Question on recommended practice with FMA

[leerho]

>
> we can't let issues like this (e.g. overly-pedantic checks triggered by a
> certain IDE) drive
> API design.
>

I absolutely agree.  I suspect that what the Eclipse compiler is doing is
that it detects an AutoCloseable object that is not inside a TWR statement
and not closed at the end of the method so it triggers the warning.

I guess I should try to target 17 instead of 16, as this issue is impacting
many places in our code.

Lee.

On Wed, Apr 14, 2021 at 2:04 PM Maurizio Cimadamore <
maurizio.cimadamore at oracle.com> wrote:

>
> On 14/04/2021 21:48, Maurizio Cimadamore wrote:
> > As Remi suggested, this seems to be an issue in the Eclipse compiler
> > being overly pedantic - we can't help, sorry.
>
> And, I should have added, as I'm sure we understand, we can't let issues
> like this (e.g. overly-pedantic checks triggered by a certain IDE) drive
> API design.
>
> So, I suggest we stick to the real issue here - which is the semantics
> of MemorySegment::close when slices are involved (which has been
> rectified in the latest API).
>
> Maurizio
>
>