Option 2 has the advantage that a schema validator can catch the mistake if you omit it.
From: <> On Behalf Of Larry Golding (Comcast)
Sent: Wednesday, June 6, 2018 3:52 PM
To:
Subject: [sarif] When columnKind is required
Importance: High
We agreed this morning that there was no default for columnKind. But it doesn’t make sense to require it for tools that don’t analyze text files.
I see two options:
Say that “If the analysis tool analyzes text files, the run object SHALL contain … columnKind … If the analysis tool does not analyze text files, columnKind SHALL be absent.
Defined another enumerated value "none" for columnKind, and tell analysis tools that don’t analyze text files to use it.
Thoughts?
I’m going to merge the change as we approved it, but I think we have to fix this.
Larry
