Request for classic confinement: mate-wayland


#1

I’m trying to publish mate-wayland. It is a demo of the work I’ve done porting the MATE desktop to Wayland with MIr. The snap is meant to function as a full desktop environment, so it needs to be able to run all the apps installed on the system, not just the software in the snap. For this, it requires classic confinement.


#2

@Wimpress ping. Also, this is currently broken but should get you started: https://github.com/wmww/sway-snap


#3

See Can't run snap commands in Konsole on Plasma snap for my problems running apps even on the devmode Plasma snap, have you managed to get it working on devmode on MATE?! :slight_smile:


#4

@Ads20000 as you can read in the comments on your post, a confined snap has limited access to the system. A classic snap has wider access to the system and that’s why classic confinement is being requested here. (devmode is not equivalent to classic confinement.)

You can see the same difference with egmde (a classic snap) compared with egmde-confined-desktop (a confined snap).


#5

I realise I am slightly biased on this topic, but I’m :+1: for granting classic for this snap. It is not currently possible to launch arbitrary applications when a snap is strictly confined, and this is a central feature of a desktop shell and the mate-panel in particular.

Having spoken with @wmww, @alan_g and @jamesh we are all in agreement that this snap will transition to classic confinement as snapd offers interfaces/facilities to support the launching of arbitrary applications.

Can I request the other @reviewers take a look at this and share their thoughts?


#6

@reviewers this request has been up for almost three weeks with one +1, could someone else take a look and expedite a decision?


#7

+1 to classic for mate-wayland since it needs to launch arbitrary programs.


#8

The requirements are understood (please understand that classic does not use a voting process. Instead requirements are gathered, suitability for the store is assessed and the publisher is vetted, but reviewers are free to express their opinions.

@Wimpress - you mentioned your bias, but is the publisher vetted?


#9

I work for Canonical (William Wold on the Mir team). This is a Canonical project.


#10

I don’t think that working for Canonical should ever get a project or person a free pass. There are processes and standards in place for a reason, and bypassing them for favouritism leads to rightful accusations of double standards. People already like to moan about things, so let’s not give them a valid reason, please.


#11

@daniel I absolutely agree! The reason I mentioned it was because @jdstrand asked if the publisher (me) was “vetted”. I assumed my employment with canonical would be an indicator that I’m unlikely to intentionally push malware. My apologies if there is a standard publisher vetting process I’m unaware of (@Wimpress didn’t mention any such process when we talked in Malta).


#12

Don’t worry about it. :slight_smile: The process will probably be a case of @Wimpress or @popeydc making a note that yes you are definitely employed by Canonical and thus are vetted, but we need to be sure that all the I’s are dotted and all the T’s a crossed to ensure that there’s a log of things that can be referred to in case someone moans :slight_smile:


#13

I vetted the publisher. Granting use of classic. This is now live.


#14

@wmww - you may either request a manual review or upload a new revision, after which it should pass automated review.