Re: im-config | Use im-config on Wayland without uninstalling IBus (!5)
Hi,
Let me comment to clarify.
On Fri, 2020-07-10 at 22:13 +0200, Gunnar Hjalmarsson wrote:
...
> On 2020-07-10 05:23, Osamu Aoki wrote:
...
> > Gunnar,
> >
> > Please help me understand situation assuming this MR is accepted.
> >
> > == Testing/Sid system ==
> >
> > * Are you sure just waiting this fix to migrate is enough to fix
> > this
> > Korean input problem?
>
> Well, I don't speak Korean and don't know how to really use ibus-
> hangul.
> But, as mentioned at the MR, I tested the proposed changes
> successfully,
> which means that I can no longer reproduce the problem as described
> in
> the upstream mutter issue 948.
>
> This is the mutter build I used for testing:
>
> https://launchpad.net/~gunnarhj/+archive/ubuntu/mutter/+packages
>
> Possibly you can install from that PPA in Debian and test yourself
> (no
> guarantee, but the Debian/Ubuntu delta seems to be small and
> unrelated).
>
> > (If so, I will wait without touching im-config on testing/sid)
>
> What's in im-config's devel branch right now would be nice to see
> uploaded to unstable as a first step IMO. It does not make the
> ibus-hangul/mutter bug worse, and it would make im-config more easy
> to
> use on Wayland for non-IBus frameworks.
>
> Then, together with the pending mutter fix, we would improve the
> situation significantly.
Thanks.
> > == Stable system ==
> >
> > * Does this fix have chance to be included in the next stable
> > update?
> > (I doubt this and we should not count on this)
> >
> > * Does this fix have chance to be included in the backport?
> > (I think this may happen but this is no excuse for not fixing
> > this
> > bad situation)
>
> Assuming those questions were not meant for me. I'm not familiar
> enough
> with the procedures for stable updates in Debian. With that said, it
> would be highly desirable to get the fix in, of course.
Actually, I meant this to you. You effectively answered anyway.
> > * What exactly is the concern to add workaround approach proposed
> > by
> > Changwoo to the im-config. Breaking virtual keyboard is non-
> > issue
> > since this is useless for Korean input at this moment. Since
> > "auto"
> > is the place where most desirable choice is recommended, I see
> > no
> > reason to reject Changwoo's proposal.
>
> The most important reason IMO is the way that change would affect the
> UI
> for explicitly setting the desired IM framework. The change would
> affect
> all IM users,
Let's work on solution which doesn't do such thing.
...
> If you know that
>
> * GNOME launches and configures IBus by itself when im-config is
> disabled ("none"),
>
> * enabling im-config by setting "ibus" explicitly changes the
> configuration so IBus works differently behind the scenes compared
> with
> how GNOME is designed to make use of IBus, and
I don't think we want "ibus" selection to do something else behund
anyone. I want "auto" to select "ibus" or "none"
Osamu
Reply to: