I'm trying to visualize how a typical P2SH script gets executed, as I realize that it's a special form of script that gets executed slightly differently to a standard script.
Before P2SH added new validation rules, nodes would execute the script like this:
In other words, an old node would just be happy to check the the HASH160 of the redeem script matches the HASH160 in the scriptPubKey. No MULTISIG validation would need to take place, because OP_1 (TRUE) would be left on top of the stack.
With the new validation rules, I'm guessing the script executes something like this:
Does P2SH validation create a second stack using a copy of the redeem script, and then validate that stack before moving on to deserializing the redeem script (and then validating that script as normal)?
I'm hoping that my visualization of P2SH execution isn't misleading.