build: bump the memtrack pin to 1.4.0

[adriencaccia]

Point the runner at the memtrack-v1.4.0 release, so runs pick up eBPF loading through a delegated BPF token and the mapped-library classification fix for environments without init-userns privilege (both from #456).

Heads up on CI: all_pinned_binaries_match_their_declared_sha256 currently fails on an unrelated pin. GitHub serves a persistent 500 for releases/download/exec-harness-v1.3.0/exec-harness-installer.sh while the same asset over the API (releases/assets/407413621) returns 200, and every other asset on that release is fine. Nothing to do with this diff, but it also means runner installs of exec-harness are failing right now.

[greptile-apps]

Greptile Summary

Updates the pinned memtrack installer from 1.3.0 to 1.4.0 and replaces its SHA-256 checksum so Linux memory-mode runs install the new release.

Confidence Score: 5/5

The PR appears safe to merge, with no actionable defects identified in the memtrack pin update.

The new version continues to use the existing release URL, checksum-verification, installer, and semantic-version compatibility paths, and the repository’s memtrack crate already declares version 1.4.0.

Important Files Changed

Filename	Overview
src/binary_pins.rs	The memtrack release version and installer checksum are updated together, with no changed interface or concrete compatibility defect identified.

_{Reviews (1): Last reviewed commit: "build: bump the memtrack pin to 1.4.0" | Re-trigger Greptile}

[codspeed-hq]

Merging this PR will not alter performance

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

✅ 17 untouched benchmarks

---

_{Comparing build/bump-memtrack-pin-to-1.4.0 (8d6fffb) with main (bdd5040)}