[nas] (resend) hard hang in libaoss.so

[Erik Auerswald]

Hi,

On 07/04/2013 03:34 PM, Paul Fox wrote:
> erik wrote:
>   > On Thu, Jul 04, 2013 at 08:44:49AM -0400, Paul Fox wrote:
>   > > erik wrote:
>   > >  > On Wed, Jul 03, 2013 at 04:09:30PM -0400, Paul Fox wrote:
[...]
>   > >  > Does the following patch work as well?
>   > >
>   > > i'm sure it would.  but i was actually going to suggest a different,
>   > > second patch.  i was going to suggest removing the sco ifdefs entirely.
>   > >
>   > > thoughts?
>   >
>   > IMHO the sco #ifdefs should be removed. They complicate reading the code,
>   > encumbering maintenance (as in this case).
>   >
>   > I still think we should first address the issue at hand (reducing the race
>   > window), perhaps even making a maintenance release with this fix.
>   >
>   > Afterwards we should remove the sco #ifdefs in one go from the
>   > voxware server.
>
> agreed on all of the above.  another alternative to complete removal
> would be to consolidate the ifdefs into far fewer chunks, to remove
> clutter from the linux path.  but if it can't be tested (and i doubt
> it can), then it should just go away.

I agree that code changes need to be tested.

The problem I see with consolidating the SCO #ifdefs is that mistakes
will slip in even if only "obviously correct" changes are made. Without
actually testing the SCO code it will break (might be broken already).

Unless someone with an SCO machine to test NAS on steps up, we cannot
test that code and should thus remove it -- because it has just shown to
be a maintainability problem.

Anyway, I don't see a need to rush this at all.

Erik