<ToyKeeper>
Reading scrollback, I hope in the future Wayback will be able to map Wayland windows to allow the X11 window manager to manage them.
norwoodites has joined #wayback
<ToyKeeper>
... and by "future" I mean maybe within a few years. I'm not in a hurry. There are no Wayland-specific apps I need to run and I don't expect to stop using native X11 until I no longer have a choice.
pinskia has quit [Ping timeout: 276 seconds]
<ToyKeeper>
I should probably catch up on technical debt for my chat infrastructure first... like I've been meaning to set up a private Matrix just to puppet to other networks and merge everything into one place. Have just been dragging my feet because Matrix administration is kind of a pain, and I haven't finished my logging client yet.
<whitequark[cis]1>
i'm excited about wayback because i have my i3 setup and i like it
<whitequark[cis]1>
and i basically don't care about wayland until wine has excellent wayland support
pinskia has joined #wayback
norwoodites has quit [Ping timeout: 244 seconds]
<ToyKeeper>
I used i3 for a while, and enjoyed it. Pretty decent window manager. But the one I ended up sticking with is Sawfish, basically the Emacs of window managers. Been using it for literally decades now, and it's intensely customized. Nothing else like it really exists.
<ToyKeeper>
In particular, it has been really handy being able to rewrite its code on the fly while it's running, without having to log out. I only log out like... once every year or two, when it's time to upgrade the entire OS.
<britney>
Emacs, nice
<britney>
I'm writing an emacs operating system
<britney>
it'll just be the linux kernel + emacs
<britney>
Emacs/Linux
<britney>
thats all you need
<britney>
when you boot up, it doesnt have the bloat of nonsense like init
<britney>
it just boots directly into emacs
fitrh has joined #wayback
arraybolt3_ has joined #wayback
arraybolt3 has quit [Ping timeout: 260 seconds]
fitrh has quit [Remote host closed the connection]
fitrh has joined #wayback
<ToyKeeper>
I enjoyed my time with Emacs for several years, but I had to write and maintain a lot of custom code to make it do what I wanted, and it still was only okay-ish for me. So I tried a bunch of others, and when I got to Vim, it only took a few hours for me to know I had found my new home. But YMMV.
<ToyKeeper>
For a window manager though... Lisp has been great. It's not my favorite language, but I love how it lets me completely change the code while it's running. Very useful for a process which is expected to run for like a year at a time.
fitrh has quit [Remote host closed the connection]
norwoodites has joined #wayback
pinskia has quit [Ping timeout: 244 seconds]
norwoodites has quit [Read error: Connection reset by peer]
pinskia has joined #wayback
pinskia has quit [Read error: Connection reset by peer]
pinskia has joined #wayback
<f_>
Heh
<f_>
Re <Ariadne> i just cancelled my irccloud subscription :p
<f_>
whitequark[cis]1 Heard of sway?
f_[m]`[m] is now known as f_|cis
<f_|cis>
<whitequark[cis]1> "on the irc side your nick..." <- nick fixed
f_|ic has quit [Quit: Connection closed for inactivity]
mebious has joined #wayback
<whitequark[cis]1>
<ToyKeeper> "In particular, it has been..." <- i really don't want my window manager to be programmable. i tried that with awesome and it was not awesome
<whitequark[cis]1>
i know about sway but i avoided it because of some deranged drama about nvidia's egl implementation
<f_>
whitequark[cis]1: I use sway with nouveau and it works perfectly fine
<f_>
well, as fine as nouveau on an old nvidia fermi GPU can get
<navi>
any gpu limitation sway has wayback will also have, due to using wlroots
<whitequark[cis]1>
i only use proprietary nvidia drivers
<navi>
though now a days nvidia proprietary has gbm support afaik
<whitequark[cis]1>
navi: oh yeah, my issue was less "i will be impacted by this" (not really since i've only ever used nvidia gpus for render offload, which on wayland i think doesn't involve the compositor basically at all?) and more "is the development process okay"
<whitequark[cis]1>
the big wayland blocker for me is wine support. wine is a key part of my workflow so things like "window positioning doesn't work" are nonstarters
<navi>
it's okay afaik, the "drama" was simply refusing to implement a whole alternative api just for nvidia, and well, that worked enough to make nvidia implement gbm like everyone else
<navi>
as for wine, yeah, wayback will work for that
<navi>
i hope that one day we can have a protocol that allows wine to simulate positioning (w/o the whole global coordinate issue)
<navi>
(actually, does that work via Xwayland? since sway acts like a x11 wm for xwayland, dunno)
<f_>
Wine not working in wayland is news to me
<f_>
But it's been a while since I last needed it
<navi>
it works pretty well under xwayland, on the other hand, the waylanddrv, so native wayland, is quite buggy still
<f_|cis>
If it works in Xwayland I reckon that's already okay?
