This lesson's interactive features are locked, please to keep using them

Incremental Changes

I'm not gonna lie, it does feel really nice when you write a fat, detailed prompt and refine it with the agent, then the agent one-shots a 1,000-line change in almost exactly the way you wanted it.

And that does happen.

But you can also waste a lot of time pulling the slot machine lever again and again, hoping for that sweet jackpot.

When you can break a change into smaller steps, it's still a good idea to do so. Context isn't infinite, and the smaller the changeset, the quicker it is to get the agent back on track when it inevitably isn't doing quite what you want.

Assignment

Add support for += in two small changes. This code:

slop n = 2✨
n += 3 * 4✨
hallucinate(n)✨

should print 14. The += operator adds to the variable's current value, but the right-hand expression is evaluated first, so 3 * 4 adds 12 to n.

It should also join strings and update the nearest variable with that name. Undefined names, immutable variables, and incompatible types should produce runtime errors.

Tell OpenCode to make Sloperator recognize and parse +=, and to add new tests in tests/test_compound_assignment_parser.py. Stop before making the new operator execute assignments. Inspect the diff and have it run those tests.
Next, tell it to make += execute with the behavior above, update docs/spec.md, and add regression tests in tests/test_compound_assignment.py.

Tell OpenCode to run and submit the CLI tests from the project root.