this makes sense, really: close_others() requests UI exclusivity, and
that is something that generally happens at the same places we emit a
ButtonRequest
- CLSAG signature scheme added
- type hints added
xmr: optimize protocol, send only required data
- real_out_additional_tx_keys contains only one element as nothing more is needed during signature
- only src_entr.outputs[index] is HMACed and always present. Other outputs are present only if needed which reduces comm and CPU overhead.
- getting rid of subaddresses dictionary (memory requirements), now subaddr indices are present per source entry so keys are computed when needed
xmr: prepare for permutation sending removal, specify index
- specify source entry ordering index prior sorting by key images as original HMAC keys are generated based on these.
- permutation checked just by valid HMACs, size of the set, key image sort order
- sending permutation is now deprecated, will be removed in the following protocol versions
- more strict state transition checks, guard strict check with respect to steps ordering
Apart from making the code more correct for its users in
apps.common.confirm and elsewhere, this fixes a problem where the
confirm_signal would be scheduled before the dialog is rendered.
By making sure that handle_rendering is scheduled (i.e., listed in
create_tasks) before confirm_signal, we can be sure to render at least
once and thus appear in the UI test results.
Based on the testing feedback I've made slight modifications to the UI.
Main points:
- running on the full brightness during the process (previously dimmed, felt broken)
- overlaying layout fix after transaction confirmation
