Avoid a potential underflow when iterating over invalid inputs

[dirkmueller]

The forward iteration logic already bounds-check for m_it != m_string->end(), do the same for the backward iteration. The issue with the assert is that the assert() might not be compiled in, and it is happening after the dereference, so it was too late.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 91.26%. Comparing base (300c5d3) to head (3fa9890).
⚠️ Report is 13 commits behind head on devel.

Additional details and impacted files

@@            Coverage Diff             @@
##            devel    #3096      +/-   ##
==========================================
- Coverage   91.36%   91.26%   -0.10%     
==========================================
  Files         204      204              
  Lines        8903     8905       +2     
==========================================
- Hits         8134     8127       -7     
- Misses        769      778       +9     

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[horenmar]

Thanks, but the change should have tests.

[dirkmueller]

Thanks, but the change should have tests.

testcase added.

[dirkmueller]

ok, the testcase is failing some check, no idea what to do now. any hints?

[horenmar]

To quote from contributing.md.

If you added new tests, you will likely see ApprovalTests failure. After you check that the output difference is expected, you should run tools/scripts/approve.py to confirm them, and include these changes in your commit.

Note that I am not sure tests with broken unicode will have portable serialization, in which case you need to add [approvals] tag to the TEST_CASE, which excludes the test from approval testing.