Account::LEN incorrect for 8-byte alignment

[tomjohn1028]

When using the @codama/renderers-vixen-parser package to generate account parsers, it generates a match with arms reliant on LEN of an account, where LEN is also generated by codama. In anchor account cases where data structures are not explicitly written proper 8-byte aligned values, the generated LEN is under that actual account length. This causes the try_unpack method to fail.

The std::mem::size_of::<MyAccount>() calculation on an account is correct. A potential solution could be to build those match arms with the std::mem::size_of::<MyAccount>(), but I have not thought through or tested implications of that change. I would like to see the use of discriminators in a flexible pattern, but I'll leave that for a separate issue.

[kespinola]

@tomjohn1028 do you have a minimal example you can share so we can reproduce? Is this you can override in the codama.mjs?