RFC Errata
RFC 8620, "The JSON Meta Application Protocol (JMAP)", July 2019
Note: This RFC has been updated by RFC 9404
Source of RFC: jmap (art)
Errata ID: 6603
Status: Reported
Type: Technical
Publication Format(s) : TEXT
Reported By: Neil Jenkins
Date Reported: 2021-06-09
Section 5.6 says:
In the "upToId" request argument definition: If the sort and filter are both only on immutable properties, this allows the server to omit changes after this point in the results, which can significantly increase efficiency. If they are not immutable, this argument is ignored. In the "removed" response argument definition: If the sort and filter are both only on immutable properties and an "upToId" is supplied and exists in the results, any ids that were removed but have a higher index than "upToId" SHOULD be omitted. In the "added" response argument definition: If the sort and filter are both only on immutable properties and an "upToId" is supplied and exists in the results, any ids that were added but have a higher index than "upToId" SHOULD be omitted.
It should say:
In the upToId argument definition: The server may be able to omit added or removed items that are after the client's last cached id, making the update more efficient. In the "removed" response argument definition: If an "upToId" is supplied and existed in the old results, any ids that were removed but had a higher index than "upToId" in those results SHOULD be omitted. If the server cannot calculate this, the "upToId" MUST be ignored. In the "added" response argument definition: If an "upToId" is supplied and exists in the new results, any ids that were added but have a higher index than "upToId" SHOULD be omitted.
Notes:
This errata fixes two issues with the upToId definition:
1. Using upToId doesn't require immutable properties in some server implementations; this is an implementation detail. The important thing is it's an optional optimisation that the server can ignore if it does not have the data to calculate it. The text has been updated to reflect this.
2. Clarify that for the "removed" argument, the indexes we are comparing for the upToId optimisation are of the ids in the *old* results. The original text is unclear and seems to imply you might compare with the index of the id in the new results, which will give an incorrect result.