Re: [Tails-project] Report from GUADEC

このメッセージを削除

このメッセージに返信
著者: intrigeri
日付:  
To: Public mailing list about the Tails project
題目: Re: [Tails-project] Report from GUADEC
Hi,

Alan:
> Here is a report from GUADEC, the GNOME Conference.


Thanks a lot for this very detailed report. It was a pleasant read :)

I've watched all the talks you've recommended plus a few other ones,
and I've loved it to the point I'm regretting I wasn't there in
person. Reading Didier Roche's blog posts about how they're migrating
Ubuntu to GNOME Shell, and learning more about Endless, made me
realize that we may not be as involved in GNOME as we could (and maybe
"should"), compared e.g. to our two other major upstreams i.e.
Debian and Tor. I'd like to attend GUADEC next year, ideally with you
and one of our UX people.

> The GNOME community puts a lot of effort on Flatpak[1] their application
> distribution technology, so there were a lot of talks and discussions
> are about it.


> Flatpak framework run applications in sandboxes, and have portals to
> communicate with the system. It uses cgroups and namespaces with bubblewrap.


I've looked quite a bit into it recently and with BitingBird we've
initiated a number of interesting conversations on this topic during
DebConf17.

We've discussed the general "sandboxing GUI apps" topic at the Tails
summit and it felt obvious that technology such as Flatpak/Portals is
*the* long-term way to address the shortcomings of our current
sandboxing model. Even better, the IBus vs. Flatpak problem was fixed
a few days ago, and there's good progress on the a11n vs.
sandboxing front. Now, we felt it was too early to put much energy
into it, so it wasn't added to our roadmap for 2018-2019 (Wayland made
it there though :) I'm convinced this should be one of our major
priorities for 2020-2021 though. Until then I'll keep playing with it
and discussing this topic within Debian.

> In "Resurrecting dinosaurs, what can possibly go wrong",
> Richard Brown was discussing the limits of Flatpak, especially in terms
> of security support and embedded libraries. I found it healthy to have
> this questions asked.


Absolutely. I recommend it to any Flatpak fan-person, such as myself.

> ## Let GNOME Disks upstream know we intend to add TrueCrypt support


> (https://labs.riseup.net/code/issues/12275)


> I discussed with GNOME Disks maintainer Kaï Luke. They are interested by
> the feature. There are already support for opening/closing TrueCrypt
> volumes in libblockdev:


>     https://github.com/storaged-project/libblockdev/issues/200
>     https://github.com/storaged-project/libblockdev/issues/240


> The missing part is UDisks:


>     https://github.com/storaged-project/udisks/issues/282


> He offered to work on the GNOME Disk part and can help getting patches
> accepted elsewhere.


Amazing, thanks a lot for bringing this topic there!

I'll forward this info to segfault and anonym (I *think* I have already,
but I'm not sure anymore).

> ## Consider replacing Florence with GNOME's own on-screen keyboard


> (https://labs.riseup.net/code/issues/8281#note-26)


> I discussed the state of on-screen keyboard with Shell maintainer Carlos
> Garnacho. Their current plan is to drop caribou entirely and to enter
> characters instead of key codes. They are interested to work with us on
> the layout issue.


>     https://bugzilla.gnome.org/show_bug.cgi?id=660368
>     https://bugzilla.gnome.org/show_bug.cgi?id=785677


> There are repositories of layouts at
> <http://www.unicode.org/repos/cldr/tags/latest/keyboards/android/>. I've
> worked on adapting an import script from caribou source, while Carlos
> will use these instead of Caribou models in GNOME Shell on-screen keyboard.


Amazing! Was there progress since the hackfest? Is it something you
plan to keep working on, and if yes, what's the timeline you have
in mind?

> After we've worked on that, someone proposed to use onboard instead of
> caribou on the GNOME bugzilla... so stay tuned.


Interesting. I've been using a computer with a touchscreen¹ on a daily
basis since a couple weeks; it's not able to play videos in GNOME
Shell due to lack of DRI module for Mesa, so I'm running GNOME
flashback (boooh, that's soooo painful, especially on a touchscreen,
compared to pristine GNOME!)… and the best on-screen keyboard I've
found was precisely onboard. The funny thing is that we used to ship
onboard in Tails before we switched to Florence, for reasons
I can't recall.

¹ https://tails.boum.org/blueprint/ARM_platforms/Acer_Chromebook_R_13_CB5-312T/

> One student finishing their GSOC was interested to work on the parts of
> GNOME that interests Tails. I discussed with them our list of bugs on
> the GNOME bug tracker, as well as our contribute page.


Any follow-up on this? I'd be happy to reach out to this person again,
with my Foundations Team hat on (I see "welcoming new code
contributors" on our mission description). But if you want to keep
being the interface between them and Tails, fine by me :)

Many thanks again!

Cheers,
--
intrigeri