VAR arrays not propagated correctly for MIRI MRS

[stscijgbot-jp]

Issue JP-3909 was created on JIRA by David Law:

A user reported issues with the VAR extensions in MIRI MRS data that I've run back to two sources:

1. The fringe correction step is applying the fringe flat to the SCI and ERR arrays, but not to VAR_POISSON, VAR_FLAT, or VAR_RNOISE.

2. The time-dependent sensitivity correction in the photom step is being applied to the SCI and ERR arrays, but not to the VAR arrays.

This isn't critical since only the ERR values are passed through cube building, but we should fix these steps to ensure consistency in the VAR arrays.