Meaning of Report Type in plain text output

[corneliusroemer]

I had a look at the plain text output of the parser as available here: https://ctt.pfstr.de/plaintext/2020-12-22.txt

I'm a bit perplexed by the "Report type" saying "Self diagnosis" and "Recursive". You didn't choose those names yourself, right? These are definitions made by CWA. Self report surely isn't a thing as of now, that must be a mistake somewhere.

Are the official names decoupled from how the transmission levels are actually determined?

Or is the problem that verification key version 1 is being parsed, but version 2 definitions are used?

Can you say a little about v1 vs v2? Are both available, the CWA now uses v2 but the parser is still using v1?

[mh-]

Hi @corneliusroemer
Please have a look at https://developers.google.com/android/exposure-notifications/exposure-key-file-format.
This is where the file format and

enum ReportType {
    UNKNOWN = 0;  // Never returned by the client API.
    CONFIRMED_TEST = 1;
    CONFIRMED_CLINICAL_DIAGNOSIS = 2;
    SELF_REPORT = 3;
    RECURSIVE = 4;  // Reserved for future use.
    REVOKED = 5;  // Used to revoke a key, never returned by client API.
  }

is defined.
This is not the Transmission Level, see

  // Varying risk associated with a key depending on diagnosis method
  optional int32 transmission_risk_level = 2 [deprecated = true];

for this.

There's not really a "verification key version 1", but the protobuf definition has evolved a bit. It's still somewhat backward compatible.
The v2 you refer to is supposed to use

  // Number of days elapsed between symptom onset and the TEK being used.
  // E.g. 2 means TEK is 2 days after onset of symptoms.
  optional sint32 days_since_onset_of_symptoms = 6;

instead of transmission_risk_level.

[corneliusroemer]

Understood, there is Report Type and TRL. Is CWA abusing the ReportType enum definition by Google or is @janpf parsing TEKs somehow wrong? Is this related to https://github.com/mh-/diagnosis-keys/issues/12?