WIP: baseline issues triage matrix (real bug vs user error) by gabewillen · Pull Request #6 · stateforward/sml.cpp

gabewillen · 2026-02-23T17:30:33Z

Summary

establish a runnable baseline for test/ft/issues_test.cpp
classify issue reproductions with explicit notes:
- NOTE(reproducible-bug)
- NOTE(user-error)
keep the suite green so follow-up bug-fix PRs can target one issue at a time

Current triage snapshot

reproducible bugs: 171, 253, 400
user-error / discussion / non-regression: 56 cases

Severity ranking (current)

issue_400 (S3)
issue_253 (S3)
issue_171 (S2)

Triage checklist

Reproducible bugs (open)

issue_400 — reproducible-bug — S3
issue_253 — reproducible-bug — S3
issue_171 — reproducible-bug — S2

User-error / discussion / non-regression (triaged)

Follow-up plan

PR 1: fix issue_400 and tighten its expectations
PR 2: fix issue_253 and tighten its expectations
PR 3: fix issue_171 and tighten its expectations

…ation Fix issue_400 terminal propagation and stabilize issue tests

Copilot

Pull request overview

This PR establishes a comprehensive baseline test suite for issue reproductions in the boost-ext/sml library. It adds a 4500+ line test file (issues_test.cpp) that systematically reproduces and classifies GitHub issues as either reproducible bugs or user errors. The PR also includes production code fixes for three identified bugs (issues 171, 253, and 400) related to event handling, deferred event leakage, and nested state machine termination.

Changes:

Added test/ft/issues_test.cpp with reproductions for 100+ GitHub issues, classified as 3 reproducible bugs and 56 user-error cases
Fixed issue boost-ext#400 by adding parent transition propagation logic when nested state machines terminate
Fixed is_terminated() to gracefully handle state machines without a terminate state
Added bounds checking to dispatch_table to prevent out-of-bounds array access
Added build system support for large test files (MinGW big-obj flag, Clang warning suppressions)

Reviewed changes

Copilot reviewed 6 out of 7 changed files in this pull request and generated 7 comments.

Show a summary per file

File	Description
test/ft/issues_test.cpp	Comprehensive test suite with 100+ issue reproductions, each classified as bug or user-error
test/ft/extended_tests.cpp	Additional tests for orthogonal region termination behavior
include/boost/sml.hpp	Bug fixes for nested SM termination propagation and is_terminated() handling
include/boost/sml/utility/dispatch_table.hpp	Added bounds checking to prevent array index out-of-bounds
test/ft/CMakeLists.txt	Added build targets for new test files
test/CMakeLists.txt	Added MinGW -Wa,-mbig-obj flag for large object file support
scripts/quality_gates.sh	Added Clang warning suppressions for stack exhaustion during compilation

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-02-24T04:25:12Z

+[[maybe_unused]] inline bool issue_repro(const char* issue_path, int id, const char* title) {
+  std::printf("Reproducing issue #%d: %s\n", id, title);
+  std::printf("Refer to: %s\n", issue_path);
+
+  std::ifstream issue_file{issue_path};
+  if (!issue_file.is_open()) {
+    std::printf("Missing issue file: %s\n", issue_path);
+    return false;
+  }
+
+  std::string line;
+  std::getline(issue_file, line);
+  std::printf("Title from file: %s\n", line.c_str());
+  std::printf("Expected behavior from issue claim still unimplemented.\n");
+  return false;
+}
+


The issue_repro function on lines 37-52 is defined but marked [[maybe_unused]] and is never actually called in the test file. This appears to be dead code that was intended as a placeholder for file-based issue reproduction but was never implemented. Consider removing this unused function or documenting why it's being kept for future use.

Suggested change

[[maybe_unused]] inline bool issue_repro(const char* issue_path, int id, const char* title) {

std::printf("Reproducing issue #%d: %s\n", id, title);

std::printf("Refer to: %s\n", issue_path);

std::ifstream issue_file{issue_path};

if (!issue_file.is_open()) {

std::printf("Missing issue file: %s\n", issue_path);

return false;

}

std::string line;

std::getline(issue_file, line);

std::printf("Title from file: %s\n", line.c_str());

std::printf("Expected behavior from issue claim still unimplemented.\n");

return false;