<whitequark[cis]1>
yeah i was looking at waylanddrv support
<whitequark[cis]1>
if i'm going to be using xwayland i might as well just continue using x11, which works and requires spending zero additional effort
<whitequark[cis]1>
(i think i'm missing out on better hidpi scaling, but i've seen friends with wayland sessions and they have different dpi scaling issues with xwayland, so it's not clear there's any real benefit)
<f_>
DPI scaling works just fine with Xwayland, it's just making Xwayland apps blurry but .. not much you can do
<f_>
it's X and X never had good DPI support
<whitequark[cis]1>
i have DPI scaling working fine on X11 right now, so going to blurry Xwayland windows will be a downgrade
<whitequark[cis]1>
(yes, I know this is courtesy of GTK, Qt, and Chromium and not X11 itself, but the point is that nothing on my screen is blurry)
<whitequark[cis]1>
(mostly Qt and Electron to be honest, I never got scaling to work well with GTK and I mostly gave up. for some reason I can't get the text scale to line up with UI element scale and I'm really tired of trying. Qt just needs an environment variable set, which for some reason startplasma does not do right, but that's at least easy to fix)
<f_>
whitequark[cis]1: oh really? I remember it being an extreme pain to get right
<whitequark[cis]1>
yep, it improved a lot over the years
<f_>
I remember needing lots of xrandr hackery and such to get it working well
<f_>
which would result in blurriness depending on how you do it
<whitequark[cis]1>
for electron apps i add --force-device-scale-factor=2 and for Qt apps i add QT_SCREEN_SCALE_FACTORS= QT_SCALE_FACTOR=2
<whitequark[cis]1>
KDE apps pick it up automatically from the environment
<whitequark[cis]1>
that's it
<f_>
oh, wouldn't it be nice to just set this once in sway config and for all
<f_>
:P
<f_>
(it even works w/gtk)
<whitequark[cis]1>
not at the cost of suffering blurriness elsewhere
<whitequark[cis]1>
i think it should (as in, was supposed to) work automatically with non-KDE Qt apps, but there's an obnoxious interaction where i can set up my environment in a way where KDE xor Qt applications have correct scaling
<f_|cis>
fair enough
<whitequark[cis]1>
this is clearly just a bug, and i could fix it, but i couldn't find time to do so yet
<whitequark[cis]1>
(actually i think that's a KDE nand Qt, strictly speaking?)
<f_|cis>
wayback == using Wayland while prograstinating on actually moving to a Wayland desktop ;)
<whitequark[cis]1>
yes, exactly! that's why i like it
<f_|cis>
s/prograstinating/procrastinating/
<whitequark[cis]1>
procrastination is good: it gives other people time to fix the problems i'm having
<f_>
lol
<whitequark[cis]1>
if i didn't procrastinate as much, i'd have to fix many more bugs in a much smaller timeframe. and i already maintain far too much shit
<f_>
for me it was a pretty smooth transition
<f_>
from dwm to dwl
<f_>
but suckless-style things tend to break on every system upgrade, so I moved to sway to be able to stop caring once and for all
<whitequark[cis]1>
e.g. i'm familiar enough with wine to probably be able to fix waylanddrv issues (setting aside the part where this would probably run into conflict with the gnome compositor values), but i already have a lifestyle where i build my own lineageos images, package an entire FPGA toolchain for several platforms, maintain a hardware definition language, and so on
<whitequark[cis]1>
at least i stopped maintaining an mCAD
<whitequark[cis]1>
re: blurriness: i started using "hidpi" displays the same year they became available outside of apple devices. i think 2012 or so? it was ... challenging, absolutely nothing on linux supported it, we were still dealing with gtk 2 and qt 4 (with a bunch of qt 3 applications, i think?). i want my desktop to look good: i'm looking at it 10+ hours a day easily. so i don't really see a downgrade to "yeah it's blurry again", after over 10
<whitequark[cis]1>
years of making it not blurry, as acceptable
<whitequark[cis]1>
i like wayland in theory! i was really excited to see it get ready for use. but i don't like, well, actually using it, at least right now
<whitequark[cis]1>
also i don't think you can run sway and have plasma panels displayed in it, can you?
