Set-Builtin.html