Drop the dead integral clause in the packer after #1716 and #1712 land

[sbryngelson]

#1716 removes integral output entirely (integral_wrt, num_integrals, and the integral(i)%... region parameters). #1712 adds a packer branch in toolchain/mfc/packer/pack.py that keeps every column of probe and integral output:

elif "probe" in short_filepath or "integral" in short_filepath:
    # Probe and integral output are multi-column time series, not spatial
    # fields: each column is a distinct physical quantity (probe: rho, vel,
    # pres, gamma, pi_inf, qv, c, accel, ...; integral: int_pres, max_pres)
    # and the set varies by configuration. Treating them as <x> <y> <z>
    # <value> would keep only the last column and silently drop the rest, so
    # every column is retained. Neither writes a header line to skip.
    doubles = _extract_doubles(content)

Once both PRs are merged, nothing writes integral*.dat any more, so:

• the or "integral" in short_filepath clause is dead — the match can never fire
• the comment's integral: int_pres, max_pres parenthetical describes parameters that no longer exist
• Probe and integral output are / Neither writes a header line should become singular

This is harmless if left as-is: the branch simply never matches. It is purely a readability issue — the code reads as though it supports a feature that was removed, which is exactly the kind of stale reference that misleads someone grepping for integral later.

Fix: reduce the condition to elif "probe" in short_filepath: and rewrite the comment to describe probe output only.

The two PRs do not conflict and can merge in either order; this cleanup applies once whichever lands second is in. Also noted in #1712 (comment).