<whitequark[cis]1>
this is the current setup i have: i3 as a WM, KDE for basically everything else on the desktop
<f_>
whitequark[cis]1: Oh? You use i3 inside KDE Plasma?
<f_>
Interesting
<whitequark[cis]1>
yep! works like a charm, i just have a line in the config to move the background plasma window (where you'd put desktop widgets, which i do not use) to a scratchpad
<f_>
cool
<navi>
i've done that once a good while ago as well, though i gave up due to some popups misbehaving
<navi>
i've also seen people doing that with xfce
<whitequark[cis]1>
i used to have a 200 line .xsession file until the migration to KDE 6, which broke it. except then i discovered that KDE 6's startplasma-x11 script (which replaced the shell script that was there before) actually makes my configuration Just Work, and even better than the .xsession hack i had before
<whitequark[cis]1>
this is what motivated me to start giving KDE $100/year in donations
<whitequark[cis]1>
navi: i've hit some occasional issues, and i've reported one bug to... i don't remember if it was i3 or kde, maybe both? but it wasn't time-consuming or difficult to work around the few times when i hit them
<whitequark[cis]1>
(also by "startplasma-x11 script" i mean "c++ executable that hardcodes most of the startup process")
norwoodites is now known as pinskia
<axtlos>
i used to always have sway running nested whenever i needed to tab windows together, its a real shame that tabbed windows isnt a more widespread feature
awfawd has joined #wayback
<awfawd>
hi (from 129.222.151.*)
awfawd has quit [Client Quit]
<f_>
awfawd: ?
<whitequark[cis]1>
they're bragging about using starlink in california
<whitequark[cis]1>
i mean, i assume they're bragging. i can't imagine what other purpose would be achieved here
<f_>
22:20 ::: awfawd (root) [~root@45.32.136.24] has joined #wayback
<f_>
that's not even a 129.222.151 ip
<f_>
129.222.151.0/24*
<axtlos>
congrats on supporting musk i guess? lol
<f_>
anyway
<leah>
arraybolt3_: yeah i knew about wayback when ariadne started working on it
<leah>
i saw in her readme that this channel is a thing, so i thought i'd check it out
<Ariadne>
actually, before. i talked about it a few times at fosdem with several people
<leah>
well, yes, but you discussed it with me in librespeech a day or two before you started working on the project
<leah>
arraybolt3_ said they did a "double-take" when they saw me join; and i had to look up what that term means
<leah>
hardly the reaction i'd expect, since i don't consider myself to be special in any way, but i felt it might be nice to tell the person how i found out about wayback
<whitequark[cis]1>
are you leah neukirchen, or another leah?
<leah>
oh, no, i'm not cool enough to be german leah
<leah>
leah neukirchen is leah2 on libera
<leah>
i'm leah rowe of libreboot
<leah>
leah2 is way cooler than me
<arraybolt3_>
leah: I mean, Libreboot is an extremely cool project :)
<whitequark[cis]1>
yeah!
<arraybolt3_>
(I had a Librebooted desktop, unfortunately something on the motherboard spontaneously fried and now that system is dead :( )
<leah>
i tell you what is cool. taking a bite out of rotten food /s
<leah>
i need to go afk to buy not-rotten food
<f_>
leah rowe? not cool enough?
<arraybolt3_>
ugh, sorry to hear that
<leah>
brb. but yeah, german-leah is way cooler
<leah>
(that's what i call leah2. german-leah)
<f_>
I have some cookies for when you come back
<f_>
:P
<arraybolt3_>
I have my browser set to reject cookies
<arraybolt3_>
/s
* f_
brings a box of cookies for arraybolt3
<f_>
I have many more cookies if anyone else wants ;p
<leah>
i have returned
<leah>
and hopefully without food poisoning
<f_>
welcome back
<arraybolt3_>
hopefully not as bad as the time I unintentionally attempted to drink fermented cranberry juice
* f_
brings a box of cookies for leah
<arraybolt3_>
key word: attempted. Did not succeed.
<leah>
didn't think to buy cookies. shop was closing. i grabbed a micromave curry, some potato chips, some instant noodles and a few litres of milk
<leah>
i normally instinctively grab salty things, not sugary things
<leah>
i already have cornflakes. that's what the milk is for.
