Re: [Xen-devel] [PATCH RFC V2 05/10] libxl_json: introduce parser functions for builtin types

[Ian Campbell <Ian.Campbell@xxxxxxxxxx>]

On Wed, 2014-04-23 at 12:14 +0100, Wei Liu wrote:
> On Wed, Apr 23, 2014 at 11:42:42AM +0100, Wei Liu wrote:
> > On Wed, Apr 23, 2014 at 11:31:19AM +0100, Ian Campbell wrote:
> > > On Wed, 2014-04-23 at 11:19 +0100, Wei Liu wrote:
> > > > On Wed, Apr 23, 2014 at 11:12:54AM +0100, Ian Campbell wrote:
> > > > > On Wed, 2014-04-23 at 11:01 +0100, Wei Liu wrote:
> > > > > 
> > > > > > The aformentioned "calloc" falls into #1 but not this one.
> > > > > > 
> > > > > > My comment is a bit confusing. The comment here actually refers to 
> > > > > > all
> > > > > > the malloc'ed memory in the loop, which will be freed by
> > > > > > libxl_cpuid_dispose.
> > > > > 
> > > > > You mean when the caller eventually gets rid of the result, or the 
> > > > > error
> > > > > path here does it?
> > > > > 
> > > > 
> > > > The former. Caller will regardlessly call dispose_fn of a type, wouldn't
> > > > it?
> > > 
> > > Hrm, I'm not sure what we expect the caller to do with the output of a
> > > failed operation. I think I would expect that the failing operation
> > > would clean up any partial result. Ian?
> > > 
> > > It might be worth having a poke around and seeing what other libxl
> > > functions do under these circumstances.
> > > 
> > 
> > The way I see it is that in current libxl code, the caller , regardless
> 
> Sorry I meant "xl code".

It does look that way doesn't it. Lets go for consistency for sure!

Any chance you could add a comment to a relevant section of libxl.h to
describe this expectation (in the general case, not just for this
interface)?

Ian.



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