<Ariadne>
axtlos: i merged your work, we can clean it up with additional PRs, though the wayback-session stuff needs a rebase
<Son_Goku>
Ariadne: what did I do?
<dviola>
question: when running a regular X11 WM via wayback, do you get composite by default? i.e. double buffer, no tearing, similar to running a full wayland session? or is it more like a pure X11 session (i.e. tearing by default)?
<Ariadne>
dviola: composite-only for now
<dviola>
nice
ConanKudo[m]1 has joined #wayback
<ConanKudo[m]1>
👋
<Ariadne>
in fact, everything is composited whether we want it to be or not...
<Ariadne>
video overlays for example, are composited...
* ConanKudo[m]1
goes back through the scrollback
<dviola>
Ariadne: sounds like it would get rid of artifacts that one might get with X without a compositor, I had plenty of those with X in the past
<Ariadne>
dviola: yes, but there's still a lot of work to be done
<Ariadne>
you're trading one set of bugs for another at the moment :p
<dviola>
ah
<dviola>
really cool project
<Ariadne>
we are getting almost to the point where i would be comfortable cutting a 0.1 release, but i need to clean up some of the code i just merged first. i also want to figure out why we need -geometry for xwayland which insists we are 640x480 without it
<Ariadne>
still targeting mid-2026 for something actually solid...
<britney>
very cool project
<whitequark[cis]>
tearing annoys me so much. i think there was some work done on modesetting in xorg to do double buffering, but it was never released or something
<whitequark[cis]>
i had this nightmare issue of "xf86-video-intel gives me no tearing, xf86-video-modesetting gives me render offload"
<Ariadne>
it's okay, you can get double-buffering whether you want it or not, right now, on wayback :p
handlerug has joined #wayback
fitrh has joined #wayback
<dviola>
I'll give it a try soon and report back issues
fitrh has quit [Ping timeout: 272 seconds]
fitrh has joined #wayback
<dviola>
Ariadne: do you think that things like scaling would eventually be possible as well?
<dviola>
fractional scaling, etc
<Ariadne>
no, the X dpi model is really bad
<dviola>
I see
<Ariadne>
you can ask the toolkits to scale though
<Ariadne>
and xrandr to resize the root window
<Ariadne>
which gives you "screen resolutions"
<dviola>
fair enough
<Ariadne>
right now the objective is to implement surfaces xwayland can already use, and then plumb new things in both xwayland and wayback
<dviola>
cool
<Ariadne>
adding new features to X involves the first two steps, but also plumbing X itself, e.g. Xlib, XCB
<dviola>
makes sense
<dviola>
well, just the compositing thing alone without something like picom is already pretty cool I think
<dviola>
there's something nice about these old wms continuing to exist for years to come, and even better if they could inherit some of the wayland advantages
fitrh has quit [Remote host closed the connection]
<whitequark[cis]>
oh, does wayback need specifically wlroots-0.19? it seems unbuildable on stock debian then
<whitequark[cis]>
> `(void*)NULL`
<whitequark[cis]>
isn't NULL already a void*?
<whitequark[cis]>
* > `(void*)NULL`
<whitequark[cis]>
isn't NULL already a void\*?
<Ariadne>
not everywhere, sadly
<Ariadne>
on everything that actually matters? probably
<whitequark[cis]>
infuriatingly, you are right and c99 defines NULL in a way where both 0 and (void*)0 are both acceptable
<whitequark[cis]>
ok, it seems to work with wlroots-0.18
<whitequark[cis]>
i've only tested with the X11 backend though
<whitequark[cis]>
Ariadne: do you want a bot here that would send a message for commits/PRs/etc?
