#๐ฅ-vesktop-development
1 messages ยท Page 32 of 1
ur so mad lol
no i actually am
im pushing 2000 commits to ur repo
just tried, it broke
stop lying
i dont like that the auto updater script runs on a platform where it can't function but wtv
@worn harness i think we should move vesktop to /opt
why husk
thats not the proper path
how
it was there at first but then i fixed it later
why is it not right
its the lazy way
@humble mortar i had an idea
each package has the same desktop file so couldnt we just put one in the root of the repo and use that instead
and vesktop.sh
@vernal lintel can u retroactively push .pacman binaries for 1.5.5 or does it have to wait for 1.5.6
no afaik
@humble mortar does symlink work
im not sure
i dont think aurpublish has any symlink handling
afterPack wouldnt be needed if vee just merged the stupid icon PR
im literally sobbing
i wanna change a line to see whats happening there
is vesktop-bin even maintained by us
no
guhh steal vees login
it literally is ๐ญ
..,.
it would be wrong to install it to /usr/share because /usr/share is supposed to be architecture independent
wha
electron IS architecture independent
arch electron packages are in /usr/lib
where vesktop is
if you only install the .asar file and nothing else sure
the only thing the pkgbuild puts in /usr/share is the desktop file and icons and license
if you install the entire electron app including the binary then yes
/usr/lib/vesktop is where its installed rn
and is that just the .asar file or also the electron binary
oh my goddd
well system electron is /usr/lib so idk
it works leave it alone
well im asking because the .pacman builds to /opt
/usr/lib is only for libraries / object files
ill have to do some moving around
if you want to put it in /usr then the correct way would be to split it up into /usr/share, /usr/lib and /usr/bin
but that's insane and not how electron works
that's why electron packages make most sense in /opt
that makes sense
putting an entire app in /usr/lib is just wrong, you're violating spec
but it also doesn't really matter much
.
it is not the entire app
/lib not /share
brother this is the entire app in /usr/lib
those are libraries are they not
/usr/lib is only for libarires (.so files and secondary binaries)
vesktop binary right there
i'll leave it for now i guess
if someone complains it can be moved
if there wasnt a vesktop.sh script then vesktop binary could go in /usr/bin
if you wanna put these files in /usr then the "right" way would be something like this:
/usr/share: resources, locales, resources.pak, icudtl.dat
/usr/share/licenses: LICENSES.chromium.html, LICENSE.electron.txt
/usr/lib: vk_swiftshader, vesktop, *.bin, *.so, chrome-sandbox, chrome_crashpad_handler
/usr/bin: vesktop.sh
but electron wasn't made for splitting like that
that's why you put it in /opt cause opt is like the misc app folder
@vernal lintel do youknow how to debug a specific line in electron-builder
node_modules electron builder is the built output
different
make a pnpm patch
does that work if its compiled already
yeah
if it doesn't have to be inside the chroot then you can literally just edit the file in place
u just patch the js
assuming it's a js file
i cant
yop just pnpm patch electron-builder
but i dont have a patch
what r u even trying to debug
cookie that just gives me the built binary
show error or whatever
im investigating somethiong on my own
huhh?/..
it doesnt give u a dir with all the js files?
no
guh?
saves bytes!!!!!
i wish npm banned pushing minified js
you love
now you know what vencord plugin devs go through
cookie will make plugins
my enemy
what does hoisted mean
ur hoisted
actually we might not actually need it anymore
because we use esbuild to bundle dependencies
if i comment out node-linker=hoisted then it builds without npm fully
can we do that
try it and see
in the past we bundled some dependencies and without hoisted, it bundles symlinks instead of the actual files
it work
but we don't do that anymore i think
i think its working, i installed and im on it right now
i know ;)
how
๐ดโโ ๏ธ ๐
veeeeee how should i format itt
not like that
6502 etc
giant button
you will design a Download on AUR button
dont be lazy I did that too a while ago
i made the UserScript button
@vernal lintel will you gonna remove hoisting from the npmrc
will you gonna
im looking for an existing design i can steal
AI generate
ill chop off my head
SOOOOOO UGLYT
List of package versions for project vesktop in all repositories
why does winget have them all seperate
so insane
ohh thats so husk
why did they think this was good
its so stupid
adobe ahh
@humble mortar sniffs crumbs
chatgpt has been going for minutes, why is it so slow
since when could it run react
that looks so asss
@humble mortar sooo judgemental
yupp
u look ass ๐
rude
chatgpt slower than cookie
Ill probably kill u if u use that badge
we gotta kill Kylie first
Kylieeee ๐ญ
@vernal lintel make svg renderer plugin
๐ฐ๐พ๐ฑ๐ฎ๐ช๐ช๐ช๐ช๐ช๐ช๐ช๐ช ๐ญ
๐ฐ๐พ๐ฑ๐ฎ๐ช๐ช๐ช๐ช๐ช๐ช๐ช๐ช ๐ญ
ok i brb
we win
i just realized what this is
ur so insane
@humble mortar dms
how
what would it link to
you don't go to the aur to download tihngs

This was the only issue I could find relating to failed startups that wasn't closed as a dupe. Maybe I should risk filing my own issue to see it closed as a dupe too.
