← Back to team overview

kicad-developers team mailing list archive

Re: V6 merge priority


On 11/03/2019 14:41, Wayne Stambaugh wrote:
> On 3/11/2019 9:17 AM, Tomasz Wlostowski wrote:
>> On 11/03/2019 13:57, Wayne Stambaugh wrote:
>>> Hi Tom,
>>> On 3/9/2019 7:59 PM, Tomasz Wlostowski wrote:
>>>> On 09/03/2019 19:03, Wayne Stambaugh wrote:
>>>>> Rather than try to figure out every possible merge combination, I'm
>>>>> going to prioritize things serially per editor.  The highest priority
>>>>> are changes those that will or potentially impact later features.
>>>> Hi Wayne,
>>>> If I may add my 5 cents - let's merge the stuff affecting the schematic
>>>> and board model first so that any refactor later on would not make
>>>> rebases of the already written features a nightmare. On my side, I would
>>>> like to:
>>>> - replace DLISTs<> with Rtrees (BOARD::MarkTrace takes ages on larger
>>>> boards)
>>>> - remove remaining references to PCB_EDIT_FRAME (or anything directly
>>>> GUI-related) from the tool and model code in pcbnew. This way, the tools
>>>> will be fully portable to any toolkit we might choose in the future and
>>>> much easier to test.
>>> Are these changes ready to go and will you be readily available should
>>> things go sideways?  How much testing did you do on the rtree changes?
>>> Replacing DLIST has the potential to be disruptive.  I agree that low
>>> level changes like this should happen before we start merging the other
>>> new features.
>> Hi Wayne,
>> I don't have any patches yet. It's mostly the legacy code that depends
>> on the DLIST<> interface, everything newer relies already on iterators
>> (DLIST_ITERATOR_WRAPPER<>). Should we remove the legacy code from pcbnew
>> sometime early during V6 cycle?
> Doing so would definitely expose anything missing in the new tool
> framework ;)  It will certainly be easier to maintain pcbnew without the
> duplicate tool interfaces.  I think it's time to finally pull the plug
> on the legacy tool code in pcbnew.

OK. Expect some patches soon.