<leah>
"double take". i like it. never heard this term before. i knew what the expression it describes was, just didn't knew it had a name
<leah>
now to be on-topic: i thought occured to me, regarding wayback
<leah>
and this is not to discourage the development of such a project, but let's say in N months from now it's completely stable and complete, and is a perfect drop-in replacement for X - and then by next year, every distro is using it instead of xorg
<leah>
the current problem with xorg is just because of how big it is, and how much of a burden that is alongside wayland
<leah>
wayback solves this problem, but creates another:
<leah>
i forsee that it'll work so well, and be so perfect, and so easily maintainable, so as to create a completely frictionless experiencing for linux distro maintainers
<leah>
and that itself is not a problem, but it creates another problem, that i can very reliably predict:
<leah>
it may in fact slow down the adoption of wayland, which is otherwise technically superior
<leah>
i mean, it uses wayland
<leah>
but a lot of people probably still don't bother with wayland, simply because what they're using still works - and the fear that xorg will die off in the next few years prompts many people to think about native wayland support, for their DE, their applications, etc
<leah>
xwayland is also a thing, but yeah, i forsee that possibility
<leah>
and to be clear, this project is needed, and should continue. and i will do all i can to help it.
<leah>
what do others here think of that?
<whitequark[cis]1>
isn't xwayland guaranteed to be used for foreseeable future by basically everyone involved?
<leah>
it's probably not as big of a problem as i imagine, but it's a possibility i thought of
<whitequark[cis]1>
sorry, s/used/maintained/
<whitequark[cis]1>
(but the two are of course linked)
<leah>
well, if a larger number of people continue using and maintained x11-based window managers and desktops, it could be that they would otherwise spend more time developing wayland compositors, and contributing to the development of the wayland protocol
<leah>
it has been my experience that most people don't want to change until they have to
<leah>
this is a human problem, not a technical problem
<whitequark[cis]1>
i'm not sure that annoying or cornering people into change is a particularly great way to solve the adoption of a functionally similar technical protocol
<whitequark[cis]1>
for one, this is how you get reactionary projects like devuan or xlibre
<leah>
yeah like, take MATE for example. it works on wayland now. if wayback existed 5 years ago, maybe MATE wouldn't have bothered porting to wayland
<whitequark[cis]1>
but more generally, if i [general] am forced to change, i may well adopt a strategy of looking for the shortest available path, rather than getting invested in putting in lots of effort into a great protocol RFC or the like. why would i? the new thing has, after all, been nothing but annoyance
<whitequark[cis]1>
that seems fine to me. if wayland is clearly technically superior people will migrate to it eventually anyway
<leah>
i'm the same way. for example, i recently configured a new linux-based router, with a special tunnel service that i use. i set up a firewall, and i already know how to use iptables. will i spend time learning nftables? hell no. i'm using iptables. i know for 20 years how to use it.
<whitequark[cis]1>
for example, i use ip instead of ifconfig because (for all the hate it gets) iproute2 is a clearly technically superior solution
<leah>
(and the fact that i'm actually using nftables, and the iptables program i used is just a frontend for nftables, is besides the point)
<whitequark[cis]1>
re: iptables example: if you are using two different frontends for the same thing anyway, what difference it makes what their adoption rate is?
<whitequark[cis]1>
* re: iptables example: if you are using two different frontends for the same thing anyway, what difference does it make what their adoption rate is?
<leah>
i brought up that example because it perfectly correlates with wayback in regard to wayland
<leah>
if debian simply forced me to use nftables commands, and didn't provide a compatibility layer so that i could still use iptables commands, i would have spent a few hours teaching myself how to directly interact with nftables
<axtlos>
I dont think it'll a big issue with distros dropping X support in many desktops, like I dont think that fedora will stop pushing wayland because wayback exists now
<leah>
(the router is running on debian)
<whitequark[cis]1>
if i wanted my operating system to give me homework every upgrade cycle i would have been running windows
<leah>
humans are generally quite stubborn about interfaces, and it's not just in software
<jvvv>
leah: this may or may not be associated to what you are saying: i'm interested in this project because the wm i prefer (bspwm) has no exact analogy in wayland. so to come up to speed with trying to help with wayback, i am in the process of learning wayland programming... so in turn i am concidering writing a bspwm like compositor
<axtlos>
i dont think the iptables and nftables example is comparable here, since they provide very different user interfaces, but with desktops users will just get gnome/kde/xfce, generally they are the same no matter if running on wayland or xorg
<axtlos>
well they arent the same right now, but the effort is to have them behave the same way
<whitequark[cis]1>
i don't understand this argument. i have a busy life, filled with things like "chronic pain" and "doing things that aren't involved in the minutae of interfaces in the linux guts". how is forcing me to participate in the wayland ecosystem a good thing?
