Yes, the step you're missing is "and build the table". Dynamic memory allocation is something you want to avoid, as well as any artifical restrictions to number of inputs or outputs. Current solution is slow, but there's really no limitation on tx size.

Plus there're significant restrictions to memory in embedded world. Actually TREZOR uses pretty powerful (and expensive) MCU just because it needs to do such validations and calculate such hashes. With SIGHASH_WITHINPUTVALUE or similar we may cut hardware cost significantly.

Marek

On Fri, Jan 23, 2015 at 5:52 PM, Gregory Maxwell <gmaxwell@gmail.com> wrote:

I'm not sure where the ^2 is coming from. So what I'd understand that
you'd do is stream in the input txid:vouts which you spend, then you'd
stream the actual inputs which would just be hashed and value
extracted (but no other verification), and you'd build a table of
txid:vout->value, then the actual transaction to be signed.

This should have O(inputs) hashing and communications overhead. Is
there a step I'm missing?