[Bug 275885] x11/lightdm: few small improvements

From: <bugzilla-noreply_at_freebsd.org>
Date: Fri, 22 Dec 2023 16:44:30 UTC
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=275885

Guido Falsi <madpilot@FreeBSD.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |madpilot@FreeBSD.org

--- Comment #1 from Guido Falsi <madpilot@FreeBSD.org> ---
Hi,

I like some of these ideas, but I'm not sure this is so easy to implement right
away. Some thoughts:

- As you state at
https://github.com/canonical/lightdm/pull/287#issuecomment-1867217795 we are
not completely sure the patch you import will work fine, it also has an error
in it, so we should not import it, but create a local patch without that error.

I'm not sure it will work, because I'm not sure the XDG variable4s will be
correctly populated at the point in time when this code runs.

I also have an issue with moving the .xsession-error file. I have no attachment
to it or its location; in fact I agree it is a poor solution to the problem. I
guess its history can be traced back to the origins of X11

But it is always been there and expected to be there. All other display
managers and most X startup scripts will put it in there, so changing only the
behaviour of lightdm will hardly fix the real issue.

We need to have a consistent behaviour of the while ports system.

An even better solution could be making this user configurable, not
autoconfigured by guessing based on XDG variables. But such a patch would be
more complicated.

We could also make it configurable at the port level maybe, changing the
hardcoded path based on some ports variable.

I'm not sure who we should discuss this with, if it should be done at a ports
infrastructure level.


- I have no strong objection to disabling qt5 and libaudit, but why you want to
do that?

I think it would be better to put those under some options, in case someone
needs them. I agree maybe this should have been done since the start of this
port.

-- 
You are receiving this mail because:
You are the assignee for the bug.