<Ariadne>
eh, maybe :)
<whitequark[cis]>
i run an instance of notifico that i could point here easily enough
<whitequark[cis]>
testing results: (1) only one monitor is picked up and it's not the one i'd like, (2) startplasma-x11 doesn't seem to like xwayback, (3) ctrl+alt+bksp doesn't work to kill the broken session
<whitequark[cis]>
it also complains about not being able to find a seatd socket
<Ariadne>
yeah all of that is in line with current expectations :p
<Ariadne>
right now the kill switch is alt+esc
<Ariadne>
we should probably fix that to be the one people expect :P
<whitequark[cis]>
lemme fix that
<Ariadne>
i just did it
<whitequark[cis]>
oh, nice
<whitequark[cis]>
okay i think i'm done messing with PR #32
<whitequark[cis]>
although i'm now realizing that i might've misunderstood the type of command line arguments that Xwayland should be parsing
<whitequark[cis]>
s/Xwayland/Xwayback/
<whitequark[cis]>
re testing: it looks like startplasma-x11 works fine, and kwin works fine, but plasmashell does not: it just displays nothing
<whitequark[cis]>
also, while wayback detects both monitors, only one of them (the primary one) has any windows displayed, the other one is just black with a cursor on it
Izzy has joined #wayback
<Ariadne>
strange re: plasmashell
<Ariadne>
and yes, at the moment we are having to tell xwayland about display geometry manually. there is a bug, or perhaps output layout detection isn't in any xwayland release yet.
<Ariadne>
so we only tell it about the first, because there's no way to tell it about more than 1 screen yet :D
<whitequark[cis]>
because of the display issue i kinda don't want to debug the plasmashell issue (it's physically difficult with the monitor setup i have), otherwise i'd look closer
<Ariadne>
well since wayback is self-hosting now, and i already have kde plasma 6.4, i can try to debug it tomorrow
<whitequark[cis]>
oh nice! thanks ^^
<whitequark[cis]>
the moment it can run plasma and i3 i'd strongly consider daily-driving it
<whitequark[cis]>
(i assume i3 already works, it doesn't seem like that'd be broken)
<Ariadne>
yeah most things should be working, but kde/gnome-shell probably do weird stuff
<Ariadne>
i need to overhaul the way we pass the environment in the session launcher
<Ariadne>
it will probably fix kde
<dviola>
wayback reminds me to the pipewire approach in some ways
<whitequark[cis]>
the nice thing about pipewire is that i didn't realize it exists well after i started using it
<whitequark[cis]>
more software should be like that
<dviola>
heh
<dviola>
indeed
<Ariadne>
that is the goal here
<dviola>
awesome
scadu[m] has joined #wayback
<scadu[m]>
Just reading here, but wanted to let anyone involved know it’s amazing to see it happening. Wasn’t this stoked for a while now, heh.
scadu[m] has quit [Client Quit]
scadu[m] has joined #wayback
scadu[m] is now known as scadu
paper has joined #wayback
<sq[m]>
I’ve got a bad migraine
<sq[m]>
Ariadne: 👋
<sq[m]>
Sorry I’ve been meaning to say hello
dviola has quit [Ping timeout: 244 seconds]
diego has joined #wayback
diego is now known as Guest5478
prahou has joined #wayback
<axtlos>
Ariadne: RE: wayback-session rebase; sorry i was already asleep when you sent that, i see that you merged the PR already so i assume you did it already?
<Ariadne>
yes
<Ariadne>
all set
<axtlos>
alright great
<f_>
f_[m][m][m] is stuck and thinks it isn't joined can anyone kick it out?
dramforever[m]1 has joined #wayback
<dramforever[m]1>
(Just hypothetically speaking since this probably isn't ready just yet but...) If i were to package wayback for nixpkgs, we can't call that wayback because the name is already taken by another random tool. Is there any other name you think you'll like? I'm thinking wayback-session since the "main program" is called and it tab completes from wayback
<dramforever[m]1>
thanks, that's enough precedent for me
<f_>
you're welcome
<axtlos>
stacksize: Regarding the screen lockers, slock, xscreensaver and xfce4-screensaver all work fine for me now
<axtlos>
although for slock i had to run `xhost +local:`
dok has joined #wayback
<dok>
hi, what's the proper way to use wayback? it worked for me few days ago but doesn't work with the lastests changes (black screen and no way to return to a session tty)
<axtlos>
`wayback-session <session command>`, if you have an .xinitrc you can also just run `wayback-session` and itll use it
<axtlos>
to exit wayback press ctrl+backspace
<dok>
ctrl alt backspace i think
<dok>
thanks i will try this
<axtlos>
just ctrl+backspace works too
<axtlos>
which does sound like a bug actually
<dok>
does wayback-session needs to be properly installed ? because i am trying directly from the _build dir
<axtlos>
xwayback and wayback-compositor have to be in PATH
<dok>
okay
<dok>
uh i got "ERROR: failed to launch Xwayland"
<axtlos>
do you have Xwayland installed?
