Re: New features tests fails
On Tue, Aug 20, 2024 at 11:28:05 +0200, Anthony Harivel wrote:
Hi Peter,
> These test cases parse an XML and format it back.
>
> You've got a bug in the code formatting the socket into the XML:
>
> if (def->kvm_features->socket == NULL) {
> virBufferAsprintf(&childBuf, "
socket='%s'/>\n",
> def->kvm_features->socket);
> } else {
> virBufferAddLit(&childBuf, "/>\n");
> }
>
>
> The above hunk formats the socket only if it's NULL, and skips it if
> it's non-null (inverted condition).
>
Doh! I think it's time for me to take some PTO...
Thank you !
However, I tried my best to follow your advice using
virBufferEscapeString() so that i can get rid of the condition on the
socket but it fails for "feature-off".
But this is passing the test suite:
I meant like this:
if (def->kvm_features->features[j] != VIR_TRISTATE_SWITCH_ABSENT) {
virBufferAsprintf(&childBuf, "<%s state='%s'",
virDomainKVMTypeToString(j),
virTristateSwitchTypeToString(def->kvm_features->features[j]));
virBufferEscapeString(&childBuf, " socket='%s'",
def->kvm_features->socket);
virBufferAddLit(&childBuf, "/>\n");
}
(note also the fixed alinment)
Note that the parser should also disallow the socket path if the feature
is explicitly off.