feat: update sails-rs to 1.0.0

[m62624]

No description provided.

[vobradovich]

Code Review: feat/awesome-sails-beta vs master

Branch: feat/awesome-sails-beta
Base: master (7f47b1a)
Head: 99b4fd3
Date: 2026-04-27

What Was Reviewed

~45 commits covering:

• Allocation-free sorted array RBAC storage (const generics, SmallVec)
• Binary search and swap_remove for linear storage
• Benchmarking infrastructure with CI integration
• Capacity tests and CapacityExceeded typed error
• TypeInfo updates for sails-rs beta.4 compatibility
• RBAC role update

---

Strengths

• Clean separation between AccessControlStorage (pure data) and AccessControl (service wrapper).
• CapacityExceeded as a typed, encodable error instead of a panic is a meaningful correctness improvement.
• Benchmark scope reduction (10 000 → 100 members) is appropriately calibrated to new capacity constraints.
• assert_str_panic removed in favour of typed Error comparison — strictly better.
• No unsafe code introduced in the access-control crate.

---

Critical Issues

1. require_role logic broken for role_id == default_admin_role()

File: crates/awesome-sails/access-control/src/lib.rs (lines 428–451)

When role_id == default_admin_role(), the role_id != admin guard skips has_role entirely. A valid default-admin member calling require_role(default_admin_role(), actor) will get AccessDenied if the fast-path fails. The previous implementation used has_role(DEFAULT_ADMIN_ROLE) || has_role(role_id) — the new fast-path dropped this fallback.

Fix: Replace the fast-path with a proper has_role(admin, account_id) call, remove the role_id != admin guard, and add an explicit test for require_role(default_admin_role(), actor_in_admin_role).

2. roles_capacity_exceeded test doesn't verify the last valid role was accepted

File: tests/access-control-test/app/tests/gtest.rs (line 708)

Loop iterates 1..ROLES_LIMIT, but never asserts that the 32 custom-role grants succeeded. The test only checks that the ROLES_LIMIT + 1th role fails — the ROLES_LIMITth slot is never verified as successfully accepted.

---

Important Issues

3. data_idx: u16 silently overflows if N > 65535

File: crates/awesome-sails/access-control/src/lib.rs (lines 95, 294)

self.role_data.len() as u16 wraps silently for N > 65535. The capacity check prevents this in practice only if N ≤ 65535. No compile-time assertion exists.

Fix:

const _: () = assert!(N <= u16::MAX as usize, "N must fit in u16");

4. RS parameter undocumented for role_data SmallVec

File: crates/awesome-sails/access-control/src/lib.rs (line 89)

role_data uses RS inline slots (same as descriptors). This works since both grow in lockstep, but is not documented. A reader expects a separate parameter or an explanatory comment.

5. set_role_admin consumes role capacity without granting members

File: crates/awesome-sails/access-control/src/lib.rs (line 715)

set_role_admin_unchecked calls ensure_role_mut(role_id), allocating a new slot even for non-existent roles. With explicit capacity limits this side-effect is now observable — worth documenting or guarding.

6. Stale DEFAULT_ADMIN_ROLE references in doc comments

DEFAULT_ADMIN_ROLE is no longer a public symbol (replaced by default_admin_role()), but doc comments throughout lib.rs and vft-admin still reference the old name. Update to default_admin_role().

7. grant_initial_admin silently swallows CapacityExceeded when N == 0

File: crates/awesome-sails/access-control/src/lib.rs (line 279)

if let Ok(role) = self.ensure_role_mut(default_admin_role()) {
    let _ = role.add_member(deployer);
}

If N == 0, the deployer never gets the admin role and the program silently misbehaves. Add a debug_assert or document N >= 1 as a precondition.

---

Suggestions

8. Enforce index-0 invariant in require_role with a debug_assert

The assumption "default admin role is always at index 0 in a sorted array" is correct but load-bearing. Add:

debug_assert_eq!(storage.descriptors[0].role_id, admin);

9. Benchmark baseline reduction eliminates high-scale regression detection

The bench_data.json went from 5000/10000-member baselines to 100. Regressions that only appear at higher scales are no longer tracked. Document this trade-off in benchmarks/README.md.

10. stress_test_max_members doesn't guard against role-slot exhaustion during setup

The test doesn't assert that the MINTER_ROLE grant succeeded before iterating members. If ROLES_LIMIT were 1, it would fail with CapacityExceeded on the role grant rather than a member add.

11. vft-admin RS == N and MS == M is intentional but unexplained

File: crates/awesome-sails/vft-admin/src/lib.rs (lines 45–48)

With ROLES_LIMIT = 4 and MEMBERS_LIMIT = 17, the SmallVec always holds everything on the stack. Valid intent — but a comment explaining why RS == N and MS == M are correct would help readers.

12. README code example has a formatting/indentation error

File: crates/awesome-sails/access-control/README.md (line 56)

The pub fn access_control( line is missing leading indentation inside the impl Program block, making the example non-compiling as shown.