<dok>
nop, i think i've uninstalled it yesterday
<axtlos>
wayback needs Xwayland
<dok>
just reinstalled it, still doesn't work but behavior has changed a bit
<dok>
things is i also installed wayback from apk, and xwayland doesnt seems to be into the deps
<axtlos>
apk like alpine packages?
<axtlos>
ah yeah, the package is missing the dependency
<f_>
mmmhmmm
<f_>
fossdd: ^
<f_>
I could fix it though
<f_>
eh no I need to backup&migrate my things first so that'll take a while
<dok>
axtlos: okay i am trying to start wayback-session logged on /dev/tty2 (instead of usually /dev/tty1) and i had warnings saying something like failed to open display :0
<axtlos>
what exactly is the command you're running
<dok>
by hard coding display :1 instead it goes a bit further and start X (although without my usual DM)
<f_>
dok: you don't need to hardcode display :1
<f_>
there's a -d option, I usually do "Xwaybar -d :3"
<dok>
PATH=... ./wayback-session/wayback-session |& tee log
<dok>
it now uses display :1 but previously failed with :0 before (i didn't started any X before)
<axtlos>
do you have a login manager running?
<axtlos>
display manager*
<dok>
no i dont think so, what's that ?
<axtlos>
like a graphical login screen
<axtlos>
lightdm, gdm, sddm
<dok>
no
<axtlos>
hm k
<axtlos>
and with display :1 it gets how far?
<dok>
i got to a "solid" dither-like gray background with the X cursor
<axtlos>
but dwm didnt launch?
<dok>
no
<dok>
i can try without the dbus-launch
<axtlos>
yeah ill do some testing later too
<dok>
same with only `exec dwm` in my .xinitrc
<dok>
maybe i messed up something with the PATH
<axtlos>
maybe try installing wayback
<axtlos>
`sudo meson install`
<dok>
...
<axtlos>
hm?
<dok>
maybe my mistake is that i put the local path at the _end_ of PATH... while having wayback already installed in /usr/bin
<axtlos>
ah yeah that could be it
<dok>
still the same behavior, but this was wrong...
<axtlos>
ok i see, im a but busy rn but ill test a bit more later
<axtlos>
you're using alpine, right?
<dok>
yes
<dok>
no worries, you have been of great help
<dok>
i will maybe do something else for a bit
<dok>
okay i got it working
<dok>
but only temporarily -_-
<dok>
okay, i cannot put any `xset ...` before `exec dwm` in .xinitrc
<f_>
I need to test wayback more once I get alpine setup on this laptop
<fossdd>
f_: thanks for the note
<fossdd>
above
<fossdd>
i'll add it as a dep
<f_>
fossdd: you're welcome :)
<f_>
sorry I couldn't do it myself
<fossdd>
its always missing dependencies i get reported, because my local system has like half of all alpine packages installed
<f_>
I'm migrating things so this laptop will be wiped entirely and all that so I'm a bit busy with that rn
thomas_adam has joined #wayback
<f_>
fossdd: oh yes, lol
<f_>
imagine having all the alpine packages installed on your system
<f_>
lots of conflicts but hey
<navi>
axtlos xwayback and wayback-compositor have to be in PATH -- honestly we probably should put wayback-compositor in lib or libexec
<navi>
users have no reason to ever start it manually
<navi>
Xwayback should be in path tho, and distros could even symlink X -> Xwayback eventually
<axtlos>
yeah having it in lib or libexec makes sense
<ConanKudo[m]1>
I was going to bring up moving it to libexec too 😅
<axtlos>
would be a simple change in the meson.build, is libexec always in the same spot? for me its in /usr/libexec, idk if its different on rhel based distros
<ConanKudo[m]1>
it is /usr/libexec in all distributions except Arch
<ConanKudo[m]1>
the Debian family moved to /usr/libexec in 2021
<axtlos>
ok then arch can fix that themselves i guess
<ConanKudo[m]1>
the SUSE family moved in 2018
<ConanKudo[m]1>
and Red Hat family has always used it
<axtlos>
I was also thinking of having a X compat flag in the build, to set up the symlinks automatically
<f_>
in arch it's /usr/lib
<axtlos>
but idk if thats wanted
<axtlos>
f_: so like /usr/lib/libexec, or do they just put everything directly in /usr/lib?
<f_>
everything in /usr/lib
<f_>
it's a mess
<axtlos>
interesting choice
<f_>
Maybe just let distros install it how they want manually? Like e.g. in alpine you can have your APKBUILD run `install` for binaries and such
<f_>
but it would seem that libexec exists everywhere except arch so ..
<axtlos>
that would still require some way to tell xwayback where wayback-compositor is
<navi>
axtlos: use meson's `libexec` optoin
<navi>
option*
<axtlos>
yeah thats what i was gonna do
<axtlos>
i assume arch has a way to overwrite that
<navi>
so that arch can just -Dlibexec=/usr/lib
<navi>
yes
<axtlos>
alright opened a PR
* ConanKudo[m]1
is not sure why Arch hasn't added /usr/libexec at this point
<axtlos>
some "organization" maybe? the guidelines say to put stuff in /usr/lib/<packagename>/
<ConanKudo[m]1>
if we have multiple binaries, /usr/libexec/wayback directory makes sense
<axtlos>
i dont think we'll be getting any extra executables, so `/usr/libexec/wayback-compositor` should be fine
<ConanKudo[m]1>
fine with me
<ConanKudo[m]1>
hmm, maybe it would make sense to move wayback to its own github org so that multiple contributing maintainers can have maintainer privileges
<ConanKudo[m]1>
I did this with PackageKit years ago when I took over
<ConanKudo[m]1>
I know Ariadne doesn't want to be the sole maintainer over time, so having group maintainership would make sense
<axtlos>
Yeah I was thinking the same, but someone offered to host it on fd.o today so maybe we can wait for that if it happens
<ConanKudo[m]1>
I would have also offered to get it set up on fdo as well
<ConanKudo[m]1>
yeah, I don't know why there's a "no compositor" rule on fdo
<ConanKudo[m]1>
it's not a rule I agree with either
<axtlos>
i mean like Ariadne already said, it could be argued that wayback isnt really a compositor like others so maybe it could be an exception
<ConanKudo[m]1>
longer term, I wonder if we'll even keep the wlroots dependency
<ConanKudo[m]1>
wlroots is very churny which is kind of unpleasant
<f_>
oh that'd be great if wayback moves to git.fdo
<axtlos>
im not sure if reimplementing a lot of wayland stuff will be worth it just to get rid of wlrotos tbh
* ConanKudo[m]1
shrugs
<ConanKudo[m]1>
I'm not demanding it or anything
<ConanKudo[m]1>
I just wish wlroots was less painful as a dependency
<f_>
How painful is it for you ConanKudo[m]1?
<ConanKudo[m]1>
well at the moment I can only build it on rawhide, not any fedora stable release
<ConanKudo[m]1>
my general worry is how we handle new wlroots versions and compatibility ranges
<f_>
ah, you need a very new wlroots to compile wayback is the pain I suppose?
<ConanKudo[m]1>
yes
<f_>
mmhmm I see
<thomas_adam>
Are you currently treating wlroots as a subproject in wayback? That might be the best way for now?
<ConanKudo[m]1>
no
<f_>
no not right now
<ConanKudo[m]1>
distros generally don't like that model so we don't do it that way
<ConanKudo[m]1>
and it's not really concern I have right now, since it's fine wayback only works on rawhide
<navi>
subprojects are good for development, not good for distribution
<thomas_adam>
Sure.
<ConanKudo[m]1>
my greater concern is what happens when the next version of wlroots is released
<f_>
pick an old wlroots and stick to it? 🙃
<ConanKudo[m]1>
we're going to need a policy that balances stability with wlroots' release and lifecycle policy
<navi>
wlroots is versioned for parallel install now, so we can just keep using 0.19 until 0.20 is stable everywhere
<f_>
Then "Oh, well it's EoL now"
<ConanKudo[m]1>
which may mean supporting multiple wlroots versions
<f_>
some compat layer thing maybe?
<ConanKudo[m]1>
actually I should rephrase... supporting multiple wlroots versions somehow is going to be a requirement
<f_>
which really just means a hack with a header file with an awful lot of #define's .. nah that won't look good at all
<f_>
ConanKudo[m]1: I agree that'd be nice too
<ConanKudo[m]1>
alright, looks like the xf-bod is okay with wayback moving to fdo gitlab, so I'm going to file a ticket for this
<f_>
yay!
tomate has joined #wayback
<axtlos>
are wlroots releases backwards compatible? I was thinking of seeing how hard it'd be to 'downgrade' wayback to wlroots 0.18, hoping that wlroots 0.19 works with 0.18 stuff
<navi>
not guaranteed
<navi>
(to be backwards compat)
<navi>
(afaik)
<axtlos>
oh hm
<ConanKudo[m]1>
yeah there's no guarantee at all
<axtlos>
how does sway handle it
<f_>
If it were backwards compatible it wouldn't be a pain :P
<navi>
ConanKudo[m]1: does fedora not package wlroots in parallel?
<f_>
that's one of the issues I had with using dwl back then
<navi>
they changed the pc name just to allow that
<f_>
it broke every single time and also required me to install some other version of wlroots
<ConanKudo[m]1>
we do, but we aggressively drop older versions because they're basically EOL
<ConanKudo[m]1>
it can be a security issue to depend on unmaintained wlroots branches
<navi>
i don't think wayback needs to support multiple wlroots at once, but just, somewhat match wlroots as it gets stabilized in distros
<navi>
which is a pain but oh well
<navi>
since our use of wayland is just "whatever Xwayland needs" and nothing else, i doubt there'll be many breakages on wlroots bumps
<axtlos>
yeah that may be the best option
<axtlos>
sway seems to bump the required wlroots for every release
<navi>
sway is developed lockstep with wlroots
<ConanKudo[m]1>
sway is wlroots' reference compositor
<axtlos>
ah right
<ConanKudo[m]1>
weston is libwayland's reference compositor
<navi>
yea, to build sway-git you need wlroots-git
<navi>
ConanKudo[m]1: more like libweston no?
<navi>
libwayland isn't really a "compositor" library but just wayland plumbing
<axtlos>
hmm maybe we could keep wlroots at one version below the latest
<ConanKudo[m]1>
well, libweston came later :)
<ConanKudo[m]1>
but yes
<ConanKudo[m]1>
axtlos: we should evaluate once wlroots 0.20 arrives how we want to handle multiple versions
Achill[m] has joined #wayback
<Achill[m]>
axtlos: generally its best practice to keep using the latest supported version
<Achill[m]>
i see no reason to always keep one below the latest
<ConanKudo[m]1>
if it turns out to be relatively trivial to support 0.19 and 0.20, then let's do that
<navi>
+1
<axtlos>
yeah that sounds good
<Achill[m]>
wlroots dev do a very nice summary every release of api changes
<tsundoku>
hi all, a friend shared Ariadne's blog post with me and I decided to hop in here to keep up with how things are going. I'm an incorrigible FVWM user who has been wondering what I'll need to do when Xorg finally starts disappearing from distros, so I'm really happy that people are working on the problem!
<ConanKudo[m]1>
oddly enough the only wayland compositor library to have a stable ABI and API is mir
<ConanKudo[m]1>
with the miral library
<axtlos>
the canonical mir?
<ConanKudo[m]1>
yup
<ConanKudo[m]1>
that one
<axtlos>
oh wow
<axtlos>
thats unexpected lol
arraybolt3 has joined #wayback
<f_>
tsundoku: fvwm works just fine in wayback
<tsundoku>
great
<f_>
The main issues are basically multimonitor support and monitor management in general
<thomas_adam>
Well, it's limited by what xwayland provides.
<tsundoku>
good to know
<f_>
thomas_adam: the things I mentioned are some of its current limits :)
<tsundoku>
one thing I'm curious about with a wayback setup is how compositing is handled. do I still need to run a separate compositor like picom or does it work differently?
<f_>
eventually as you know already the goal is for it to support X11 entirely so eventually it'll be mostly on-par :)
<f_>
tsundoku: yes
<f_>
Well. You don't need picom in X11 to begin with
<f_>
but I think it should work just fine in wayback, though I'll admit I haven't tried that
<tsundoku>
got it. I'll be interested in following that
<axtlos>
i dont see why it shouldn't work, but im testing it rn
<tsundoku>
and yes I know it's not necessary but I like having compositing
<f_>
Yeah no reason it wouldn't work
<f_>
but if it indeed does not work I think a bug report would be welcome :>
<tsundoku>
OK
<tsundoku>
I may have to set something up with wayback to test at some point
<axtlos>
tsundoku: yeah picom works without any issues
<tsundoku>
great, thanks!
