Not that it’s probably anything to do with it, but Gerbview is not free anymore (I think). Maybe try another viewer or re-import the gerbers into EPC and see if that creates @n error ?
By all means try an alternative Gerber Viewer, but do NOT rely on EasyPC Gerber importer in this instance as this is like a schoolboy marking his own homework!
As an example, EasyPC can make "rounding errors" in the Gerber plots when less than 5 decimal place resolution is specified in the Gerber Output plot settings. One Forum member ended up with a batch of pcbs where the flood area covered an isolated pad, -despite EasyPC importer showing the intended flood area correctly. Many (All?) other Gerber importers showed up this rounding error.
There are many free Gerber viewers out there, I have use Pentalogix "Viewmate" for many many years, and there is the free definitive Ucamco online viewer on https://gerber.ucamco.com Hope you resolve your query.
PS. G81 is Excellon FMAT1 for "turn on drill mode".
"One Forum member ended up with a batch of pcbs where the flood area covered an isolated pad, -despite EasyPC importer showing the intended flood area correctly. Many (All?) other Gerber importers showed up this rounding error."
However this is only ok if you can visually "see" the short ..... very very difficult if not impossible if your looking at a very complex board. If it happens on a 2 layer board its bad enough but sometimes easy to fix with a scalpel ... but on a 4 layer and above it can spell disaster. I therefore check this in a completely different way. I export the netlist and gerbers from EPC. I then import the gerbers into FAB3000. FAB3000 allows me to generate a netlist from the actual gerbers. This netlist and the netlist from EPC are the compared and automatically report any discrepancies. This is the only safe way of establishing if the gerbers are correct in every way.