Re: Orca autostart problem

2017-09-24 Thread Samuel Thibault
Samuel Thibault, on lun. 25 sept. 2017 01:37:19 +0200, wrote:
> Samuel Thibault, on dim. 17 sept. 2017 23:11:17 +0200, wrote:
> > Ok, it seems there is also incompatibility with lightdm.
> 
> Ah, no. The registryd daemon remains behind, but it does not actually
> hurt the Orca startup, at least on systems installed with the Debian
> Installer. I'll now check with live-installed systems.

It does work with the live-installed system.

So I really guess the only issue was that the documented dbus command
were typed as root instead of as normal user.

Samuel



Re: Orca autostart problem

2017-09-24 Thread Samuel Thibault
Samuel Thibault, on dim. 17 sept. 2017 23:11:17 +0200, wrote:
> Ok, it seems there is also incompatibility with lightdm.

Ah, no. The registryd daemon remains behind, but it does not actually
hurt the Orca startup, at least on systems installed with the Debian
Installer. I'll now check with live-installed systems.

Samuel



Re: Orca autostart problem

2017-09-24 Thread Samuel Thibault
Hello,

Luke Yelavich, on lun. 18 sept. 2017 07:44:06 +1000, wrote:
> On Mon, Sep 18, 2017 at 07:11:17AM AEST, Samuel Thibault wrote:
> > Ok, it seems there is also incompatibility with lightdm.
> 
> I wonder whether this linked bug is the culpret.
> 
> https://bugs.launchpad.net/bugs/1670933

Well, yes and no. Of course all these issues about starting at-spi
within lightdm are related, and the approach you propose should be
solving everything, but here Egon is talking about Debian stable, which
AIUI doesn't have a systemd service file for the registry, so the
approach wouldn't work.

But patches/01_at-spi.patch in stable should be handling this, by just
always starting at-spi from lightdm itself, allowing it to shut it down
properly before the user session. It does work when installing with
brltty or speakup enabled, I don't know yet why it doesn't otherwise.

Samuel