cask/audit: also check binary for minimum OS

[EricFromCanada]

• Have you followed the guidelines in our Contributing document?
• Have you checked to ensure there aren't other open Pull Requests for the same change?
• Have you added an explanation of what your changes do and why you'd like us to include them?
• Have you written new tests for your changes? Here's an example.
• Have you successfully run brew style with your changes locally?
• Have you successfully run brew typecheck with your changes locally?
• Have you successfully run brew tests with your changes locally?

---

Adds logic to audit_min_os for checking the load commands in an application's binary for its minimum macOS requirement if not specified by LSMinimumSystemVersion of its Info.plist. It does this by fetching from Info.plist the binary path, ensuring it exists and is not a shell script (e.g. wings3d), then looks in each slice for LC_VERSION_MIN_MACOSX (older SDKs) or LC_BUILD_VERSION (newer SDKs) for their minimum macOS version value, and returns the lowest value found. This will surface a few more minimum OS suggestions for casks that don't define any, e.g. alacritty (multi-arch download) and contour (arch-specific downloads).

Since this audit checks more than Info.plist, changed plist_ references to bundle_.

Expect this audit to warn about:

• cases where a cask's stated requirements don't line up with what the audit detects because upstream hasn't made the necessary changes, e.g. calibre (site says latest requires 14.x, but it'll run on 13.x) or texstudio.
• casks that specify an older Intel build twice in order to make something available to newer Intel macOS versions (gpgfrontend: support newer Intel macOS homebrew-cask#215778, ghdl: support older ARM macOS homebrew-cask#215766, x-moto: support newer Intel macOS homebrew-cask#215752, stolendata-mpv: support newer Intel macOS homebrew-cask#215694)
• casks with arch-specific downloads where a) the Intel build's minimum OS is at least 11.x, b) the ARM build requires a different version, and c) it's not using on_os blocks for some reason.

But as this is a strict audit, these won't trip up CI.

[MikeMcQuaid]

Smart! Love the idea.