grufwub has quit [Remote host closed the connection]
grufwub has joined #wayback
Guest59 has joined #wayback
Guest59 has quit [Client Quit]
pinskia has quit [Read error: Connection reset by peer]
pinskia has joined #wayback
_whitelogger has joined #wayback
pinskia has quit [Read error: Connection reset by peer]
pinskia has joined #wayback
mojo_ has joined #wayback
<f_>
I would like to avoid having to depend on logind
<f_>
But I guess it's fine to have some support for it, Xorg also has some integration if it's installed
norwoodites has joined #wayback
pinskia has quit [Ping timeout: 240 seconds]
pinskia has joined #wayback
norwoodites has quit [Ping timeout: 248 seconds]
__0x1eaf has joined #wayback
<Ariadne>
to be absolutely clear: i consider hard depending on logind to be unacceptable
<Ariadne>
some optional support is fine, but an SUID helper that does the vt switch would be the way to go here for cases where libseat cannot do it
<f_>
Yeah Xorg doesn't hard-depend on logind so there's no reason wayback should
<f_>
Even if it's elogind
<axtlos>
I agree
<axtlos>
but all that is hypothetical until we figure out how to switch the vt wlroots renders on
<axtlos>
I'll have to look into that more tomorrow
<f_>
23:38 <dramforever[m]1> mccd: no hurry but did you try xwayland-satellite? i'm curious to see if it works for you. it should be non-blurry... i think < I suspect they have a HiDPI display and have 2x+/fractional scaling in their compositor
<f_>
"Irrespective of everything else about this, their web page starts with a box that is styled to resemble WindowMaker, and so it is automatically an excellent project."
<f_>
:> glad they love it
<axtlos>
huh what's lobste.rs
<axtlos>
looks like a reddit frontend
<axtlos>
though either way I'm happy that people like it
nemoto has quit [Remote host closed the connection]
<axtlos>
honestly I was worried that wayback was gonna get some "redhat is getting rid of our xorg now!!!!!!" reaction
<dok>
i am personally intrested in wayback so i run classic Xorg desktkops on SoCs which i think have a better wayland support than plain Xorg
vkareh has joined #wayback
wink has joined #wayback
<mojo_>
same, we develop a apps on SoCs, X is deprecated and it is only a weston with downstream patches, its hard to explain to my colleagues, that thei are wrong if they want to position their application
<f_>
axtlos: lobsters is a link aggregrator
<f_>
Similar to HackerNewa
<f_>
*HackerNews
<f_>
Not a reddit frontend
<f_>
(And I still hate mobile keyboards)
<navi>
note, iirc, that Xorg *also* fails the vt switch if it doesn't have perms and logind support is disabled
<navi>
so there's no need for the suid helper, at least not to match behaviour with Xorg
<hiya>
Is Ariadne Conill the founder of wayback?
<hiya>
Is she an alpine linux developer?
__0x1eaf has quit [Quit: Lost terminal]
arraybolt3 has quit [Quit: Konversation terminated!]
arraybolt3 has joined #wayback
<f_>
hiya: yes and yes
<f_>
navi: oh haha
<hiya>
Thank you
<dramforever[m]1>
<f_> Wayback does not handle dpi at all
<dramforever[m]1>
i genuinely don't know what dark magic hidpi xwayland is but it (mostly) works for me...
<dramforever[m]1>
what kde calls "Apply scaling themselves"
<f_>
hiya yw
<hiya>
f_: needed this info for bulletin :)
<hiya>
First for our site
<hiya>
issue 0
<f_>
dramforever[m]1......idk it doesn't do anything related to DPI at all for the moment
<f_>
hiya: nice
<bigmac>
hiya: sweet, share the link after it's done
<axtlos>
dramforever[m]1: it means that the apps do the scaling, not the compositor
<axtlos>
xwayland has nothing to do with it
<bigmac>
By the way, I see that there's an issue board. However, do we track anywhere on which issues is being worked on and which are free to be picked
<axtlos>
no we don't, though rn the general rule is that if there is no open pr and it wasn't mentioned here no one's working on it
mojo_ has quit [Ping timeout: 252 seconds]
<Consolatis[m]1>
> xwayland has nothing to do with it
<Consolatis[m]1>
well, afair it scales the internal x11 display to surface size * fractional scale factor and then uses the viewporter protocol to tell the compositor that its buffer is already scaled to the correct dimensions. resulting in the compositor not doing any scaling on top.
<Consolatis[m]1>
no clue if (and how) xwayland exposes some "please use this DPI scale" kind of setting to its X11 clients though, otherwise they would just end up being very small
<ConanKudo[m]1>
they do wind up being very small generally I think
<ConanKudo[m]1>
that's why KDE has that warning in system settings
<bigmac>
f_: Have you played with #49 more afterwards? You reproduced it by launching wayback-session on a new tty -> launching cage w/ foot inside of a xwm and then when something new is launched from the xwm (not within cage) it opens up in cage?
<dramforever[m]1>
<bigmac> f_: Have you played with #49 more afterwards? You reproduced it by launching wayback-session on a new tty -> launching cage w/ foot inside of a xwm and then when something new is launched from the xwm (not within cage) it opens up in cage?
<dramforever[m]1>
wayland clients are hard-coded to try wayland-0 if WAYLAND_DISPLAY is not set
phantop has quit [Client Quit]
<Consolatis[m]1>
bad ones, yes
<dramforever[m]1>
if by bad ones you mean anything that calls wl_display_connect(NULL);
<ConanKudo[m]1>
yeah it's not "bad ones"
<ConanKudo[m]1>
it's the expected failover check
<ConanKudo[m]1>
things that don't do that are doing it wrong
<dramforever[m]1>
this is documented, intended behavior of wl_display_connect()
<ConanKudo[m]1>
yup
<ConanKudo[m]1>
if there's a Wayland client that doesn't do that, it's broken
<Consolatis[m]1>
applications are supposed to use wl_display_connect(NULL) unconditionally, even if XDG_SESSION_TYPE is not set to wayland and no WAYLAND_DISPLAY is set?
<dramforever[m]1>
i think yes
<Consolatis[m]1>
hm.. well, then let me retract that "bad" statement
<dramforever[m]1>
in the future where we also ship 12to11 or something for x11 sessions
<dramforever[m]1>
12to11 should run on wayland-0
<Consolatis[m]1>
feels very wrong to me though. I think some compositors actually started to not use wayland-0 to work around such apps but I don't quite remember the details
<dramforever[m]1>
yes, i think cage should not run on wayland-0 unless it's the top compositor
<dramforever[m]1>
uh i mean it should not use wayland-0 on the x11 or wayland backend
<ConanKudo[m]1>
yeah there's a few cases where you don't want this, particularly if you are nesting compositors
caskd has quit [Ping timeout: 260 seconds]
caskd has joined #wayback
simendsjo has joined #wayback
pinskia has quit [Read error: Connection reset by peer]
<f_>
bigmac: I think we should close that issue given it's not wayback-specific