<f_|cis>
awesome
<tsundoku>
I don't know what I'd do if I couldn't get things like shadows and translucent terminals with the window manager of my choice
<tsundoku>
it was earth-shattering when I found out about the standalone X compositors
<tsundoku>
I kind of wish I could get everything compiz does together with other window managers, but
Guest5478 has quit [Quit: WeeChat 4.6.3]
dviola has joined #wayback
hl has joined #wayback
<f_>
hi hl :p
<ConanKudo[m]1>
axtlos: before wayback was launched, I was contemplating making a similar project using the miral library
<ConanKudo[m]1>
simply because of the stable API/ABI thing and needing very little from the compositor library anyway
<Achill[m]>
i genuinly think mir is harder to package than wlroots
<Achill[m]>
also, mir also changes the ABI a lot
<Achill[m]>
afaik also the api quite noticable
<Achill[m]>
especially with stuff like miracle using internal apis
<hl>
hi
ari[m] has quit [Quit: Idle timeout reached: 172800s]
<axtlos>
Just tried building wayback with wlroots0.18 for fun, it seems like it just works?
<axtlos>
builds without any errors and launches fine
<ConanKudo[m]1>
<Achill[m]> "especially with stuff like..." <- miracle is a special case, but miriway has been fine since it only uses miral
<ConanKudo[m]1>
but the other option for a ABI-style one would be QtWaylandCompositor
f_[m][m][m] has quit [Quit: Idle timeout reached: 172800s]
caskd has quit [Ping timeout: 252 seconds]
caskd has joined #wayback
rbrue has joined #wayback
<Son_Goku>
axtlos: I kind of expect that it would at this stage, as I don't think we use much from wlr api
<axtlos>
yeah true
<axtlos>
maybe itd make sense to lower the wlroots version right now to have it work on non rawhide fedora
<Ariadne>
yeah let’s add a fallback
rbrue has quit [Ping timeout: 245 seconds]
<Ariadne>
ConanKudo[m]1: github acked
<Ariadne>
err gitlab
<Ariadne>
whatever :)
<ConanKudo[m]1>
cool
<Ariadne>
sorry i am still waking up
<ConanKudo[m]1>
I can do the work to move things once we've got merges done on github
<ConanKudo[m]1>
could you add me to the wayback repo so that I can use the importer (assuming we want to keep the issues)
<Ariadne>
yeah give me a sec
<Ariadne>
we also need to update the code of conduct document from the one i just normally plop down to the freedesktop one
<ConanKudo[m]1>
yeah I plan to do that just before importing
<Ariadne>
(the one i normally use basically sums up to "don't be an asshole, or i'll be forced to ban you")
<ConanKudo[m]1>
sensible :)
<ConanKudo[m]1>
Ariadne: are you considering attending XDC this year?
<Ariadne>
i have concerns about leaving the US right now
<ConanKudo[m]1>
that's fair
<Ariadne>
e.g. not being able to get back in
<ConanKudo[m]1>
that's very fair
<Ariadne>
i think my plan is to sign up for global entry as a vibe check
<Ariadne>
if they approve it, i'll probably go
<ConanKudo[m]1>
basically wondering if we want to do a talk proposal about wayback for XDC
<Ariadne>
ConanKudo[m]1: what is your github handle
<ConanKudo[m]1>
Conan-Kudo
<Ariadne>
invite sent
<ConanKudo[m]1>
accepted
<ConanKudo[m]1>
hmm I don't have access to repo settings
<Ariadne>
humm
<ConanKudo[m]1>
this is probably a dumb github thing :/
<axtlos>
make sure the role is set to maintain or admin in the collaboraters settings
<Ariadne>
there isn't any roles because it's not an organization repo :/
<Ariadne>
ugh
<Ariadne>
do i want to create a github org that we will probably be discarding immediately
<Ariadne>
i guess i have no choice... github sure is a platform of all time
<axtlos>
wow there really is no way to give people access to the repo settings?
