Inclusion of KDE code in WS

[Roland Knall <rknall () gmail com>]

Hi

Short question:
Would it be ok, to include 3 KDE classes (from git://
anongit.kde.org/kitemmodels.git) with the WS sourcecode, after asking the
author (single author in this case) if this is ok? The code is licensed
with GPL 2.1

Long question:
I am running into a rather difficult issue with the view/model concept. As
the new interface tree makes excessive use of the proxymodels (as it should
do), it makes sense to share selections between views. We do this right now
using signal/slots, which leads to an overhead in the programmcode, but
also an overhead in time, as each signal/slot is run in a different thread.

The solution for this is rather simple in theory, selectionmodels (which
handle selection in a model) can be shared. This is done on quite a few
occasions in the code already. Now, with proxy models sharing the
selectionmodels use different indeces, as their underlying datamodel
differs. But all those data models have the same parent (although after
quite a few models in between). In the above mentioned source, Stephen
Kelly wrote a KModelIndexProxyMapper which is portable, independent and
GPL-licensed, which achieves exactly what is needed here.

We could implement our own version (heavily written off the KDE version but
mentioning and referring to it), or just copy their code, as it is quite
good.

regards,
Roland

___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <wireshark-dev () wireshark org>
Archives:    https://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev
             mailto:wireshark-dev-request () wireshark org?subject=unsubscribe