fix(gnovm): match Go's call-time dispatch for interface-bound method values

[ltzmaxwell]

• Interface-bound method values (g := i.M, defer i.M()) now resolve their concrete method and receiver at call time, matching Go: nil-panic timing, value snapshot, embedded promotion (value/pointer/interface fields, any depth), field re-read through a boxed pointer, and dynamic re-dispatch. Also fixes two VM crashes: defer i.M() on a non-nil interface, and a nil embedded pointer-receiver (c5).
• Mechanism: BoundMethodValue gains a Method field; Func == nil marks a lazy bind resolved at the call (resolveLazyBound); deferred callables are unified onto Defer.Callable (type-switched like doOpPrecall). Concrete/pointer binds (pt.M, t.M) are unchanged.
• Hard fork: Method is persisted (proto field 4; pb3_gen regenerated), _allocBoundMethodValue 200→216 → wire/merkle/gas change; coordinated-upgrade / fresh-genesis only (like ADR-5544), old state still decodes.
• Gas: the dispatch walk moved bind→call, so both consts were re-fit — OpCPUSelectorInterface 751→276 (the lazy bind no longer walks) and new OpCPULazyBoundResolve=529 (the call-time walk); net ≈ +54/interface call. Ratio-scaled (reference HW unavailable) — calibration TODO before the fork.
• Tests: classed filetests (timing / nil / embedded / operand-capture / dynamic / defer) + persist and boxed-cur realm txtars; rationale in gnovm/adr/pr5737_nil_value_method_panic_timing.md. Pre-existing orthogonal divergences filed separately: method expressions don't support interface, promoted, or mixed-receiver forms #5787, GnoVM: calling a method on a nil interface panics uncatchably instead of a recoverable nil-pointer panic #5850.

[Gno2D2]

🛠 PR Checks Summary

All Automated Checks passed. ✅

Manual Checks (for Reviewers):

• IGNORE the bot requirements for this PR (force green CI check)

Read More

🤖 This bot helps streamline PR reviews by verifying automated checks and providing guidance for contributors and reviewers.

✅ Automated Checks (for Contributors):

🟢 Maintainers must be able to edit this pull request (more info)

☑️ Contributor Actions:

1. Fix any issues flagged by automated checks.
2. Follow the Contributor Checklist to ensure your PR is ready for review.
   • Add new tests, or document why they are unnecessary.
   • Provide clear examples/screenshots, if necessary.
   • Update documentation, if required.
   • Ensure no breaking changes, or include BREAKING CHANGE notes.
   • Link related issues/PRs, where applicable.

☑️ Reviewer Actions:

1. Complete manual checks for the PR, including the guidelines and additional checks if applicable.

📚 Resources:

• Report a bug with the bot.
• View the bot’s configuration file.

Debug

Automated Checks

Maintainers must be able to edit this pull request (more info)

If

🟢 Condition met
└── 🟢 And
    ├── 🟢 The base branch matches this pattern: ^master$
    └── 🟢 The pull request was created from a fork (head branch repo: ltzmaxwell/gno)

Then

🟢 Requirement satisfied
└── 🟢 Maintainer can modify this pull request

Manual Checks

**IGNORE** the bot requirements for this PR (force green CI check)

If

🟢 Condition met
└── 🟢 On every pull request

Can be checked by

• Any user with comment edit permission

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[davd-gzl]

This is a design problem, not a tweak. The root cause sits a level deeper than the single flag site, so as written the change fixes the interface case but regresses the concrete one. Go has two distinct timings: a concrete *T(nil) derefs *pt when the method value is formed, so it panics eagerly at bind, while an interface-boxed *T only derefs inside the call, so its panic lands at call time. The nilReceiverPanic flag is read at the single VPDerefValMethod site, but that site runs after the interface is unwrapped to a bare *T, so it can't tell the two apart and defers both, when only the interface case should defer. To hold both, the decision needs to live where the receiver kind is still known (interface dispatch, or marked in the preprocessor) rather than at that one late site.

receiver form        Go            master         this PR
-------------------  ------------  -------------  --------------
defer pt.M() concrete eager(=0)    eager(=0) ✓    call-time(=1) ✗   <- regressed
defer i.M()  iface    call-time(=1) eager(=0) ✗   call-time(=1) ✓   <- fixed
G=i.M; persist; G()   panic         (binds eager)  no panic ✗        <- new hole

The current filetest covers the interface case; the concrete case and the persistence hole below are both unexercised today, so they'd be worth adding. An ADR pinning the intended semantics would also help anchor the fix.

Full review: https://github.com/samouraiworld/gno-agent-workspace/blob/main/reviews/pr/5xxx/5737-defer-nil-receiver-panic/1-4c57c37e4/claude-opus-4-8_davd-gzl.md

Repros run at 4c57c37.