Cinnamon 2.0 released [LWN.net]

Cinnamon 2.0 released

Posted Oct 14, 2013 13:27 UTC (Mon) by torquay (guest, #92428)
In reply to: Cinnamon 2.0 released by ovitters
Parent article: Cinnamon 2.0 released

1. Currently Cinnamon forked things and doesn't do much maintenance on their fork. E.g. bugfixes made in gnome-shell are not merged in Cinnamon. 2. If you do fork, then there is an increased maintenance burden (merge of bugfixes, merge of changes in infrastructure). 3. If you fork, then there will be a bit of maintenance you have to do. It is not going to be free.

I agree that forking has downsides such as increased maintenance effort, which can be considered as a form of cost. It shouldn't be done without weighing up the pros and cons.

The point I was trying to make is that perhaps the overall cost of forking (resulting in control over the API) in this case is actually lower than the cost of working with a continuously shifting/unstable upstream API. The cost of coordination with other organisations / developers should not be assumed to be trivial, especially when development directions are not well aligned.

On top of that, the current reputation of Gnome developers (whether deserved or not) is that they're not exactly open to outside ideas. This dissuades people from submitting upstream. It would take time and effort to dispel this.

to post comments

Cinnamon 2.0 released

Posted Oct 14, 2013 17:01 UTC (Mon) by Rehdon (guest, #45440) [Link] (1 responses)

Also I think you're spot on with your remark about Gnome 2 being abandoned when 3 came out, it surely is a powerful caveat (although, to be fair, moving to a new paradigm makes sense ... throwing the baby with the dirty water less so :).

Rehdon

Cinnamon 2.0 released

Posted Oct 14, 2013 18:13 UTC (Mon) by ovitters (guest, #27950) [Link]

Didn't you read my reply? GNOME never supported anything longer than 6 months. Aside from this, a lot of work MATE is doing could (is?) have been done by taking the code developed after they forked it (they forked too early and too much).

Now for some interesting tidbit:
There is a conflict in the way MATE forked gnome-keyring into mate-keyring which makes d-bus start mate-keyring while the program was expecting gnome-keyring.

This bug was reported last week in Mageia. Bugs of course happen, but there are persons who often suggest that we went out of our way to not allow both to be used at the same time. However, we never supported that. And for the most part, we continued maintenance in the new release. It had gnome-shell instead of gnome-panel. The biggest technical changes were done partly in 2.x releases (moving to gsettings) and 3.8 (killing fallback mode). In 3.0 the major new thing was gtk+3.x. However, the preparation work was also done during 2.x.