Bug: deterministic simulator stack overflow hides DB divergence

[yupme-bot]

Deterministic repro. Happens 100% of runs.

Reproduction

cargo run --bin limbo_sim -- --seed 123 --maximum-time 20 --doublecheck --keep-files

Environment

Commit: 57a2e66d44da8c5a479d585e9217bdbbb958121c

Rust: rustc 1.88.0

Cargo: cargo 1.88.0

OS: Windows 10 x64

Observed Behavior

With seed 123 and --doublecheck, I consistently observe two problems:

ParseError("table affectionate_lacazeduthiers_27 already exists") appears twice (likely once per pass: main + doublecheck).

Shortly after, the process terminates with:

thread 'main' has overflowed its stack
(exit code 0xc00000fd, STATUS_STACK_OVERFLOW)

Notes

The ParseError suggests the doublecheck pass may not be starting from a clean state (or state is leaking between passes).
However, I haven’t confirmed whether the stack overflow is independent or triggered by error handling/logging.

Artifacts include:

• seed123_full.txt
• test.db
• doublecheck.db
• screenshot.png

Expected Behavior

Simulator should complete without stack overflow and report divergence cleanly.

seed123_bug_report.zip

[yupme-bot]

Update: Repro also happens without --doublecheck.
Command: cargo run --bin limbo_sim -- --seed 123 --maximum-time 20 --keep-files
Observed: same ParseError("table affectionate_lacazeduthiers_27 already exists") followed by stack overflow (STATUS_STACK_OVERFLOW).