Reject PSBT inputs that carry no utxo data

[ruipereira1]

Note: this touches the same region of psbt_parser.py as #965, so it is based on
that branch and includes its commit. Only the second commit
("Reject PSBT inputs that carry no utxo data") belongs to this PR — please review #965
first. Happy to rebase once that one lands, or to fold the two into a single
"PSBT parser hardening" PR if you'd prefer fewer moving parts.

Description

Problem or Issue being addressed

PSBTParser._parse_inputs() assigns script_pubkey from whichever utxo field an input
carries:

for inp in self.psbt.inputs:
    if inp.witness_utxo:
        self.input_amount += inp.witness_utxo.value
        script_pubkey = inp.witness_utxo.script_pubkey
    elif inp.non_witness_utxo:
        self.input_amount += inp.utxo.value
        script_pubkey = inp.script_pubkey

    inp_policy = PSBTParser._get_policy(inp, script_pubkey, self.psbt.xpubs)

An input with neither field leaves the local variable bound to the previous iteration's
value, so that input's policy is derived from a different input's script. On the first
input it raises UnboundLocalError instead — confirmed by the added test, which strips
witness_utxo/non_witness_utxo from a single-input PSBT.

Solution

An explicit else that raises, rather than reading a stale value.

This is defensive rather than a fix for an observed failure: psbt.fee() later in
_parse_outputs() would most likely raise first on such a PSBT. But silently reusing the
last loop value does not belong in the code that decides whether a transaction has mixed
inputs.

Additional Information

The error message is short because Controller.handle_exception renders it on the error
screen.

Screenshots

No new or modified screens.

---

This pull request is categorized as a:

• New feature
• Bug fix
• Code refactor
• Documentation
• Other

---

Checklist

I ran pytest locally

• All tests passed before submitting the PR
• I couldn't run the tests
• N/A

187 passed with this change. 4 tests in test_seedqr.py that decode QR images could not run in my environment because libzbar does not load on Windows; they fail identically on an unmodified dev checkout (baseline: 184 passed, same 4 failures) and are unrelated to this change. They pass in CI, which installs libzbar0.

---

I included screenshots of any new or modified screens

• Yes
• No
• N/A

---

I added or updated tests

• Yes
• No, I'm a fool
• N/A

---

I tested this PR hands-on on the following platform(s):

• Raspberry Pi OS Manual Build
• SeedSigner OS on a Pi0/Pi0W board
• Emulator

Not tested on hardware or the emulator — I don't have a device. Covered by a unit test. Happy for a maintainer to verify on-device.

---

I have reviewed these notes:

• I understand