<leah>
well, my thinking was from the perspective of a developer (of a desktop environment), not a user
<axtlos>
i think that'll also resolve itself simply by wayland getting better features
<leah>
i mean, several distros now are dropping xorg, regardless of other implementations. and that might be good too. jolting users into adapting
<whitequark[cis]1>
can't you talk to the developers of desktop environments (this is probably like one big room worth of people) and convince them to use wayland instead of doing this vaguely coercive upgrade dance?
<leah>
but yeah, humans always choose the easiest most familiar thing
<axtlos>
like hdr for example, i dont think xwayback will be able to provide hdr, any desktop that wants to support hdr would move to wayland
<leah>
take newspapers for example. when the internet came around, people found it scary at first. 30 years later, you look at any news site and it still looks like a newspaper
<whitequark[cis]1>
just learned that i'm not human (because i frequently choose the long and unfamiliar thing) :D
<leah>
we used to listen to the radio
<leah>
now we listen to podcasts
<leah>
and that's basically the same interface
<leah>
when don't dislike change, but they want everything to feel the same
<leah>
people don't*
<leah>
so if you give people something that is exactly the same, they'll use that
<leah>
they only change when they have to
<leah>
which is why i like this project. because i'm an lxde user and refuse to change.
<whitequark[cis]1>
the thing about x11 and wayland is that they're basically the same thing
<whitequark[cis]1>
just like how gnome, kde, xfce, and the windows shell are basically the same thing
<leah>
most people doen't interact directly with x11 or wayland anyway
<leah>
they're using some library e.g. qt/gtk and that interacts with it
<whitequark[cis]1>
so if you give people homework (whether that would be end users, or application developers) you get to justify it
<whitequark[cis]1>
otherwise it's not change, it's just churn
<whitequark[cis]1>
people in technical communities seem to stratify by how much tolerance for churn they have. e.g. javascript developers for the most part have an astronomical degree of tolerance to churn in dependencies. c developers much less so. this is something you see everywhere and it's most likely inevitable
<whitequark[cis]1>
those who have a high degree of tolerance for churn (maybe they don't have much to do with their life but to fiddle with display servers--which can be a very useful thing!) are the early adopters. and you can speed up adoption by turning churn into change in the minds of those affected, by justifying it with the positive outcomes they'll see from adoption
<f_>
23:23 <leah> i brought up that example because it perfectly correlates with wayback in regard to wayland
<f_>
leah: so perfectly iptables is now a backwards compat command for nftables
<leah>
yes, and because of that, i don't need to learn nftables commands
<leah>
that was the point i was trying to make in comparison to wayback/wayland. a lot of devs will probably just keep doing x11 stuff forever
<f_|cis>
yep
<f_|cis>
If they still want to do x11 stuff then so be it, I'd say
<leah>
and that's a good thing. wayback makes this choice viable. but my point was that without such a choice, wayback would be adopted much faster by everyone.
<leah>
sorry i meannt: without this choice (wayback), wayland (natively) would be adopted faster
<leah>
since it was clear that xorg is dying. wayback removes the problem entirely.
<f_|cis>
nah I don't think so
<f_|cis>
xorg is still being somewhat a little maintained
<f_|cis>
in that if you report serious security issues they eventually get fixed
<whitequark[cis]1>
also, isn't xwayland technically a configuration of xorg?
