On 7/8/22 13:23, Daniel P. Berrangé wrote:
> On Fri, Jul 08, 2022 at 12:49:13PM +0200, Michal Privoznik wrote:
>> Historically, the dumpxml command reject any unknown arguments,
>> for instance:
>>
>> virsh dumpxml fedora xxx
>>
>> However, after v8.5.0-rc1~31 the second argument ('xxx') is
>> treated as an XPath, but it's not that clearly visible.
>> Therefore, require the --xpath switch, like this:
>>
>> virsh dumpxml fedora --xpath xxx
>>
>> Yes, this breaks already released virsh, but I think we can argue
>> that the pool of users of this particular function is very small.
>> The sooner we do this change, the better.
>>
>> Resolves:
https://bugzilla.redhat.com/show_bug.cgi?id=2103524
>> Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com>
>> ---
>> tools/virsh-domain.c | 1 +
>> 1 file changed, 1 insertion(+)
>>
>> diff --git a/tools/virsh-domain.c b/tools/virsh-domain.c
>> index 43034f2f81..a181ae005f 100644
>> --- a/tools/virsh-domain.c
>> +++ b/tools/virsh-domain.c
>> @@ -10461,6 +10461,7 @@ static const vshCmdOptDef opts_dumpxml[] = {
>> },
>> {.name = "xpath",
>> .type = VSH_OT_STRING,
>> + .flags = VSH_OFLAG_REQ_OPT,
>> .completer = virshCompleteEmpty,
>> .help = N_("xpath expression to filter the XML document")
>> },
>
> Opps, shouldn't we do the same change for all the other 'xpath'
> parameters.
>
I guess we should. But I worry that we might break some users scripts.
But I guess if you're okay with it then so am I.
It was documented as --xpath in the man page. That you could give it
positionally was an oversight of mine, thanks to our crazy homegrown
CLI parser behaviour. Did I mention I wished we just used getopt...
With regards,
Daniel
--
|: