caskd has quit [Remote host closed the connection]
caskd has joined #wayback
caskd has quit [Remote host closed the connection]
caskd has joined #wayback
caskd has quit [Remote host closed the connection]
caskd has joined #wayback
<f_>
is joborun on IRC
<dramforever[m]1>
hmm, sounds like that should be repro'able mostly easily
<dramforever[m]1>
... if i knew what consolekit2 is...
<navi>
logind1 before systemd, it's a login/seat/session manager
<navi>
currently developed by freebsd folks iirc
<dramforever[m]1>
ah, and consolekit2 is a fork that kept that alive
<dramforever[m]1>
i wonder how startx usually works with consolekit
<dramforever[m]1>
there's an ominous comment here on arch wiki
<dramforever[m]1>
> If you are not using a display manager and ConsoleKit is not working (i.e. ck-list-sessions command showing active = FALSE), you should start your window manager using the bash_profile method: Xinit#Autostart X at login.
<f_>
and I did this while being half-asleep, so that's why it's Draft.
mccd has joined #wayback
caskd has joined #wayback
<navi>
i'll review in a moment
<bigmac>
Hey. I am attempting to reproduce https://gitlab.freedesktop.org/wayback/wayback/-/issues/50. However, so far I had little to no success doing so. Could anyone try reproducing it (I suppose just switching ttys and checking for the specific error will suffice). My meson configuration: https://paste.centos.org/view/4cb875dc I am not seeing the mentioned error with WAYLAND_DEBUG=1 while switching
<bigmac>
ttys from and to the one defined with the -d flag
<dramforever[m]1>
switching terminal is suspicious
<dramforever[m]1>
maybe i'm overreading it but is "mini" possibly a mac mini
<f_>
bigmac: I can reproduce it just fine on my laptop
<f_>
with alpine package wayback-0_git20250714-r0
caskd has quit [Ping timeout: 276 seconds]
<bigmac>
f_: Mhm. And could I ask you to grep for "set_destination", are there any instances where the destination would be 0, 0 ?
<f_>
oh I don't have any logs right now
<f_>
I will try to get some logs later
caskd has joined #wayback
__0x1eaf has joined #wayback
caskd has quit [Read error: Connection reset by peer]
caskd has joined #wayback
caskd has quit [Ping timeout: 240 seconds]
caskd has joined #wayback
<dramforever[m]1>
f_: i wonder if it's possible your laptop is doing one of those disconnect reconnect things on resolution change, and just reinitializing the output
<thomas_adam>
f_: Is this ""getopt"" implementation worth carrying around?
cow has quit [Ping timeout: 265 seconds]
<axtlos>
yes absolutely, it allows us to easily specify ignored options, and we were using a custom argument parser either way, this just makes it more convenient
<thomas_adam>
axtlos: I can see that from the diff.
<thomas_adam>
There's other ways though.
<axtlos>
there are a lot of ways to do things
<axtlos>
if you think a different approach is better suited feel free to send patches ;)
<spicywolf>
is there some reason not to use std getopt and just make cases to ignore some?
<axtlos>
because Xorg has weird arguments that also are `+nya` instead of all being `-opt`
<ConanKudo[m]1>
because it predates GNU getopt
<ConanKudo[m]1>
which standardized the convention for enable/disable and long opts
<ConanKudo[m]1>
that is not an interface we can change
<thomas_adam>
axtlos: I shall.
<thomas_adam>
I think it might set a weird order otherwise. wayback isn't special in this regard.
<navi>
X options are not flags
<navi>
they're all options by themselves, basically matched directly
<navi>
the annoying thing of blocking or ignoring xwayland flags, is that it becomes a bit of a race, every time Xwayland add a new flag
nemoto has joined #wayback
<axtlos>
i mean i dont think we should be blocking them, there can be valid reasons why users might want to use some argument
mccd has quit [Ping timeout: 240 seconds]
<navi>
a warning might be enough maybe? "Xwayland-only flag dected, you're on your own"?
<navi>
though there's a few flags we probably should ignore, `-rootless`, and the ones we use internally
<axtlos>
yeah
<axtlos>
yep, i retracted my review for that lol
grufwub has quit [Ping timeout: 240 seconds]
<axtlos>
but yes a warning is probably the way to go
mccd has joined #wayback
<axtlos>
thats what i commonly see in other projects
<spicywolf>
axtlos: oh yeah I forgot. lol
<dramforever[m]1>
wondering if "unrecognized flag, passing to ..." makes sense
<axtlos>
nah i dont think we need that, the point is to replace Xorg, warning for unrecognized options when getting an X option would be confusing
<axtlos>
probably going to reopen it, it didnt follow the issue template
<dramforever[m]1>
fwiw wayback is in the pipelines to being shipped to nixpkgs, which should go to users in a few days depending on build server load and backlog https://github.com/NixOS/nixpkgs/pull/426859
<dramforever[m]1>
i'll probably add a mr to the readme when it definitively lands
<axtlos>
oh awesome
<axtlos>
reminds me that i wanna make a void and gentoo package
__0x1eaf has quit [Quit: Lost terminal]
<f_>
19:19 <thomas_adam> f_: Is this ""getopt"" implementation worth carrying around? < yeah definitely. Right now only Xwayback uses it but eventually maybe wayback-session will make use of it too
<f_>
19:29 <thomas_adam> There's other ways though. < tell me?
<f_>
19:32 <spicywolf> is there some reason not to use std getopt and just make cases to ignore some? < Xorg does arguments (e.g. "+xinerama") getopt does not support, and we need to maintain compatibility
<f_>
20:05 <navi> the annoying thing of blocking or ignoring xwayland flags, is that it becomes a bit of a race, every time Xwayland add a new flag. < eh. Some flags are useful, some others are meh for Wayback purposes
<f_>
20:07 <navi> a warning might be enough maybe? "Xwayland-only flag dected, you're on your own"? < it does warn when ignoring args
<f_>
20:09 <dramforever[m]1> wondering if "unrecognized flag, passing to ..." makes sense < nah, no need, there's a lot of args Wayback will just pass to Xwl
<f_>
So having that warning would be weird
<f_>
axtlos: I'm not in front of my computer, but RE "Any reason this doesn't use wayback_log?" This is just because the other args handling stuff we did don't use wb_log
<f_>
Should I switch to it for usage?
<axtlos>
im not sure, maybe theres a reason not to use it?
<f_>
I don't see any
<f_>
This is the only reason why I used regular fprintf
<axtlos>
then we should switch to wb_log
<navi>
i didn't use wb_log on the old code bc i forgot
<navi>
and no one called it out ~w~
fossdd has quit [Ping timeout: 268 seconds]
norwoodites is now known as pinskia
<f_>
axtlos I replaced fprintf with wb_log
<axtlos>
nice
<f_>
As for dok's issue about xset I think I might know why it happens
<f_>
-terminate might be the cause
<f_>
Maybe it should be set to some timeout or something.
<axtlos>
we should get rid of terminate all together imo
<axtlos>
it was more of a temp solution with the custom child handler not working properly