Re: GNOME Online Accounts 3.34 won't have documents support

[Bastien Nocera <hadess hadess net>]

On Wed, 2019-01-23 at 20:10 +0000, Debarshi Ray wrote:
> On Wed, Jan 23, 2019 at 08:02:16PM +0100, Bastien Nocera wrote:
> > On Wed, 2019-01-23 at 18:52 +0000, Debarshi Ray wrote:
> > > Grep for "Future of Pocket in GNOME" from 24th August 2018 in
> > > your
> > > inbox.
> >
> > Which solves what? You're removing the Documents and Mail
> > categories,
> > you're removing Pocket support.
>
> Two things.
>
> First, we are talking about GNOME Documents, GNOME Online Accounts,
> GNOME ..., etc.. We are talking about GNOME.
>
> If you are not going to respond to a thread that has run for six
> months, on a topic that you care about, then I am sorry, I can't
> assume good faith. Those who participated agreed that they don't have
> time to work on Pocket, nor is the current state of affairs very
> good. That's how we decided to drop it.

Having to constantly repeat that, yes, this was of interest to me was
soul-sapping.

> > Reading that mail you mentioned won't bring those features back.
> > It's
> > far from the first time you've wanted to remove Pocket support from
> > GOA, as if it was a time sink, and a maintenance pain.
> >
> > I just don't understand the strategy of disabling/removing features
> > and
> > services, breaking apps on newer hosts.
>
> We have always, since the very first days when David Zeuthen was
> around, pushed back against adding random accounts to GOA. We have
> always said that the integration should be meaningful to a good cross
> section of users, that there should be a default application or OS
> component, etc..
>
> This is nothing new. We eventually wrote it down as
> https://wiki.gnome.org/Projects/GnomeOnlineAccounts/Goals
>
> The Pocket integration doesn't do much. You yourself said that you
> seldom use it.

Mostly because the various integration points didn't exist. We (GNOME)
were pining for a sharing interface which didn't, and still doesn't,
exist, and epiphany maintainers didn't want to integrate the patch that
was already written.

> We regularly refuse requests to add random accounts to GOA. Just look
> at the enhancement requests left open on Bugzilla. There are lots
> more
> that I have dealt with from other channels. It isn't fair that we
> turn
> down other requests, but keep Pocket in.

I don't think that feature requests should be treated on the same level
as existing, merged, features, and I don't see any merge requests for
new account types, or providers.

> It certainly is a maintenance burden. It's a burden when one has to
> port away from deprecated GLib and WebKit APIs, it's a burden
> when someone has to tweak the base-classes to accommodate yet
> another quirky service provider or to just repay some technical debt
> and clean things up.

I wish that was brought forward earlier in the thread. Seeing as it's a
maintenance burden, and it's impossible to write an application with
some reasonable degree of expectation that Pocket support will be
present in the host gnome-online-accounts framework, I think it best
for it to be removed completely, so that applications don't try to use
it in lieu of in-application independent support:
https://gitlab.gnome.org/GNOME/gnome-online-accounts/merge_requests/18

> But it's a burden we can carry as long as we have someone committed
> to
> push it forward towards a better future. That future need not be in
> two months time, but it has to be something that's more realistic
> than
> unicorns.

Sorry, I don't understand what unicorns have to do with this
discussion.