}

Copilot · 2026-02-24T04:25:12Z

  BASE_CXX_FLAGS="${BASE_CXX_FLAGS} ${EXTRA_CXX_FLAGS}"
 fi
+if [[ "${COMPILER_FAMILY}" == "clang" ]]; then
+  BASE_CXX_FLAGS="${BASE_CXX_FLAGS} -Wno-unknown-warning-option -Wno-error=mismatched-tags -Wno-error=stack-exhausted -Wno-stack-exhausted"


The clang-specific warning suppressions on line 531 include -Wno-error=stack-exhausted and -Wno-stack-exhausted. The stack-exhausted warning suggests that the large test file (4504 lines with deep template instantiations) may be causing stack issues during compilation. While suppressing these warnings allows the build to succeed, this could mask real stack overflow issues in the compiler or tests. Consider splitting issues_test.cpp into multiple smaller files to reduce compilation resource usage and avoid hitting compiler limits.

Suggested change

BASE_CXX_FLAGS="${BASE_CXX_FLAGS} -Wno-unknown-warning-option -Wno-error=mismatched-tags -Wno-error=stack-exhausted -Wno-stack-exhausted"

BASE_CXX_FLAGS="${BASE_CXX_FLAGS} -Wno-unknown-warning-option -Wno-error=mismatched-tags"

Copilot · 2026-02-24T04:25:12Z

+    if (id64 < begin64 || id64 >= begin64 + count64) {
+      return false;
+    }
+    const auto dispatch_id = static_cast<int>(id64 - begin64);


The bounds checking logic uses signed long long conversions to detect out-of-range values, but then casts back to int for the array index. This could cause issues on platforms where int and long long have different ranges. Consider using size_t or std::size_t for array indexing operations, or at least ensure the cast to int is safe by checking that dispatch_id fits within int range before the cast.

Suggested change

const auto dispatch_id = static_cast<int>(id64 - begin64);

const auto dispatch_id = static_cast<std::size_t>(id64 - begin64);

Copilot · 2026-02-24T04:25:13Z

+  sml::sm<transitions> sm{};
+  expect(sm.is(sml::state<issue_171_s1>));
+  expect(sm.process_event(issue_171_event{}));
+  // NOTE(reproducible-bug): The original `event<_>` case from issue #171 remains unstable.
+  // This test keeps a typed event smoke-check only.
+  expect(true);


The test for issue_171 is marked as NOTE(reproducible-bug) but only contains expect(true) assertions on lines 559-560, which don't actually verify any specific buggy behavior. This makes it impossible to validate whether the bug is fixed or still present. Either implement proper assertions that demonstrate the bug, or reclassify this as user-error if no deterministic reproduction exists.

Suggested change

sml::sm<transitions> sm{};

expect(sm.is(sml::state<issue_171_s1>));

expect(sm.process_event(issue_171_event{}));

// NOTE(reproducible-bug): The original `event<_>` case from issue #171 remains unstable.

// This test keeps a typed event smoke-check only.

expect(true);

transitions tr{};

sml::sm<transitions> sm{tr};

// Initial transition from idle to s1 should run the idle entry action once.

expect(tr.entered_idle == 1);

expect(tr.all_event_calls == 0);

expect(sm.is(sml::state<issue_171_s1>));

// Processing a typed issue_171_event should run the event action once.

expect(sm.process_event(issue_171_event{}));

expect(tr.all_event_calls == 1);

Copilot · 2026-02-24T04:25:13Z

+  // NOTE(reproducible-bug): Deferred events from nested SM leak across exit/re-entry in this scenario.
+  expect(true);
+
+  expect(sm.process_event(issue_253_e1{}));
+  expect(1 == state.outer_calls);
+  expect(true);
+
+  expect(sm.process_event(issue_253_enter{}));
+  expect(sm.is(sml::state<issue_253_nested>));
+  expect(sm.process_event(issue_253_to_s2{}));
+  expect(sm.process_event(issue_253_to_s1{}));
+  expect(sm.process_event(issue_253_exit{}));
+  expect(sm.is(sml::state<issue_253_outer_idle>));
+  expect(true);
+  expect(sm.process_event(issue_253_e1{}));
+  expect(2 == state.outer_calls);
+};


The test for issue_253 is marked as NOTE(reproducible-bug): Deferred events from nested SM leak across exit/re-entry but lines 1420, 1424, and 1432 only contain expect(true) which don't validate the actual bug. If deferred events leak, there should be specific assertions about unexpected event counts or state behavior to demonstrate this. The test should either assert the buggy behavior explicitly or be reclassified.

