Re: libostree v2017.3
- From: Cosimo Cecchi <cosimo endlessm com>
- To: Robert McQueen <rob endlessm com>
- Cc: Richard Hughes <hughsient gmail com>, Colin Walters <walters verbum org>, ostree-list gnome org, Philip Withnall <withnall endlessm com>, Mario Sanchez Prada <mario endlessm com>, Joaquim Rocha <jrocha endlessm com>
- Subject: Re: libostree v2017.3
- Date: Fri, 17 Mar 2017 18:31:49 -0700
On Mon, Mar 13, 2017 at 8:32 AM, Robert McQueen <rob endlessm com> wrote:
On 11/03/17 09:44, Richard Hughes wrote:
Endless has a [custom update daemon](https://github.com/endlessm/eos-updater)
which...now that I look has certainly advanced a lot. They may have
ideas/requests/design input.
Hey Endless people! How do you show OS release notes to the user?
We don't at the moment. :( Typically our release notes are finalised in the days between tagging the
release in our staging repo but before it's pushed to our production repo, in parallel with the image and
static delta builds, etc.
So if we had to presume/assume that the release notes were ready and translated to be checked in to ostree,
we would probably be stuck. I know we do want to have the release notes shown to people before/after (?) an
upgrade, don't know if we'd got as far as the mechanism - not sure if that's planned yet. Cosimo might know.
Arriving a bit late here, and I haven't given this too much thought,
but in my mind there are two distinct use cases:
- showing a changelog paragraph in e.g. gnome-software before the
update is applied
- showing a tour of "what's new" (imagine something along the lines of
the GNOME release notes) with screenshots or videos, after the user
has updated to a new major release
I think that only the former would be covered by a mechanism built in
the commit history itself, but the latter is more interesting to the
end user.
It sounds like for the former case, what Philip proposed using
detached metadata would possibly work, but factoring in things like
translations and formatting, would we not end up with something that
looks very similar to the AppStream XML format that Richard mentioned?
At the end of the day, I feel we may get a bigger return focusing our
energy on solving the latter case at a higher layer of the stack (even
though solving that properly may require some addition to OSTree).
Cosimo
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]