<f_|cis>
It seems to be somewhat a little separate
<f_|cis>
it has its own releases and such, but I think otherwise is a branch of xorg
<whitequark[cis]1>
my understanding was that it's a configuration of xorg built with just the wayland backend
<whitequark[cis]1>
so all of the x11-per-se code remains, it's just the display drivers that get yeeted
<f_>
That's already a pretty good improvement I think
<whitequark[cis]1>
oh yes. but my point is that even in the "best possible case" for the purpose of this conversation (wayback doesn't exist, every display driver is git rm'd) application developers still get to use x11 as long as they please
<f_|cis>
sure
<whitequark[cis]1>
so the only people who would be "forced" to do anything about it are window manager and desktop environment developers, most of whom seem to be at the forefront of wayland adoption already
<whitequark[cis]1>
e.g. i believe that kde soft-deprecated running it on x11, in that it's going to remain a part of kde only insofar as people are fixing bugs and generally updating it, and not a lot of people do this even now
<leah>
interesting
<leah>
someone mentioned devuan earlier
<whitequark[cis]1>
i haven't specifically read anything about gnome's stance on x11 but i find it unlikely that it's more welcoming
<f_>
GNOME is killing x11 entirely
<f_>
KDE seems a bit more lax
<f_|cis>
leah: this? 21:20 <whitequark[cis]1> for one, this is how you get reactionary projects like devuan or xlibre
<whitequark[cis]1>
f_: here we go
<f_>
whitequark[cis]1: well, gnome is killing a lot more things
<f_>
such as non-systemd support :P
<whitequark[cis]1>
(as an aside: the outcome of gnome forcing change on me--it was the gtk 4 HIG transition--was that i stopped using gnome)
<leah>
f_: yeah, that was the one. i don't really care about devuan, but i checked out their site on a whim
<leah>
they call themselves a "GNU+Linux distro". and i'm pretty sure the + was my influence
<f_>
lol
<leah>
back when i used to be in GNU Leah Mode, i regularly told everyone to use + instead of /
<leah>
if i recall, my only contact with them was in around 2021 or so. let's check the wayback machine
<leah>
systemd is awesome anyway. i don't take devuan seriously. they're a meme.
<f_|cis>
they announced it on twitter along with an image saying "Who would you give control to? Red Hat or Xlibre"
<leah>
i welcome our benevolent redhat overlords
<leah>
they make me feel safe
<f_|cis>
As if it were "Red Hat v.s. Xlibre"
<leah>
ok yeah, 2014 devuan said gnu/linux
<f_|cis>
when it really just is "freedesktop CoC team v.s. metux"
<whitequark[cis]1>
oh, is that what happened
<leah>
2015 too.
<whitequark[cis]1>
that explains a lot actually
<axtlos>
"devs who know what they're doing vs x^2 == pow(x,2)"
<leah>
2016 too
<f_>
axtlos: laughed out loud at that one, thanks
<f_>
made my day
<leah>
yep 2018.
<leah>
devuan started calling it GNU+Linux in 2018. before then, GNU/Linux
<f_>
axtlos: I suspect that metux doesn't seem to properly test his changes even.
<f_>
whitequark[cis]1: what happened?
<leah>
i'm pretty sure that was my influence. i was on a massive really annoying campaign back then getting everyone to use +. e.g. libreboot used +
<leah>
and like 2/3 of devuan people are libreboot users
<leah>
so
<f_>
gnu+leahnux
<axtlos>
f_: honestly yeah, a wrong binary operation would definetly be noticeable when testing
<leah>
i think if a distro wants to use xlibre now, they can. but using it right now seems silly
<f_>
s/right now// and I agree
<leah>
even metux himself said the recently 25 release is likely very buggy
<f_>
very buggy yes
<f_>
x^2 = pow(x,2)
<leah>
if i were so inclined, and i'm not so inclined, but if i were, i would wait a year
<whitequark[cis]1>
f_: your statement suggests that xlibre is basically a grievance project in response to coc enforcement. i am saying that that checks out because it looks like a grievance project in response to something
<f_>
whitequark[cis]1: oh, yes exactly
<f_>
metux got kicked out of FDO on CoC grounds
<axtlos>
i do kinda want to test xlibre for the sillies but i cant be assed to install some meme distro in a vm for that
<leah>
i think metux just wants to be bdfl and didn't like other people telling him what to do
<leah>
i spoke with ariadne about it, and did more research based on what she told me. mostly because i like to verify what i'm told
<leah>
he(metux) regularly gets in disputes over often very petty reasons
<f_>
it wasn't redhat or anything. Completely on CoC grounds
<leah>
he strikes me as someone who just wants to have control
<axtlos>
wasn't it also just because of very bad code quality
<leah>
the redhat thing is just a useful scapegoat for him
<leah>
i know the tactic all too well
<f_>
axtlos: I suspect both
<leah>
i've systematically reviewed metux's changes since 2024, since i became interested in it when phoronix did an article about it
<f_>
I mean, redhat isn't perfect, but also a number of people like to point fingers at redhat because of all the issues they got
<Ariadne>
i am sorry but if you do not know what the ^ operator in C does, you have no business maintaining a C program that requires elevated privilege like the X server
<leah>
some of the changes are actually good. but mostly seem to be code cleanup
<Ariadne>
like, holy shit dude
<leah>
and massive amounts of re-factoring that lead to some regressions
<f_>
All too often I see people kicked out on CoC grounds pointing their finger at redhat
<f_>
vaxry did it before, and now metux is doing it
<axtlos>
i cant wait to see the CVEs xlibre is going to produce
<leah>
i'm going to share some wisdom with you all
<leah>
code==bugs
<leah>
fewer lines of code == fewer bugs
<f_>
Xlibre™ x^2 == pow(x,2)
<leah>
on that basis, both xorg and xlibre are a liability.
