Change draft for #149 (nested graphs)

["Larry Golding \(Comcast\)" <larrygolding@comcast.net>]

I pushed a change draft for Issue #149, “Support nested graphs”:

 

                Documents/ChangeDrafts/Active/sarif-v2.0-issue-149-nested-graphs.docx

 

I will move its adoption at TC #17 on May 16^th.

 

This draft describes Luke’s design, which we’ve discussed at length in email. It removes my previous design, which defined a concept of “nested graph traversals”. That concept is now gone.

 

In addition to Luke’s design, this change draft proposes a mechanism that makes it easy for SARIF viewers to offer a “step over/step into” experience when a graph traversals spans multiple graphs. The one drawback of Luke’s design was that it made it hard for a viewer to know when a graph traversal would return to the graph it started from. This proposal addresses that problem:

3.31.3                [edgeTraversalOverTargetNodeId property

If the edgeId property (§3.31.2) of this edgeTraversal object refers to an edge object (§3.29) whose targetGraphId (§3.29.6) refers to a graph other than the current graph (see §3.30.1), and if the containing graphTraversal ultimately returns to the current graph, then the edgeTraversal object MAY contain a property named stepOverTargetNodeId whose value is a string that equals the id property (§3.28.2) of the node object (§3.28) in the current graph to which the graph traversal first returns.

A SARIF viewer MAY use this property to offer the user a “step over/step into” experience. That is, the viewer offers the user the choice of either stepping into this edge’s target graph or jumping directly to the point of return from the target graph.

For an example, see §3.29.6.

 

That’s some pretty dense spec-ese there. I recommend you look at the example to see how it works.

 

Thanks,

Larry