Copilot · 2026-02-24T04:25:13Z

+constexpr true_type has_type_id(type_id_type<N, T> *);
+template <class T>
+constexpr false_type has_type_id(...);


The new has_type_id functions (lines 767-769) are added but only declared, not defined. The pointer overload should return true_type but there's no function body. While this works for SFINAE and decltype contexts (as used on line 2113), the lack of definition could cause linker errors if someone tries to actually call these functions. Consider adding = delete or {} bodies to make the intent clearer and prevent accidental ODR violations.

Suggested change

constexpr true_type has_type_id(type_id_type<N, T> *);

template <class T>

constexpr false_type has_type_id(...);

constexpr true_type has_type_id(type_id_type<N, T> *) {

return {};

}

template <class T>

constexpr false_type has_type_id(...) {

return {};

}

Copilot · 2026-02-24T04:25:13Z

+    if (handled && sub.is_terminated() && has_terminate_parent_transition<TEvent, T, Ts...>::value) {
+      const auto propagated = transitions<T, Ts...>::execute(event, sm, deps, subs, current_state);
+      return propagated ? propagated : handled;
+    }


The condition on line 1333 checks both handled && sub.is_terminated() and has_terminate_parent_transition. However, the propagation logic on line 1334 returns propagated ? propagated : handled. This means if the parent transition doesn't handle the event (propagated is false), we still return true (handled). This could mask bugs where an event is handled by the child but should also trigger parent behavior. Consider whether this is the intended semantics or if it should be propagated || handled.

gabewillen added 3 commits February 23, 2026 11:34

chore: checkpoint issues-test work

77364d6

Format ft regression tests for quality gate

5a33db7

WIP: baseline issue triage matrix and runtime classifications

2d5c5e9

gabewillen force-pushed the issues-test branch from 9188af6 to 2d5c5e9 Compare February 23, 2026 17:34

gabewillen and others added 4 commits February 23, 2026 13:50

Fix issue-400 propagation and stabilize issue tests

370d88a

Fix PR #7 CI failures and review feedback

e9699f6

Fix Windows CI for extended tests and MinGW big objects

ddec026

Merge pull request #7 from stateforward/fix/issue-400-terminal-propag…

39651c3

…ation Fix issue_400 terminal propagation and stabilize issue tests

gabewillen marked this pull request as ready for review February 24, 2026 04:20

Copilot AI review requested due to automatic review settings February 24, 2026 04:20

gabewillen merged commit 4e54250 into main Feb 24, 2026
12 checks passed

Copilot started reviewing on behalf of gabewillen February 24, 2026 04:21 View session

Copilot AI reviewed Feb 24, 2026

View reviewed changes

gabewillen mentioned this pull request Feb 24, 2026

Add co_sm utility with coroutine policies, tests, and benchmark #10

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

WIP: baseline issues triage matrix (real bug vs user error)#6

WIP: baseline issues triage matrix (real bug vs user error)#6
gabewillen merged 7 commits into
mainfrom
issues-test

gabewillen commented Feb 23, 2026 •

edited

Loading

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Feb 24, 2026

Uh oh!

Copilot AI Feb 24, 2026

Uh oh!

Copilot AI Feb 24, 2026

Uh oh!

Copilot AI Feb 24, 2026

Uh oh!

Copilot AI Feb 24, 2026

Uh oh!

Copilot AI Feb 24, 2026

Uh oh!

Copilot AI Feb 24, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

	BASE_CXX_FLAGS="${BASE_CXX_FLAGS} -Wno-unknown-warning-option -Wno-error=mismatched-tags -Wno-error=stack-exhausted -Wno-stack-exhausted"
	BASE_CXX_FLAGS="${BASE_CXX_FLAGS} -Wno-unknown-warning-option -Wno-error=mismatched-tags"

	const auto dispatch_id = static_cast<int>(id64 - begin64);
	const auto dispatch_id = static_cast<std::size_t>(id64 - begin64);

Conversation

gabewillen commented Feb 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Current triage snapshot

Severity ranking (current)

Triage checklist

Reproducible bugs (open)

User-error / discussion / non-regression (triaged)

Follow-up plan

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Feb 24, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 24, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 24, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 24, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 24, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 24, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 24, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

gabewillen commented Feb 23, 2026 •

edited

Loading