Content
The UI was super laggy today and usually no matter what system i'm on I encounter a sort of memory leak where 10 ish minutes into screensharing I need to turn it off and on again otherwise my fps drops drastically. In the 2nd most recent update this was even worse to the point where my games straight up frameskip so I thought updating would fix it, and now I can't even screenshare at all.
My pc specs :
Manufacturer: LENOVO
Model: LNVNB161216
Form Factor: Laptop
Processor Info...
please use our support channel to get help
3c87c89 README: add homebrew as installation method (#1... - khcrysalis
vesktop now on brew :D

https://github.com/Vencord/Vesktop/issues/803#issuecomment-2771256055
@kode54, I solely wanted to know what relates that problem to this. I ask because although I presume it's the fact that you need to disable Wayland support in Electron, do you see the same error message in GNOME Abrt? It shouldn't matter much, but it helps to separate that problem from your other one about your webcam.
I'll estimate that your webcam's problem and the poor compositor support are separate. I suggest that ...
@brazen kite are you having any issues with the shortcuts pr right now? what was this about exactly?
i did test on windows btw it was seemingly fine (it also did notably need visual c++ for some reason)
It was due to upstream issue https://github.com/electron/electron/issues/44543 which has now been fixed and backported to Electron versions 33, 34, 35.
Vesktop will need to update their Electron dependencies for this to be fixed, so you may need to wait a little.
@tidal quarry guh
you dont click the link
and visit the aur
you use a package manager
husk
?
no you dont
what
aur helpers arent official
guh
@humble mortar you were the one working on this right
I'm just using vesktop-git
nothing wrong with the package
its on by default in pkgbuild conf
oh hm
sorry for bothering
np
ywww
OK uh let me try to explain
so yk the issue where the discord keybind manager can't actually register mouse inputs
even if I were to register the bind through the CLI on windows, there's no apparent way for me to be able to go and actually bind my mouse to the bind
does that make sense?
like on KDE, for example, I could go to the keybinds manager and do it through there
so what I was asking was if it's possible to be able to pass in the key you want to bind to into the CLI
KDE doesnt support mouse shortcuts like that
it's just an example
windows doesn't have a central place to do that is all I was trying to say
oh the cli triggers the shortcut doesn't have anything to do with bindings
I figured
I was trying to mess with it to try and figure it out
so do I need to like make a script that executes the CLI thing when I press my mouse?
what is discord using for key capture in the bind menu?
yeah pretty much
electron CAN detect mouse inputs so I'm a bit confused why their impl doesnt
i think the official discord app's key capture thing is actually from the native part
we'd probably also have to rewrite it if we want it to capture mouse and stuff
renderer doesnt have access to more than the 2 side buttons
a native impl is def needed
i dont have a mouse with more than two side buttons so I can't test ๐ช
actually I could prob simulate it with xdotool key
we could also just tell people to bind stuff to the cli.....
binding to the CLI with a script is ass cause no push to talk
my two side buttons doesn't work
Sorry about that. The webcam issue doesn't seem to be happening with official client. I also had to install a patch against OBS Studio to work with v4l2loopback 0.14.0.
My environment is labwc-git on Arch, with AMD Radeon RX 7700 XT, and a Logitech C615 camera that prefers to use MJPEG for its higher resolutions and frame rates. I'll post a different issue for the camera one.
I do not receive a notice when the startup fails, except for some EGL errors in the terminal if I launch it there. I...
Content
Currently, as far as I can tell, the 1080x1080 image build/icon.png is downscaled to 256x256 using some non-nearest neighbor scaling algorithm. Only having a 256x256 icon is problematic as it can be scaled to as little as 16x16 and it looks pretty awful:
I strongly suggest that a proper ICO file with multiple different sizes, [ma...
Content
This is a pretty major issue. Direct messages do not work. I'm not sure why it hasn't been reported. It seems to be tied to the new UI change.
I'm on Linux Mint and running Vesktop through flatpak. Click to copy version doesn't work or I'd paste it in here.
