Give each started Automaton instance its own timers

[simpleqt]

Problem

The @ATMT.timeout and @ATMT.timer decorators create one Timer object per decorated condition, registered on the class (scapy/automaton.py:596 def deco(f, state=state, timeout=Timer(timeout)), collected into the class-level cls.timeout dict). Instances never shadow that dict, so every running instance of the same automaton shares the same Timer objects.

A Timer's runtime state (_time, _expired, _just_expired) is per-run bookkeeping that is consumed exactly once:

• _decrement() flips _expired the first time the remaining time hits zero,
• expired() hands out the timers whose _just_expired is set and resets that flag for whoever asked first.

With two instances of the same automaton running concurrently — exactly what Automaton.spawn() ("start the automaton for each new client") and @ATMT.ioevent(as_supersocket=...) do — whichever control thread decrements a shared timer first consumes the expiration. The other instance's until_next() keeps returning None (blocking) for that timer, so its select() blocks forever: the timeout silently never fires for it. spawn()'s housekeeping only reclaims instances with isrunning() == False, so the wedged instance also leaks its sockets.

Measured on current master (two instances, one 0.2s timeout each):

unpatched: a.isrunning()=False  b.isrunning()=True   <- b never observed its timeout
patched:   both instances reach END within the timeout

Changes

_do_control copies the class-level timers into a per-instance self.timeout when the instance starts (after parse_args(), so pre-start reconfiguration through timer_by_name() — which mutates the class-level timers — is still honored; _func and the configuration stay shared, only the runtime state becomes per-instance).

Existing semantics verified unchanged: the timer-count assertions from automaton.uts (10/6 counts, timer_by_name timeouts), the pre-start "reconfigure timers" flow, and class defaults staying pristine across repeated runs.

Testing

New case in test/scapy/automaton.uts races two instances of a one-timeout automaton and asserts both observe their own timeout (fails on unpatched master by hanging one instance; passes with the patch). Locally:

$ python verify_race.py        # unpatched
a done: False a.fired: False   <- instance a wedged
$ python verify_race.py        # patched
OK: both instances observed their own timeout

flake8 (max-line 88, repo ignores) clean on the changed file aside from pre-existing F401s that only appear with pyflakes ≥3 (type-comment usage; invisible to the pinned flake8<6 CI uses).

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 80.87%. Comparing base (e4742a5) to head (336482a).

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #5199      +/-   ##
==========================================
- Coverage   80.92%   80.87%   -0.05%     
==========================================
  Files         393      393              
  Lines       98107    98114       +7     
==========================================
- Hits        79392    79354      -38     
- Misses      18715    18760      +45     

Files with missing lines	Coverage Δ
scapy/automaton.py	80.75% <100.00%> (-1.02%)	⬇️

... and 27 files with indirect coverage changes

🚀 New features to boost your workflow:

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