fix: Correct sign-check wording in lending protocol messages

[tyalymov]

High Level Overview of Change

Correct misleading sign-check wording in lending-protocol diagnostic messages. Several
vault invariant log messages and one LoanPay assert stated that a value "must be
positive" while the corresponding check only rejects negative values (zero is a valid
state). They now say "must not be negative" / "non-negative" to match the actual check.

This is diagnostic text only — there is no behavioural or consensus change.

Context of Change

Following review discussion on #7863: the vault/loan invariants use < 0 (or >= 0)
comparisons, i.e. they enforce non-negativity, but several messages read "must be
positive", which implies > 0. Zero is legitimate (e.g. an empty vault has
assetsAvailable == 0, assetsTotal == 0; assetsMaximum == 0 means "no maximum";
a vault with no impaired loans has lossUnrealized == 0), so the wording was misleading
when diagnosing invariant failures.

Reworded to match the checks:

• VaultInvariant.cpp — "assets available / assets outstanding / assets maximum must be
  positive" → "... must not be negative" (each guards < 0).
• LoanPay.cpp — assert description "positive vault and broker balances" → "non-negative
  vault and broker balances" (guards >= 0).

A full sweep of the lending-protocol files confirmed these are the only mismatches;
LoanInvariant / LoanBrokerInvariant already use "is negative" / "is zero or negative"
correctly, and the remaining "positive" mentions are code comments or genuinely
strict-positive checks (<= 0).

API Impact

• Public API: New feature (new methods and/or new fields)
• Public API: Breaking change (in general, breaking changes should only impact the next api_version)
• libxrpl change (any change that may affect libxrpl or dependents of libxrpl)
• Peer protocol change (must be backward compatible or bump the peer protocol version)

Test Plan

Updated the affected expectations in Invariants_test.cpp and confirmed the invariant
suite passes with the new wording. Both test sets are green (xrpld --unittest and
ctest --preset conan-debug). No logic changed, so no new tests were added.

[xrplf-ai-reviewer]

All clear. No issues to flag.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[Tapanito]

LGTM