Hm, with sample provided in #2267 (comment) parsing fail at closest

is.checkPos(entriesStart, "Expected first entry start");

because of additional 2 bytes, maybe it is padding, so it might be better to change this line to

is.skipToPos(entriesStart, "Failed to skip to entries start");

And, as soon as, sample also use compact entries, it still can't be parsed, so I am not sure if this is a correct solution.
To be honest, it is better to merge this PR with compact entries support, because this is the only way to verify its correctness.

This PR only adds support for 16-bit entry offsets. Not support for compact entries, not support for non-standard paddings, only 16-bit entry offsets.

But if you insist on implementing all together, then you can close the PR. I don't think I'll create a PR with compact entries support, because my current implementation of the resource parser is too different from yours.

I see.
I will try to implement compact entries support myself and will merge this PR once I managed to parse provided sample.