<ConanKudo[m]1>
I guess not anymore
<ConanKudo[m]1>
I remember having this problem with PackageKit years ago
rbrue has joined #wayback
<Ariadne>
i will suck it up and create an org, transfer wayback to the org, and grant ConanKudo[m]1 admin of the org to handle the migration i guess
<ConanKudo[m]1>
oof thanks :)
<Ariadne>
it seems silly and wasteful but that's how github works i guess
<ConanKudo[m]1>
eh the org will stick around forever since we have PR references in our git commit history anyway
<Ariadne>
github: design a good ACL system challenge, 2025, difficulty: impossible apparently
<ConanKudo[m]1>
yay merge commits :(
<Ariadne>
sorry, but i do like merge commits ;)
<ConanKudo[m]1>
I don't mind them that much
<Ariadne>
they create nice backout points
<Ariadne>
while preserving history
<ConanKudo[m]1>
I prefer semi-linear merges (gitlab parlance) where the result is fully bisectable linear history while preserving merge points
<ConanKudo[m]1>
assuming I have merge commits at all, as usually for my own projects I avoid merge commits due to preferring to have detailed logical commits with info
<Ariadne>
holy mother of god, they have ads for copilot inside the organization create tool now
<f_>
I also have f_|ic for irccloud, f_|jab for jabber/xmpp, f_|pmOS for the pmOS irc bridge etc etc
<f_>
:P
<axtlos>
makes sense
<f_>
I really don't like "[m]", makes me kinda look like a spammer tbh
<axtlos>
my first guess was cisgender but i didnt see why youd have that in the username lol
<f_>
🤣
<f_>
Speaking of topic changes from matrix to irc:
<f_>
Received an error on irc.libera.chat: err_chanoprivsneeded
<f_>
["f_|cis","#wayback","You're not a channel operator"]
<ConanKudo[m]1>
I would have fixed it, except... I'm afraid of what would happen since this is on libera.chat
<f_>
ConanKudo[m]1: fixed what?
<whitequark[cis]>
yeah, your puppet needs to be a chanop for the sync to work
<ConanKudo[m]1>
my irc nick
<f_>
whitequark[cis]: yep
<f_>
ConanKudo[m]1: nothing would happen?
<f_>
I'm not sure what you mean
<f_>
your nick would change but that's it
<f_>
Nothing wrong with that
<ConanKudo[m]1>
yes, no more [m] with the correct nick
<ConanKudo[m]1>
I have that set up on OFTC
<whitequark[cis]>
re: "[cis]" suffix, although it does refer to "catircservices", i started using [cis] because the ambiguity with "cisgender" amused me
<ConanKudo[m]1>
😆
rbrue has quit [Quit: Konversation terminated!]
rbrue_ has joined #wayback
<vyivel>
so not commonwealth of independent states after all
<axtlos>
Ariadne: you should probably also close PR #26, its outdated by now
<Ariadne>
i hate closing unresolved PRs as i don't want to demotivate them from contributing again, but yeah i suppose so
<whitequark[cis]>
tbf keeping the PR hanging open can also be quite demotivating
<Ariadne>
also true
<axtlos>
yeah thats fair, though i explained already that this would be more fit for some future "wiki" so it's probably not that bad
<ConanKudo[m]1>
well, we will eventually get a wayback-site repo set up on gitlab.fdo
<Ariadne>
anyway, once we are on fd.o, i'll grant some permissions to axtlos over there. from that point, we want to require code review IMO.
<whitequark[cis]>
i do think a usage section is needed because i didn't realize wayback-session is even a thing
<Ariadne>
i agree
<axtlos>
yeah we should probably set up Contributing guidelines too
f_|cis is now known as f_|cat
<ConanKudo[m]1>
yup
<f_|cat>
no more imbiguity
<f_|cat>
"cat" now means "catircservices" and as a bonus means "I also like cats"
<whitequark[cis]>
nice
<Ariadne>
it can however also mean that you are a cat
<Ariadne>
how many times per day do you meow
<f_|cat>
2xbonus
<whitequark[cis]>
meow
<f_|cat>
meow
<whitequark[cis]>
i am a cat. i am also cis, obviously
<f_>
meow
<f_>
all for code review obviously :)
<f_>
20:52 <ConanKudo[m]1> well, we will eventually get a wayback-site repo set up on gitlab.fdo
<f_>
who would work on the website
* ConanKudo[m]1
shrugs
<ConanKudo[m]1>
I asked for it so we can have wayback.freedesktop.org
<f_>
hand-written html is my passion :)
<axtlos>
i say we just set up a groff page generated through mandoc /j
<whitequark[cis]>
nothing wrong with handwritten html
<ConanKudo[m]1>
I'm the terrible person that likes sphinx :P
<Ariadne>
that is part of the reason why moving to fd.o now makes a lot of sense. we will be able to track everything needed across the stack in one spot
<f_>
(not mine)
<whitequark[cis]>
it's not wrong of course
<jvvv>
axtlos: thanks for sharing the 'xhost +local:' ... now my screen locker is working
<f_>
\o/
<Ariadne>
that's another thing we need to hook up... DPMS
<ConanKudo[m]1>
do we want a mailing list btw?
