We improved our code by breaking our built-in functions to the builtin/ package, and giving that 100% test-coverage.
However our testing for the main core of the interpreter, eval/eval.go, is only at 85% and is done in a very horrid way.
We should remove all tests, and start again. We want to aim for 100% coverage, though I suspect that will also require some overhauling of the code, as things like the bounds-testing (which was necessary to avoid terminate when running fuzz-tesitng) are hard to manage.
(We could add nextToken(), etc to isolate the bounds-checking for example.)
We improved our code by breaking our built-in functions to the builtin/ package, and giving that 100% test-coverage.
However our testing for the main core of the interpreter,
eval/eval.go, is only at 85% and is done in a very horrid way.We should remove all tests, and start again. We want to aim for 100% coverage, though I suspect that will also require some overhauling of the code, as things like the bounds-testing (which was necessary to avoid terminate when running fuzz-tesitng) are hard to manage.
(We could add
nextToken(), etc to isolate the bounds-checking for example.)