RE: stdin/stderr/stdout/stdoutStderr: fileLocation or physicalLocation

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

Michael agrees. I'll file the issue and push a change draft.

This one is in the gray area between "bug -- obviously subject to editorial
discretion" and "substantive change". I suggest we put this one to the TC
for approval.

Larry

-----Original Message-----
From: James A. Kupsch <kupsch@cs.wisc.edu> 
Sent: Friday, March 23, 2018 7:40 AM
To: Larry Golding (Comcast) <larrygolding@comcast.net>;
sarif@lists.oasis-open.org
Cc: 'Michael Fanning' <Michael.Fanning@microsoft.com>
Subject: Re: stdin/stderr/stdout/stdoutStderr: fileLocation or
physicalLocation

I think these could be fileLocations, and I would not object to the change.
There are scenarios where having a region might be useful, but I do not
think they would occur in practice.  For instance, process 1 writes stdout
to a log file, then process 2 appends output to the same log file.

Jim


On 03/22/2018 07:18 PM, Larry Golding (Comcast) wrote:
> The spec defines each of these properties as a physicalLocations  = { 
> fileLocation + region }.
> 
> James, remind me please why it?s not just a fileLocation. Is there a 
> scenario where one of these streams occupies just a portion of a file? 
> Do some tools embed the contents of these streams in their native log 
> file format?
> 
> Larry
>