fix: last test line index

[mark2185]

Make sure these boxes are checked before submitting your pull request:

• Add fixtures and spec when implementing or updating a test runner
• Update the README accordingly
• Update the Vim documentation in doc/test.txt

Without the rewrite of the for loop there was an off-by-two error when searching downwards.

Running just the tests without the fix in place:

-------- Testing a Vim plugin
spec/zig_spec.vim .. 1/?
not ok 5 - Zig update g:test#last_position with exact line of the test
# Expected g:test#last_position['line'] == 4 at line 4
#       Actual value: 6
#     Expected value: 4
spec/zig_spec.vim .. Failed 1/5 subtests

Test Summary Report
-------------------
spec/zig_spec.vim (Wstat: 0 Tests: 5 Failed: 1)
  Failed test:  5
Files=1, Tests=5,  1 wallclock secs ( 0.01 usr  0.00 sys +  0.05 cusr  0.04 csys =  0.10 CPU)
Result: FAIL

The last_position['line'] is used in test#visit to jump to the place where the test was invoked from, but I think it makes more sense that TestVisit actually visits the test it ran.

In case the user ran TestFile, it won't update the last_position['line'].

Note: the PR contains the PR #903 because it tests its new functionality (searching downwards) so that one should get merged first. Rebased!

Note: TestVisit for test runners other than zig will still exhibit the current behavior.

[mark2185]

@codeinabox I also have a patch that will enable the test runner to run tests that don't have a string literal as a name, but an identifier (test addOne { ... that's already in the test file), should I open up a new PR for that or just chuck it in here as a separate commit?

It's just an additional regex pattern and a spec update.

[codeinabox]

Thank you!