Reduce memory footprint of module metadata Obj instances

[bcoles]

Description

Reduces per-instance memory overhead in Msf::Modules::Metadata::Obj when loading the ~6,600 modules from the JSON metadata cache. The metadata cache is the first thing loaded on startup, so every allocation here is multiplied across every module entry.

All changes are internal to init_from_hash and the class-level helpers. No public API changes.

Three techniques are applied:

1. String deduplication via String#-@ (fstring interning)

Fields like type, platform, arch, and author are highly duplicated across modules (e.g., only 7 unique type values across 6,600+ entries). String#-@ returns a frozen, deduplicated copy from Ruby's internal string table, so identical values share a single object.

2. Shared frozen empty containers

Many modules have empty aliases, references, notes, or targets. Previously each got its own [] or {} allocation. These now share EMPTY_ARRAY and EMPTY_HASH frozen constants.

3. PlatformList caching

PlatformList.transform was called once per module (~6,600 times) despite only ~100 unique platform strings existing. A class-level cache (cached_platform_list) returns the same PlatformList object for identical platform strings, moving the private parse_platform_list instance method to a shared class method.

[Copilot]

Pull request overview

This PR reduces memory overhead when instantiating Msf::Modules::Metadata::Obj entries from the on-disk JSON metadata cache, which is loaded early during framework startup and multiplied across thousands of modules.

Changes:

• Interns highly-duplicated strings (e.g., type/platform/arch/author entries) to reduce duplicate heap allocations.
• Reuses shared frozen empty array/hash constants for empty metadata fields.
• Adds a class-level cache to reuse parsed Msf::Module::PlatformList objects for repeated platform strings.

[Copilot]

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated no new comments.