[libvirt] [PATCHv2 1/6] virsh-domain: Refactor cmdVcpucount and fix output on inactive domains

[Peter Krempa]

On 04/19/13 01:27, Eric Blake wrote:
> On 04/16/2013 08:00 AM, Peter Krempa wrote:
>> This patch factors out the vCPU count retrieval including fallback means
>> into vshCPUCountCollect() and removes the duplicated code to retrieve
>> individual counts.
>>
>
>> +#define PRINT_COUNT(VAR, MAX, STATE) if (VAR > 0) \
>> +    vshPrint(ctl, "%-12s %-12s %3d\n", _(MAX), _(STATE), VAR)
>> +        PRINT_COUNT(conf_max, "maximum", "config");
>> +        PRINT_COUNT(live_max, "maximum", "live");
>> +        PRINT_COUNT(conf_cur, "current", "config");
>> +        PRINT_COUNT(live_cur, "current", "live");
>> +#undef PRINT_COUNT
>
> Doesn't quite work with i18n.  Since MAX and STATE are not string
> literals, they do not appear in the .po file.  But since "maximum" isn't
> marked, it also doesn't appear.  So nothing gets translated.
>
> You either need: _(MAX) in the macro, and N_("maximum") as argument to
> the macro, or plain MAX in the macro, and _("maximum") as argument to
> the macro.
>
> Also, using the name MAX as a macro parameter risks collisions with a
> common macro name of MAX(); it might be better to name the macro
> parameter something less likely to collide, like WHICH.
>
> Beyond that, it looks like your conversion does the right things.  I
> assume you tested on all three of transient, persistent online, and
> persistent offline guests, and with a good set of combinations of
> arguments to make sure we aren't breaking back-compat for which numbers
> get reported.
>
> ACK with the macro fixed up.
>

Thanks, I fixed the macro and double checked that the original flag 
meanings match the stuff reported by old virsh and pushed this patch.

Peter