<leah>
xwayback is more viable
<f_>
Xlibre™ we like cves
<leah>
focus on wayland. and maintain an x11 shim for xwayland. that totally computes, as far as i'm concerned.
<whitequark[cis]1>
also "more recently touched code" == "a lot more bugs"
<leah>
yes, quite
<Ariadne>
also "more recently touched code by someone who doesn't know what they're doing" == "infinite CVE unlock"
* axtlos
desparately trying to understand xdg-output for xwayland
<f_>
AFAICT xorg seem to be trying to not touch x11 unless needed
<whitequark[cis]1>
yeah
<leah>
i don't care how good you are
<Ariadne>
axtlos: i think that is a Son_Goku thing :)
<leah>
the moment you touch something, it breaks
<whitequark[cis]1>
i don't think 2025 is a great year to start a big xorg refactoring
<leah>
no matter how good of a hacker you are
<axtlos>
Ariadne: huh what do you mean?
<f_>
whitequark[cis]1: oh no my xrandr broke on xlibre :(
<whitequark[cis]1>
f_: meow?
<Ariadne>
axtlos: neal seems to have a handle on how we can get xinerama working right
<axtlos>
oh i see
<f_>
(xrandr breaking is one bug metux introduced in xorg before getting kicked out)
<axtlos>
xinerama is very far for me right now lol, im just implementing a client xdg-output for xwayback
<f_>
anyways at the end of the day .. xlibre is just not viable
<f_>
let's focus on wayback instead ;p
<axtlos>
ah i mistyped in my original message, xwayland = xwayback
<axtlos>
both boil down to properly implementing some form of output management, but the issue itself is different
<f_>
alright
<navi>
i started writing patches for using socketpair and removing the usleep hack, though it's kind of moot to do any of that before splitting Xwayback and wayback-{compositor,session}, and since someone's already doing that, i'll wait
<axtlos>
i already have the socketpair stuff done
<axtlos>
havent started working on wayback-session yet though so i havent tackled the usleep stuff
<navi>
neat
<ToyKeeper>
whitequark[cis]1: FWIW, I used DPI scaling for GTK apps for a while, and it was really easy as long as I didn't need fractional scaling. I simply put `export GDK_SCALE=2` in my .Xsession and everything worked.
<whitequark[cis]1>
there is no "simply" when it comes to DPI scaling, unfortunately
<whitequark[cis]1>
the problem i hit is how this combines with Xft DPI
<whitequark[cis]1>
what does xrdb -query say for you?
<ToyKeeper>
Eh, xrdb -query dumps out all my .Xresources. None of it is related to widget or text scaling.
<ToyKeeper>
The gdk scale thing just doubles the size of GDK/GTK widgets and text, so things will look the right size on a screen with 2X as much resolution.
<ToyKeeper>
In my case, it was on a TV that I wanted to be able to read from far away.
<ToyKeeper>
Re "the only people who would be "forced" to do anything about it are window manager and desktop environment developers" earlier, that's not the case. Far more people are affected, because Wayland and X11 are not interchangeable flavors of the same basic thing. A lot of stuff I rely on daily is categorically forbidden in Wayland, so I can't just switch to a new tool... because there are none.
<whitequark[cis]1>
i know how GDK_SCALE works, yeah. i tend to hit issues where text and non-text (UI elements, icons, etc) are scaled in a way that they don't match each other. it's quite annoying
<whitequark[cis]1>
i ended up having export GDK_DPI_SCALE=0.6 GDK_SCALE=3 in my profile, which i could only get to work by trial-and-error