feat(core): for loops via peepoJuice, plus lists

[ohnishant]

Two language features in two commits.

For loops

peepoJuice no longer means not. It opens a loop:

peepoJuice PepoG i 0. peepoLessThan i 3. peepoCookie i peepoFriendship i 1. Wokege
    peepoChat i.
Bedge

The init and step clauses are optional PepoG or peepoCookie statements; leave both out and the loop works as a while. The condition is required, so there is no accidental infinite form. Header clauses and the loop variable live in one scope for the whole loop; each body iteration gets a child scope of it.

One semantic change came with this: Environment.Set used to always write into the local scope, which meant a loop body could never update a variable declared outside it. A loop that cannot accumulate state can't do anything observable, so Set now rebinds the nearest existing binding and only creates local bindings for new names. Fresh names still stay scoped to their block (TestBlockScoping covers that).

The old not operator is removed in the same commit since peepoJuice was its keyword. #3 (prefix vs postfix) is closed as superseded; picking a replacement emote for negation is now #26.

Lists

Thinking1 and Thinking2 wrap list literals and index expressions:

PepoG xs Thinking1 3, 1, 4 Thinking2.
peepoChat xs Thinking1 0 Thinking2.   # 3

Both emotes come from towdan's set per your pick. They lex as single tokens because identifiers may now contain digits. Trade-off worth knowing: PepegaCredit5 used to lex as PepegaCredit then 5, and is now one identifier. The lexer test asserts the new behavior.

Lists nest, chained indexes dig into inner lists, indexing out of range or with a non-integer is a runtime error, and peepoMeasure accepts lists alongside strings. The playground renders both new keywords as emotes via emotefetch.

Tests

• parser: header variants, round-trips through String(), malformed headers, unclosed brackets
• evaluator: counting loops, while form, scoping of the loop variable, list literals including nesting, indexing chains, out-of-range and type errors, and a loop summing a list via peepoMeasure
• lexer: digit-bearing identifiers, Thinking1/Thinking2 tokenization

go test ./..., go vet ./... and gofmt -l . are all clean.

[github-actions]

⚠️ Test files changed

This PR touches test code. Agents must not delete tests or remove
test cases without explicit owner approval.

	before	after
Test functions	40	58

Changed test files

M	cmd/evaluator/evaluator_test.go
M	cmd/lexer/lexer_test.go
M	cmd/parser/parser_test.go
M	cmd/repl/repl_test.go
M	cmd/token/token_keywords_test.go