fix: make stdin quit monitoring non-blocking

[MilesCranmerBot]

Fixes intermittent hangs/freezes caused by stdin quit monitoring when stdin is a pipe/non-interactive (notably on Windows).

TDD: adds a fast unit reproducer using a FakeInputStream that simulates bytesavailable>0 with a blocking read, then fixes the monitor to use non-blocking reads + robust CRLF/partial-chunk parsing.

Related PySR issues: #832, #1101, #1092.

[github-actions]

Benchmark Results (Julia v1)

Time benchmarks

	master	f3cbb63...	master / f3cbb63...
search/multithreading	23.5 ± 0.72 s	24.2 ± 1.2 s	0.969 ± 0.056
search/serial	54.9 ± 0.35 s	53.6 ± 1 s	1.02 ± 0.021
utils/best_of_sample	2.09 ± 0.77 μs	1.93 ± 0.65 μs	1.08 ± 0.54
utils/check_constraints_x10	16.8 ± 4.2 μs	16.8 ± 4.2 μs	0.998 ± 0.35
utils/compute_complexity_x10/Float64	2.16 ± 0.11 μs	2.17 ± 0.11 μs	0.996 ± 0.074
utils/compute_complexity_x10/Int64	2.11 ± 0.11 μs	2.09 ± 0.11 μs	1.01 ± 0.076
utils/compute_complexity_x10/nothing	1.65 ± 0.11 μs	1.65 ± 0.11 μs	1.01 ± 0.095
utils/insert_random_op_x10	5.89 ± 1.6 μs	5.9 ± 1.5 μs	0.999 ± 0.37
utils/next_generation_x100	0.498 ± 0.11 ms	0.492 ± 0.12 ms	1.01 ± 0.33
utils/optimize_constants_x10	0.0513 ± 0.0092 s	0.0503 ± 0.01 s	1.02 ± 0.27
utils/randomly_rotate_tree_x10	8.38 ± 0.97 μs	8.53 ± 0.97 μs	0.983 ± 0.16
time_to_load	2.9 ± 0.096 s	2.6 ± 0.097 s	1.11 ± 0.056

Memory benchmarks

	master	f3cbb63...	master / f3cbb63...
search/multithreading	0.204 G allocs: 52.1 GB	0.205 G allocs: 53.3 GB	0.978
search/serial	0.206 G allocs: 53.9 GB	0.206 G allocs: 53.9 GB	1
utils/best_of_sample	0.038 k allocs: 3.25 kB	0.038 k allocs: 3.25 kB	1
utils/check_constraints_x10	0.034 k allocs: 0.875 kB	0.034 k allocs: 0.875 kB	1
utils/compute_complexity_x10/Float64	0 allocs: 0 B	0 allocs: 0 B
utils/compute_complexity_x10/Int64	0 allocs: 0 B	0 allocs: 0 B
utils/compute_complexity_x10/nothing	0 allocs: 0 B	0 allocs: 0 B
utils/insert_random_op_x10	0.044 k allocs: 1.81 kB	0.044 k allocs: 1.81 kB	1
utils/next_generation_x100	4.62 k allocs: 0.276 MB	4.62 k allocs: 0.276 MB	1
utils/optimize_constants_x10	22.5 k allocs: 22.9 MB	21.8 k allocs: 22.5 MB	1.02
utils/randomly_rotate_tree_x10	0.042 k allocs: 1.34 kB	0.042 k allocs: 1.34 kB	1
time_to_load	0.145 k allocs: 11 kB	0.149 k allocs: 11.2 kB	0.983

[codecov]

Codecov Report

❌ Patch coverage is 85.71429% with 4 lines in your changes missing coverage. Please review.
✅ Project coverage is 95.12%. Comparing base (2f8a7ac) to head (c8b442f).

Files with missing lines	Patch %	Lines
src/SearchUtils.jl	85.71%	4 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #562      +/-   ##
==========================================
+ Coverage   95.02%   95.12%   +0.09%     
==========================================
  Files          42       42              
  Lines        3882     3900      +18     
==========================================
+ Hits         3689     3710      +21     
+ Misses        193      190       -3     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[MilesCranmer]

Looks actually pretty good.

However one thing I would like is for a test that does not rely on fake streams but on the actual REPL.

I think we can do this with Linux expect tool. Or are there other options?

@MilesCranmerBot please investigate and add such a test.

I recommend doing some mutation testing as well - try introducing silent failures - does the test catch it? Or does the test always pass?

[MilesCranmer]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Keep them coming!

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[MilesCranmer]

@MilesCranmerBot Can we do this with linux expect rather than hacking together a reimplementation

[MilesCranmerBot]

Swapped the REPL stdin nonblocking check over to a plain Linux expect script (so we’re not reimplementing an expect loop in Python).

• Added: test/repl/test_repl_stdin_nonblocking.exp
• CI: installs expect (ubuntu-latest + evaluation only) and runs the script
• Removed: test/repl/test_repl_stdin_nonblocking.py

Pushed in f3cbb63.