Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Tue, 28 Nov 2017 12:45:25 -0600

On 11/28/2017 11:42 AM, Michael Davidsaver wrote:
> The virtual dtors for RefMonitor and ChannelProviderRegistry
> are simple fixes which don't need further review.

Should ChannelProviderRegistry::~ChannelProviderRegistry() be public or
private? The ctor is private to prevent sub-classing so I suspect the
dtor should be too, but Dirk made it public.

> The change to epicsException.h also seems straightforward, though I'd
> add parenthesis after 'defined'.

Okay.

> I don't like removing FORCE_INLINE from castUnsafe().  This needs
> further analysis and a more target fix.
> 
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=54965

Awaiting your suggestions there then.

> As for the change to pva/client.h, it isn't clear to me from the diff
> just what is being changed (please use "diff -u"!).

The line numbers are a bit off against the master branch, probably
because of commit 250826ff which happened since his checkout. This is
the git diff against the current HEAD:

> diff --git a/src/client/pva/client.h b/src/client/pva/client.h
> index 08998b3..7721adb 100644
> --- a/src/client/pva/client.h
> +++ b/src/client/pva/client.h
> @@ -193,7 +193,11 @@ void registerRefTrack();
>   */
>  class epicsShareClass ClientChannel
>  {
> +#if defined __GNUC__ && __GNUC__*100+__GNUC_MINOR__ < 305
> +public:
> +#endif
>      struct Impl;
> +private:
>      std::tr1::shared_ptr<Impl> impl;
>      friend class ClientProvider;
>      friend void detail::registerRefTrack();

That looks a little strange, but could be a valid workaround to a gcc issue.

> I was expecting a change like this to pvif (in class FieldName), so
> I'm not clear what's going on here.

I don't see anything comparable in pvif.h.

- Andrew

-- 
Arguing for surveillance because you have nothing to hide is no
different than making the claim, "I don't care about freedom of
speech because I have nothing to say." -- Edward Snowdon

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System