Re: [Xen-devel] [PATCH] xentop: Adds options for tabs-separators, and including the domain ID in the output.

[Ian Jackson <ian.jackson@xxxxxxxxxxxxx>]

Wei Liu writes ("Re: [PATCH] xentop: Adds options for tabs-separators, and 
including the domain ID in the output."):
> On Sat, Aug 13, 2016 at 05:51:21PM +0000, Stefan Wieser wrote:
> > This change adds two options to xentop:
> > 
> >         -T      adds a tabulator (\t) character after each field,
> >                 to allow easier and more robust parsing. This
> >                 affects batch mode only.
...
> All in all I have no opinion for a few new options. I will wait a few
> gays so that other people can express their opinions.

In the absence of a better way of collecting this data in a
machine-readable fashion, I think something like this patch is a good
idea.

Stefan, I think you need to do something about the possibility that a
domain name could contain \t.  I suggest \-escaping the domname
(\ => \\; tab => \t; space left unchanged; bytes 0-31 and 127 encoded
with \xXX; assumption is that output may be UTF-8, which will be
passed unchanged if it contains no control chars).

I think this behaviour should be enabled by -T.

Also, I think you could profitably add something to the docs
explaining how to parse the output.  In particular, that a parser must
look at the column headings and not assume the columns will always be
the same columsn in the same order.

Ian.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel