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

[philip chimento gmail com]

On Mon, Jan 28, 2019, 00:00 Debarshi Ray, <rishi is lostca se> wrote:

On Sun, Jan 27, 2019 at 09:26:05PM -0800, philip chimento gmail com wrote:
> On Sun, Jan 27, 2019 at 1:04 PM Debarshi Ray <rishi is lostca se> wrote:
>
> > On Sun, Jan 27, 2019 at 11:24:00AM -0800, philip chimento gmail com wrote:
> > > 2. It's not possible to discontinue support for services X, Y, and Z from
> > > GOA, and yank the rug out from under apps that expected (even if that
> > > expectation was wrong) it to be part of a stable platform.
> >
> > You mean like the time when GJS broke GNOME Documents?
> >
>
> Sorry, I don't know what particular event you're referring to.

https://bugzilla.gnome.org/show_bug.cgi?id=791613

> GJS is even different from GOA because, as pointed out earlier
> in the thread, GOA is not part of a Flatpak runtime but is still used by
> Flatpak apps.

One, Flatpak is still not the primary way users consume applications.

Second, being part of the runtime is irrelevant for a session D-Bus
service like GOA. The service is meant to live on the host. There were
no D-Bus or C API breaks. What changed is the metadata associated with
an account, which applications introspect at runtime anyway.

Third, with Flatpak, a newer version of an application has to be ready
to run on an old host anyway, and the old host might not have the GOA
feature that your new application wants. Apps need to be resilient
about that. Hence the introspection.

> Are you seriously suggesting that because I committed some mistake, you
> should insist on the right to make the same mistake with eyes wide open?

And what mistake am I insisting on making?

> Literally just a few messages before mine in the thread, we heard about
> deja-dup.

And what about Deja Dup? GOA screwed over Deja Dup?

> For what it's worth, I don't appreciate the ad-hominem attack here. I
> intended to *help* you by trying to break the cycle of people ignoring your
> problem and yelling at you about theirs, and vice versa, but if you prefer
> to continue yelling then suit yourself.

You are trying to help me by making unsubstantiated innuendoes?

You said:

> PS. Yes, count me among the completely surprised that GOA is not an API
> that apps should use. It was not communicated anywhere close to the level
> it needed to be. That's on GNOME, not on those app developers. This is why
> it's our problem.

If nothing else, you are driving a wedge between "GNOME" and "those
app developers".

In the context of this thread, it also sounds as if there exists a
whole bunch of application developers whose hard work was thrown away
because of me.

You talk about communication. The very first version of the GOA home
page, created in August 2012, said:

"The objective of GOA is to provide a central place where a set of
online accounts (defined by the OS/distributor) can be configured
for use with core applications. In UX terms, GOA provides a static
list of online accounts that can be setup by users (through the
Online Accounts panel in System Settings). These accounts will then
be used by core GNOME applications.

While third party applications can access the accounts set up
through GOA, this is not its primary goal, nor does it set out to
enable third party applications to add online accounts of their own.
There are several reasons for this:
  ...
  ...
"

See:
https://wiki.gnome.org/action/recall/Projects/GnomeOnlineAccounts?action="">

That text was recently improved and lives on its own page, linked
multiple times from the main landing page:
https://wiki.gnome.org/Projects/GnomeOnlineAccounts/Goals

The current landing page says:
"GNOME Online Accounts Single sign-on framework for GNOME. It aims to
provide a way for users to setup online accounts to be used by the
core system and applications."

See: https://wiki.gnome.org/Projects/GnomeOnlineAccounts

We don't have an active PR team at our disposal to send out weekly
press releases - it's not like the GNOME Engagement Team has a ton of
resources. Blogs were written to address the most pressing issues of
the day.

There doesn't exist a single application developer whose work was
thrown away without any prior notice or discussion.

The Telepathy support was removed after multiple Empathy and Telepathy
maintainers publicly and repeatedly asked people not to use the
stack. The removal was publicly announced early in a development
cycle. Not a single soul complained when this happened.

Michael Catanzaro, Allan, Jakub and I have had multiple discussions,
often over email, with Felipe Borges, author of GNOME's Pocket client
[1], and Bastien over the years. The most recent discussion was about
toggling the Autotools and Meson options to not build the Pocket
extension by default, accompanied with a public announcement. That's
supposed to happen after March. No code was to removed until October.

Michael Catanzaro, Allan, Jakub and I had similarly discussed the
GNOME Documents situation with Cosimo.

You are tilting at windmills here. As if, I, the GOA maintainer,
screwed myself, the GNOME Documents maintainer. That's how bizarre
this situation is.

[1] http://felipeborges.github.io/bolso/

Nope, nope, nope. I'm not going to continue this discussion. Good suggestion from Britt to take a break. I thought I could help but no. Good luck.