fix: correct mempool test formatting failure

[tolga-tom-nook]

Summary

• changes a t.Fatal call with a %d formatting directive to t.Fatalf
• fixes the Go vet/build failure in the mempool test package

Root cause / acceptance proof

On the upstream BitgesellOfficial/bgld master branch, go test ./mempool currently fails before tests run because vet rejects the formatting directive passed to t.Fatal:

mempool/mempool_test.go:563:4: (*testing.common).Fatal call has possible Printf formatting directive %d
FAIL	github.com/bitgesellofficial/bgld/mempool [build failed]

This PR uses t.Fatalf, which is the intended testing helper when the message has printf-style formatting plus arguments.

Validation

• RED on upstream master: go test ./mempool fails with the vet error above
• GREEN on this branch: go test ./mempool

ok  	github.com/bitgesellofficial/bgld/mempool	0.169s

Bounty note

Submitting as a focused test/build reliability improvement for the Bitgesell bounty/improvement program, especially BitgesellOfficial/bitgesell#81. If approved for payout, USDT/EVM-compatible address: 0x4a76c7E64C08cF29B59eFC640b4ada97A270d428.

Assisted by Hermes Agent.

[MyTH-zyxeon]

Maintainers: this PR looks like a clean Bitgesell btcsuite#81 candidate because it fixes a real go test ./mempool blocker with a one-line, behavior-preserving change.

Acceptance checks I would use before merge:

• confirm the upstream RED/GREEN proof still reproduces exactly as written (go test ./mempool failing on master, then passing on this branch)
• verify the helper change is intentionally limited to formatting semantics (t.Fatal -> t.Fatalf) and does not alter orphan-processing logic or test expectations
• if you track docs-only vs build-fix payout tiers differently under save mempool to disk btcsuite/btcd#81, treat this as a build/test reliability fix rather than a docs cleanup

From the public diff, this is the sort of narrow maintainer-friendly patch that should be easy to approve if the branch still reproduces the vet failure.