@michaelcfanning and
I discussed this offline. I reminded him that the earliest invocation.startTimeUtc does
not necessarily equal the run's start time (for reasons I explained in an earlier comment). He replied by reminding me of the driving
scenario, which is:
... to identify the time of detection for a result. It is helpful for this value to be consistent, i.e., there should be a single timestamp for an analysis run. From this angle:
find the earliest timestamp from the array of invocations seems fine as a simple, reliable way to get what’s required. I understand conceptually that doing this might not precisely describe the actual
start of a run, but do we care?The real point here is: ‘this results has been raised since build XXX on date YYY’. We’re providing a slot for a date as a convenience, so that users aren’t required to go look up information associated with the
instance id.
(Emphasis added)
I find that convincing, and I'm going to revise the change draft accordingly. We don't need to restore run.startTimeUtc (or run.endTimeUtc,
which I had restored just for symmetry).
Larry
From: Michael Fanning <>
Sent: Monday, December 10, 2018 9:30 AM
To: Larry Golding (Myriad Consulting Inc) <>; OASIS SARIF TC Discussion List <>
Subject: RE: Change draft for #287: default for lastDetectionTimeUtc
Hi, Larry,
I am not sure resurrecting run.startTimeUtc makes sense, please see my most recent reply.
I led us to this situation because I forgot that we removed this property. All things considered, introducing it again seems like it brings more complexity without much incremental value.
Michael
From: <>
On Behalf Of Larry Golding (Myriad Consulting Inc)
Sent: Sunday, December 9, 2018 1:56 PM
To: OASIS SARIF TC Discussion List <>
Subject: [sarif] Change draft for #287: default for lastDetectionTimeUtc
I pushed a change draft for
Issue #287: “Define default for resultProvenance.lastDetectionTimeUtc”.
Documents/ChangeDrafts/Active/sarif-v2.0-issue-287-lastDetectionTimeUtc-default.docx
In the issue’s discussion thread, Michael and I decided that it made sense to bring back the property
run.startTimeUtc. Please read the thread to see how that idea is related to this issue. So we’ve restored that property, and for symmetry, also restored
run.endTimeUtc.
We’ll move its adoption at TC #29 on Wednesday December 12th.
Thanks,
Larry
