Automatic encoding detection vs. manual selection: memory-usage trade-off in Atom

Goal

Determine whether Atom should enable automatic character-encoding detection for unknown files, balancing the convenience of correct encoding against possible temporary memory spikes caused by re-parsing.

After the minimap memory-leak fix in Atom 1.58, residual RSS growth remains when opening and closing many tabs; enabling detection may add further short-term allocations during file load.

Does enabling automatic encoding detection produce a measurable increase in RSS compared to keeping it disabled when repeatedly opening and closing large files?

Is the short-term memory overhead from re-parsing acceptable for the benefit of accurate encoding detection on unknown files?