Re: [PATCH v2 16/19] Refactoring virDomainGraphicsDefParseXMLSpice() to use XPath
On Thu, Mar 24, 2022 at 07:16:28PM +0100, Michal Prívozník wrote:
On 3/24/22 18:02, Andrea Bolognani wrote:
> but I can't help
> wondering if there are other cases where switching to
> virXPathString() in the way seen here might have introduced undesired
> changes in behavior.
>
> Thoughts?
Unless you want to audit every virXPath*() call with its history, I say
we're good as we didn't get any bug reports. It is unfortunate, but the
master was broken only for one day.
I'm not concerned about the brief period of time between Kristina's
and Tim's patches were merged, but about any subtle changes in
behavior that might still be lingering even after a year.
That said, I have just finished auditing all uses of
virXMLPropTristate*() and I don't think I can find the strength in me
to go through the same process again for virXPath*() - at least not
right away.
So let's just hope that the lack of bug reports is an accurate
indicator of the lack of significant regressions, I guess :)
--
Andrea Bolognani / Red Hat / Virtualization