Re: [4/4] dlls/ole32: Add more traces to the storage methods.

[Vincent Povirk]

On Thu, Oct 29, 2015 at 2:41 AM, Dmitry Timoshkov <dmitry(a)baikal.ru> wrote:
> static HRESULT TransactedSnapshotImpl_GetTransactionSig(StorageBaseImpl *base,
>    ULONG* result, BOOL refresh)
>  {
> +  FIXME("stub\n");
>    return E_NOTIMPL;
>  }

This is intentional. We don't need transaction signatures because
nothing else can have our snapshot open to commit to it. Anything
calling these methods should expect and handle E_NOTIMPL.

[4/4] dlls/ole32: Add more traces to the storage methods.

[Dmitry Timoshkov]

Vincent Povirk <madewokherd(a)gmail.com> wrote:
> > static HRESULT TransactedSnapshotImpl_GetTransactionSig(StorageBaseImpl *base,
> >    ULONG* result, BOOL refresh)
> >  {
> > +  FIXME("stub\n");
> >    return E_NOTIMPL;
> >  }
>
> This is intentional. We don't need transaction signatures because
> nothing else can have our snapshot open to commit to it. Anything
> calling these methods should expect and handle E_NOTIMPL.

Still printing a FIXME wouldn't hurt, right? Or maybe you have an
application that calls it thousands of times flooding the console
output?

-- 
Dmitry.

[4/4] dlls/ole32: Add more traces to the storage methods.

[Vincent Povirk]

> Still printing a FIXME wouldn't hurt, right? Or maybe you have an
> application that calls it thousands of times flooding the console
> output?

It would hurt. It would be misleading because there's nothing to fix.

[4/4] dlls/ole32: Add more traces to the storage methods.

[Dmitry Timoshkov]

Vincent Povirk <madewokherd(a)gmail.com> wrote:
> > Still printing a FIXME wouldn't hurt, right? Or maybe you have an
> > application that calls it thousands of times flooding the console
> > output?
>
> It would hurt. It would be misleading because there's nothing to fix.

Is there a test for this?

-- 
Dmitry.