[LDR] Is SEH still needed here? A potential optimization

In ancient times learn_more committed 0.4.10-dev-136-gc7eb46d to fix issues with the mpress packer (which improved CORE-12582 and CORE-14556 without fixing them entirely). He did fix the regression CORE-12400 with that same commit (without being aware of it back then).

Unfortunately back then this very same commit broke "Antivir 6 deadlocks RtlpWaitForCriticalSection when "scanning the process list" CORE-14838 .
He then fixed that follow-up-issue by 0.4.10-dev-474-gdfff8ed (by catching the undesired exceptions)

learn_more and me discussed later, whether that SEH'ing maybe is superfluous today, because we later fixed with 0.4.15-dev-5589-g0d4a8d0 other regressions of the initial commit 0.4.10-dev-136-gc7eb46d which did cause the first exceptions in there. Up to now (2023-11-26) we are not aware of any other regressions than the mentioned ones having been introduced by that.
We had this discussion here, so here is the follow-up-ticket to do that exact investigation.

The expected Task is basically to just revert 0.4.10-dev-474-gdfff8ed and check whether that would break anything or not.
The CORE-14838 scenario can serve as a first test-case to check for the intended optimization being free of side-effects.