Re: [Xen-devel] [PATCH v2] RFC: Automatically check xen's public headers for C++ pitfalls.

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 12.03.15 at 11:03, <tim@xxxxxxx> wrote:
> At 13:16 +0100 on 05 Mar (1425557777), Tim Deegan wrote:
>> At 11:55 +0000 on 05 Mar (1425552955), Ian Campbell wrote:
>> > On Thu, 2015-03-05 at 11:41 +0000, Jan Beulich wrote:
>> > > >>> On 05.03.15 at 12:35, <ian.campbell@xxxxxxxxxx> wrote:
>> > > > On Thu, 2015-03-05 at 12:25 +0100, Tim Deegan wrote:
>> > > >> So I've seen four responses in favour of just renaming the field
>> > > >> (Andrew Cooper, Razvan Cojocaru, Don Slutz and David Vrabel) and one
>> > > >> in favour of #ifdeffing it so it's only renamed in C++ (Jan Beulich).
>> > > >> I really don't like adding more #ifdefs to an already hard-to-read
>> > > >> file; I'd rather just rename the field, or else leaving it alone and
>> > > >> letting C++ users carry the fixup in their own code.
>> > > >> 
>> > > >> CC'ing the other "THE REST" maintainers for their opinions.
>> > > > 
>> > > > Rather than ifdefs for C++, don't we need them based on
>> > > > __XEN_INTERFACE_VERSION__?
>> > > 
>> > > That's not applicable to the stuff under public/io/.
>> > 
>> > In which case I'd certainly prefer just changing the name and getting it
>> > over with.
>> > 
>> > mini-os would need checking, since that's (AFAIK) the only intree user
>> > of these headers.
>> 
>> mini-os doesn't in fact use this field; it's a blktap2-ism.
>> Here's a (non-RFC) patch to rename it and update blktap2:
> 
> Ian, Jan, can I get an Ack or Nack on this so I can clear it off my
> plate? :)

In fact I'm not enough against the change to NAK it, but I'm also not
really in favor of it, so wouldn't want to ack it. I.e. - Ian?

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel