On 03/14/2017 02:30 PM, John Ferlan wrote:
On 03/11/2017 08:16 PM, Laine Stump wrote:
> On 03/10/2017 04:10 PM, John Ferlan wrote:
>> Rather than use virXPathString, pass along an virXPathNode and alter
>> the parsing to use virXMLPropString.
> Just so I understand the reasoning correctly - you're not doing this so you can
use virXMLPropString() instead of virXPathString(), but just so you can remove the
"adapter" from the path to each attribute (and in the cases where that turns a
"path" into simply the attribute name, you're switching to
virXMLPropString() because it's presumably slightly more efficient. Right? Or is there
some other reason you prefer virXMLPropString()?
>
Missed this on the first pass through your review... Probably because
your email client has stopped believing in line wrapping or my client
isn't seeing some setting properly <sigh>
Really? Oh %&*()%$*)%$...
Late last week I encountered messages from someone which, when replied to, led to all of
*their* quoted paragraphs in the reply each being a single very long line. So I asked
about it on IRC, and Dan pointed me to some webpage claiming to know how to
"fix" Thunderbird so that it worked reasonably with git-generated emails. I made
the suggested changes, it didn't help, and then *I thought* I changed everything back.
Apparently not :-(
I had originally wanted the ability to just parse an <adapter...> since
that was how this would be represented in the domain; however, even
though that got mothballed - I still felt what I'd done so far was at
least a step up in readability, flow, etc. over what was there before.
Okay, so even more reason than what I'd assumed, i.e. more than enough :-)