kicad-developers team mailing list archive

[Seth Hillbrand <seth@xxxxxxxxxxxxx>]

On 2019-08-13 10:50, Wayne Stambaugh wrote:
> On 8/13/19 2:13 AM, jp charras wrote:
> > Le 12/08/2019 à 21:54, Wayne Stambaugh a écrit :
> > > Sounds like a plan.  If there are not bug reports against this over 
> > > the
> > > next month, I'll will cherry-pick it into 5.1.
> > >
> > > Wayne
> >
> > I am not sure this issue exists in 5.1.
> >
> > AFAIK it comes from moving code from our DLIST to std::deque, only in
> > master branch.
> >
> > Our DLIST dtor manages (when it has the ownership) the items deletion.
> > But std::deque does not delete the items, so the code using std::deque
> > has to delete these items.
>
> Maybe we should have used boost::ptr_deque instead.  You get heap
> allocation cleanup for free.

We can adjust between containers now if desired as long as they obey the 
std::random_access_iterator constraints.  The major work in moving from 
DLIST was removing all the places where we assumed the element was 
itself a list.

For now, I prefer keeping the single list cleanup in the board dtor 
until we have a standardized python interface that doesn't expose our 
internals.  After we abstract the python interface, I see no reason why 
we wouldn't utilize std::unique_ptr as Simon suggested.

-Seth