<whitequark[cis]>
please no
<Ariadne>
i'd rather not
* ConanKudo[m]1
shrugs
<ConanKudo[m]1>
I figured I'd ask
<ConanKudo[m]1>
don't have strong feelings either way
<f_>
bearable? whitequark[cis]: not sure if Mailman3 does this
<ConanKudo[m]1>
you now use your fdo gitlab account for sso
<whitequark[cis]>
i actually have zero issues with cleartext password storage by mailman
<f_>
but not sure if there's a need for a mailing list rn
<ConanKudo[m]1>
and you can associate multiple email addresses to a single account
<whitequark[cis]>
i just think mailing lists are a terrible user experience
<Ariadne>
i just hate email
* f_
git send-email to Ariadne
<Ariadne>
git send-irc
<whitequark[cis]>
i would rather send git patches over netcat or qr-code-by-post than use git-send-email
<f_>
::: git-send-irc (irc@git-scm.com) has joined #wayback
<whitequark[cis]>
i would consider microfiche or encoding them in DNA but i'd have to tool up for that first
<f_>
you're limited wrt qr codes
<f_>
I think the biggest ones are ~3kb or something.
<Ariadne>
also i guess wayback needs a logo or whatever
<f_>
art design is my passion
<f_>
(no)
<Ariadne>
i just worry about the xlibre weirdos shitting up some sort of logo thread with nazi shit
<f_>
xlibre doesn't have a logo does it
<Ariadne>
idk
<f_>
do you care?
<Ariadne>
do i care about?
<f_>
xlibre
<axtlos>
they do kinda have a logo but its pretty boring
<axtlos>
just a grey X with a drop shadow
<Ariadne>
not really, but i do care about any moderation incidents that could happen
<f_>
okay, fair enough
<f_>
axtlos: oh boring
<axtlos>
i think if we want a logo we should probably get one quickly and maybe organize it here
<axtlos>
it wont get as much of an outreach but itd be the safest
<Ariadne>
like if i post "make a logo for wayback" on fedi, we will get swastika submissions, i guarantee it
<axtlos>
oh yeah id definetly not post it on fedi
<Ariadne>
because that's how reactionary dipshits are
<ConanKudo[m]1>
wonder if a "W" stylized like the "X" would be enough here?
<Ariadne>
they're too busy "owning the woke DEI big tech wayland establishment" to realize that 2^16 is 18, not 65536
<ConanKudo[m]1>
~_~
<Ariadne>
ConanKudo[m]1: i like it
<whitequark[cis]>
You cough and sneeze. "Ah, another bug," runs the realization through your head. You put the napkin into the sequencer, plug its USB cord into your laptop, and start the awful Java application it came from. Fifteen minutes later, it beeps, trimphantly, and in the middle of the HAdV-A31 genome, you see a chunk of plaintext. Scrolling down to the one where the mutation rate hasn't overwhelmed the erasure coding, you copy it out and
<whitequark[cis]>
paste into git am. "We really should switch to a more efficient system," you muse, "latency used to be measured in milliseconds, not megaseconds."
<whitequark[cis]>
(still a better user experience than mailing lists)
<Ariadne>
the problem with mailing lists is that most people use webmail now
<ConanKudo[m]1>
�
<ConanKudo[m]1>
blech
<axtlos>
i could ask some artist friends if they have any ideas for a wayback logo
<whitequark[cis]>
💭 crop of back-to-the-future time machine front panel with "present time" being the x11r6 release date