Search This Blog

Loading...

Saturday, March 2, 2013

What is happening in QtNetworkManager (former libnm-qt)

Last year I wrote about some changes I would like to see happening in QtNetworkManager. One of the most important ones (Move the DBus parsing code to QtNetworkManager) is happening right now (just grab the code from settings branch from libnm-qt repository). Thanks Jan Grulich from RedHat for working on this issue.

Well, there are still other things to work on QtNetworkManager, just take a pick :-) If you are a student you can also take a look at my Google Summer of Code idea for improving the interface details panel in Plasma NM. Just write a proposal based on the idea and I will take a look it.

Update: I have just added a second idea for GSoC. It is about implementing SMS send/receive feature in Plasma NM.

7 comments:

Anonymous said...

I'm really interested in the UI redesign, more than anything else! The current plasmoid is rather complicated and frustrating to use, compared to something like the g-shell integrated NM front-end. Keep up the good work!

Honza Grulich said...

I think that all new changes in current networkmanagement will be pointless since we are working on a new networkmanagement with Lukas Tinkl. Our plan is to create a new connection editor and simple QML plasmoid and everything will be based on libnm-qt. So the only thing that should be improved is libmm-qt.

Lamarque Souza said...

@Anonymous, I am going to attack the UI problem again during the next solid sprint. The last time we did the redesign we used to use Plasma/QGraphicsWidget, which did make our life difficult to do UI changes. Then we started to the QML plasmoid and the redesign had to be postponed again until it was finished. Now that the foundations ofr QML plasmoid (mainly the Qt's models) are ready we can continue with the redesign. Unfortunately, some of the ideas we had for the redesign are not easy to implement in QML, so we are going to rethinkg some things during next sprint.

@Honza Grulich, what do you mean by a new networkmanagement?! If you plan to refactor networkamnagement you should talk to me first, don't you think? The connection editor requires a lot of changes to use libnm-qt. Improving the current QML plasmoid is in my plans.

Master branch of Plasma NM already uses libnm-qt for almost everything, the only exception was the create/update/delete part, which is mostly in the connection editor. All the commits I have pushed lately already uses libnm-qt or do not depend on libnm-qt, so I do not understand what is your point here.

Honza Grulich said...

@Lamarque We don't want to redesign the current networkmanagement, we want to write an entire networkmanagement from scratch, including the connection-editor and plasma applet. You know that it's not easy to fix some bug in the current version, so we decided to write something simpler. I hope we will have some basic functionality for the next solid sprint and we can discuss further steps.

I know that master branch uses libnm-qt, but it's still really complicated. Instead of creating your own d-bus interface we are planning to create a model with available connections (you can simple get the list of available connections for each device) and then simple QML plasmoid using this model.

Lamarque Souza said...

You know that "from scratch" means "from zero"? Currently there is a connection-editor branch in networkmanagement repo that reuses a lot of what already exists, that is not "from scratch".

What exactly do you mean by "creating you own d-bus interace"? There is a dbus interface in Plasma NM's kded module so the plasmoid cal talk to it. Is that d-bus interface you are talking about?

Anonymous said...

Hello! What about bridging support?

Honza Grulich said...

Bridging and bonding is also in our plan, because they are proposed as a new feature for fedora 19.