---

Summary

#	Severity	Area	Issue
1	Critical	require_role logic	Admin bypass path broken for role_id == default_admin_role()
2	Critical	Capacity test	roles_capacity_exceeded doesn't verify the last valid role was accepted
3	Important	data_idx overflow	No compile-time assertion that N <= u16::MAX
4	Important	RS parameter	role_data SmallVec uses RS but undocumented
5	Important	set_role_admin	Creates empty role slots, consuming capacity silently
6	Important	Stale docs	DEFAULT_ADMIN_ROLE in doc comments throughout
7	Important	grant_initial_admin	Silent CapacityExceeded failure when N == 0
8	Suggestion	require_role	Add debug_assert for index-0 assumption
9	Suggestion	Benchmarks	Baseline reduction eliminates high-scale regression detection
10	Suggestion	Capacity tests	stress_test_max_members doesn't guard against role-slot exhaustion
11	Suggestion	vft-admin sizing	RS == N and MS == M is intentional but unexplained
12	Suggestion	README	Indentation error in code example

[github-actions]

🔬 Click to see benchmark results vs master

🔬 Benchmark Comparison

Metric	Current	Baseline	Change	%	Status
access_control.grant_role.0	950_403_467	990_713_476	-40_310_009	-4.07%	👍
access_control.grant_role.100	1_015_694_788	852_534_024	+163_160_764	19.14%	❌
access_control.grant_role.25	984_009_704	-	-	-	✨
access_control.grant_role.50	994_648_346	-	-	-	✨
access_control.has_role.0	965_248_793	837_548_725	+127_700_068	15.25%	❌
access_control.has_role.100	812_412_203	850_763_647	-38_351_444	-4.51%	👍
access_control.has_role.25	811_950_519	-	-	-	✨
access_control.has_role.50	812_181_361	-	-	-	✨
access_control.has_role_multi.1	965_248_793	837_548_725	+127_700_068	15.25%	❌
access_control.has_role_multi.16	851_426_257	-	-	-	✨
access_control.has_role_multi.32	851_657_692	-	-	-	✨
access_control.has_role_multi.8	851_406_759	-	-	-	✨
access_control.revoke_role.0	974_464_581	844_228_279	+130_236_302	15.43%	❌
access_control.revoke_role.100	1_017_811_082	857_192_808	+160_618_274	18.74%	❌
access_control.revoke_role.25	986_081_598	-	-	-	✨
access_control.revoke_role.50	996_735_040	-	-	-	✨

Legend

• 🚀 Significant improvement (>5% reduction)
• 👍 Minor improvement (<1% reduction)
• ✅ No significant change
• ⚠️ Minor regression (>1% increase)
• ❌ Significant regression (>5% increase)
• ✨ New metric (not present in baseline)
• 🗑️ Removed metric (not present in current)

---

_{🤖 This comment was automatically generated. Baseline: branch master.}

[github-actions]

🔬 Click to see benchmark results vs master

🔬 Benchmark Comparison

Metric	Current	Baseline	Change	%	Status
access_control.grant_role.0	948_553_136	990_713_476	-42_160_340	-4.26%	👍
access_control.grant_role.100	1_013_844_457	852_534_024	+161_310_433	18.92%	❌
access_control.grant_role.25	982_159_373	-	-	-	✨
access_control.grant_role.50	992_798_015	-	-	-	✨
access_control.has_role.0	963_398_462	837_548_725	+125_849_737	15.03%	❌
access_control.has_role.100	810_561_872	850_763_647	-40_201_775	-4.73%	👍
access_control.has_role.25	810_100_188	-	-	-	✨
access_control.has_role.50	810_331_030	-	-	-	✨
access_control.has_role_multi.1	963_398_462	837_548_725	+125_849_737	15.03%	❌
access_control.has_role_multi.16	849_575_926	-	-	-	✨
access_control.has_role_multi.32	849_807_361	-	-	-	✨
access_control.has_role_multi.8	849_556_428	-	-	-	✨
access_control.revoke_role.0	972_614_250	844_228_279	+128_385_971	15.21%	❌
access_control.revoke_role.100	1_015_960_751	857_192_808	+158_767_943	18.52%	❌
access_control.revoke_role.25	984_231_267	-	-	-	✨
access_control.revoke_role.50	994_884_709	-	-	-	✨

Legend

• 🚀 Significant improvement (>5% reduction)
• 👍 Minor improvement (<1% reduction)
• ✅ No significant change
• ⚠️ Minor regression (>1% increase)
• ❌ Significant regression (>5% increase)
• ✨ New metric (not present in baseline)
• 🗑️ Removed metric (not present in current)

---

_{🤖 This comment was automatically generated. Baseline: branch master.}

[github-actions]

🔬 Click to see benchmark results vs master

🔬 Benchmark Comparison

Metric	Current	Baseline	Change	%	Status
access_control.grant_role.0	948_553_136	990_713_476	-42_160_340	-4.26%	👍
access_control.grant_role.100	1_013_844_457	852_534_024	+161_310_433	18.92%	❌
access_control.grant_role.25	982_159_373	-	-	-	✨
access_control.grant_role.50	992_798_015	-	-	-	✨
access_control.has_role.0	963_398_462	837_548_725	+125_849_737	15.03%	❌
access_control.has_role.100	810_561_872	850_763_647	-40_201_775	-4.73%	👍
access_control.has_role.25	810_100_188	-	-	-	✨
access_control.has_role.50	810_331_030	-	-	-	✨
access_control.has_role_multi.1	963_398_462	837_548_725	+125_849_737	15.03%	❌
access_control.has_role_multi.16	849_575_926	-	-	-	✨
access_control.has_role_multi.32	849_807_361	-	-	-	✨
access_control.has_role_multi.8	849_556_428	-	-	-	✨
access_control.revoke_role.0	972_614_250	844_228_279	+128_385_971	15.21%	❌
access_control.revoke_role.100	1_015_960_751	857_192_808	+158_767_943	18.52%	❌
access_control.revoke_role.25	984_231_267	-	-	-	✨
access_control.revoke_role.50	994_884_709	-	-	-	✨

Legend

• 🚀 Significant improvement (>5% reduction)
• 👍 Minor improvement (<1% reduction)
• ✅ No significant change
• ⚠️ Minor regression (>1% increase)
• ❌ Significant regression (>5% increase)
• ✨ New metric (not present in baseline)
• 🗑️ Removed metric (not present in current)

---

_{🤖 This comment was automatically generated. Baseline: branch master.}