[noise] Test vector format

[Trevor Perrin]

On Sun, Apr 10, 2016 at 8:12 PM, Alex <alex at centromere.net> wrote:
> On Sun, 10 Apr 2016 20:04:36 -0700
> Trevor Perrin <trevp at trevp.net> wrote:
>
>> On Sun, Apr 10, 2016 at 6:14 PM, Alex <alex at centromere.net> wrote:
>> > On Sat, 9 Apr 2016 12:40:46 -0700
>> > Trevor Perrin <trevp at trevp.net> wrote:
>> >
>> > Is it safe to say that there can be only one failure per vector?
>> >
>> > Is it also safe to say that, if there is a failure, it must be the
>> > last message in the "messages" array?
>>
>> Those both sound reasonable.
>>
>
> If that's the case, then putting a "fail: true" key (which defaults to
> false if not present) within the vector object (not the individual
> message object) seems reasonable to me. This could be helpful, for
> example, if you're using vim to search through the file and want to
> find which tests fail quickly. You can do a regex search for "fail",
> but it wouldn't be so easy if you assume that "pattern: null" is a
> failure.

Yeah, that sounds fine: so if "fail: true" you omit the "payload" from
the message object, and processing the ciphertext is expected to
produce an error.

Trevor