self-host: gen-c passes leading-zero, bad-binary-digit and misplaced-underscore literals to C #7319

Profile (read first)

You are working as Gamma (Compiler Core) on epic #5980: rewrite t27c in t27 itself.
The self-hosted core (specs/compiler/core/t27core.t27) must write exactly the bytes
t27c gen-c writes, so every gen-c defect is a defect the core has to refuse instead of learn.
Rules for this task:

• Change the Rust lexer only inside the Boundary below. gen/ is generated: never hand-edit it.
• Build with cargo build --release -p t27c and use ./target/release/t27c from the repository
  root, not a t27c on PATH.
• Write the regression test first and watch it fail on the unfixed compiler. Then fix, and
  revert the fix once to watch the test go red again. Say in the PR that you did both.
• Any spec that stops compiling must be listed in the PR by path with its fix. A silent corpus
  change is a failed task.
• The PR body says Closes # plus this issue's number and Part of #5980. ASCII only in source files.

Context

gen-c's number scan takes any run of digits, x X b B _ and (after 0x) hex letters, and writes
what it read to C. Several malformed literals therefore pass with exit 0:

• a decimal with a leading zero reaches C as an OCTAL constant: 010 and 0_10 are written as
  010, which C reads as 8. A test asserting D == 10 compiles and traps at run time. 08
  does not compile at all (invalid digit "8" in octal constant);
• 0b102 is written as 0b102 (invalid digit "2" in binary constant);
• 1_, 1__0 and 0x_1F are accepted and quietly normalized to 1, 10 and 0x1F.

The Zig reference refuses every one of these (std.zig.number_literal failures leading_zero,
invalid_digit, trailing_underscore, repeated_underscore, invalid_underscore_after_special),
so the fix is a lexer error, not a rewrite. 0o17 becoming 0 is the separate #2645.
The core (P1 slice 2) refuses all of these with E_LEX.

Reproduce

cargo build --release -p t27c
for v in 010 0_10 08 0b102 1_ 1__0 0x_1F; do printf 'module m;\nconst D: i64 = %s;\n' "$v" > /tmp/lit.t27; ./target/release/t27c gen-c /tmp/lit.t27 >/dev/null 2>&1; echo "$v exit $?"; done

Today every line ends in exit 0.

Acceptance criteria

• 1. The Reproduce loop prints a non-zero exit for each of the seven literals (today exit 0 for all seven)
• 2. The same loop over 9 10 1_000 0x1_F 0b1_0 0 0x00 prints exit 0 for each (today too)
• 3. cargo test --release -p t27c --test genc_number_literals passes

Boundary

• bootstrap/src/compiler.rs
• bootstrap/tests/genc_number_literals.rs

User Scenarios & Testing

• Given a spec with const D: i64 = 010;,
  when ./target/release/t27c gen-c runs on it,
  then it exits non-zero with an error naming the literal and its line, instead of writing an octal 8.

Requirements

• FR-001: the lexer MUST refuse a decimal literal with a leading zero and more digits, a binary digit other than 0 or 1, and a _ that is not between two digits of the literal's radix.
• FR-900: cargo test --release -p t27c --test genc_number_literals MUST pass, and MUST fail with the fix reverted.
• FR-901: cargo test --release -p t27c --test core_selfhost and the corpus suite (./target/release/t27c suite --repo-root .) MUST still pass.