Solution found here: https://github.com/Vencord/Vesktop/issues/1135#issuecomment-2741203553 This fixes the issue, but I imagine others less savvy are having this problem too so this is probably a major issue for Vesktop...
what
"i'm not sure why it hasn't been reported"
literally links other issue that reports the same issue

im confused xd
Duplicate of #1135 (I'm honestly kinda confused why you mentioned that issue and having found the solution there and still opened a second report of the same thing xD)
Anyway, this is just a Discord issue because they're dumb and don't test their experiments on their web app before pushing it onto poor users. Not really much to do from our side, Discord has to just fix it
๐ญ
why is that even necessary
not needed on linux & windows
why do they specifically block it on mac
but not windows
match: /"discord\.com"===location\.host/,
thats so stupid lmao
lmao
why does mac even allow that
how? @humble mortar
XFCE
has working badge
what's description (aka error.code) here? just a number? it should be a user friendly message that even tech noobs can understand
console.log(`Failed to load ${url}.`, `Retrying in ${retryDelay}ms`);
setTimeout(() => loadUrl(url), retryDelay);
webPreferences: {
preload: join(__dirname, "splashPreload.js")
}
error.code is the description
error.id is just the number
log the entire object to see
no idea
find ur lib folders for the libunity so
it might be a fake badge that does the count for the number of de notifications
I don't think so lol
it's XFCE
I use dunst for notifications
I'm honestly kinda confused why you mentioned that issue and having found the solution there and still opened a second report of the same thing xD
#1135 Doesn't mention the issue in Direct Messages, but in Nitro, so it could be a different issue. I wasn't sure if I should add to 1135, but it is marked as closed without any fix, so adding to it seems irrelevant.
Not really much to do from our side, Discord has to just fix it
Is there a valid reason Vesktop should ever prompt to open the ...
Content
when testing the master branch of vesktop the maximize and close buttons do not do anything when clicked. and when adding a print statement to the icp events only the minimize event would ever be printed. In the build I compiled a few weeks ago with the PR #1104 maximize and close worked.
https://github.com/user-attachments/assets/97d8cb87-b492-40da-b193-c91334ceef2d
known and a fix has been made and will soon be released.
for now just use the native titlebar.
also:
I am a developer, was just helping report an issue I experienced.
-------- Original Message --------
On 4/5/25 14:57, nin0dev wrote:
known and a fix has been made and will soon be released.
for now just use the native titlebar.
also:
IMG_4516.jpeg (view on web)
โ
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you authored the thread.๏ฟผMessage ID: @.***>
๏ฟผnin0-dev left a comment (Vencord/Vesktop#1148)
known and a fix has been made and will...
@vernal lintel update the wording to maintainer when
this makes it unclear because then "Hi I Vibe Coded A Website I am Developer"
Turning on NoDevtoolsWarning fixes this (for some reason)
Glad I found this, disabled the plugin a few weeks ago due to myself not caring about the warning.
i think we are basically ready for a merge. marking this ready for review.
if anyone finds any bugs or has feature requests please do tell
it only took a year and then some wooo
also tested stuff on flatpak and appimage everything works seemingly fine
how do i force the x11 shortcut method
you can't on wayland unless you go into the code change the return conditions at the beginning of each function in the patch
do you wanna do just the ui or also actually register the keybinds?
fatpak
xwayland
should probably be an env var for that case
Uhhhhh I mean that wouldn't work except in other x11 apps but yeah it checks the XDG_SESSION_TYPE and WAYLAND_DISPLAY variables
Also experiencing this on NixOS 24.11 stable in gnome (X11) where my webcam lowers to 640x360
sadan would die
Can also confirm that, after building from source, screen sharing a window still results in other people hearing an echo.
Having same issue, have built from source and used the release version to no avail.
@tuxinal Looks like Cmd on MacOS shows as "UNKUNDEFINED":
<img width="704" alt="image" src="https://github.com/user-attachments/assets/de9f9e2a-c324-4e98-bbf2-f88a58dc13d0" />
Additionally, custom keybinds don't seem to be working at all, whether they be unkundefined or have normal modifiers (control, alt, etc.).
Thoughts?
On Windows this still happens, When i search Deep Links i only find this:
set it to Not Eligible didnt help. Did i do sth wrong i didnt find the right setting
@ryanccn still doesn't work for me :(.
Vencord: v1.11.8 โข d753478 (Vesktop v1.5.5) - 4 Apr 2025
Client: stable ~ Vesktop v1.5.5
Platform: MacIntel
โ ๏ธ Activity Sharing disabled
@ryanccn still doesn't work for me :(.
@comfortablemilk4454 Are you building the app from source? The fix I made hasn't made it into a release yet.
oh thats why lol. mb thought it made it into a release!
On Windows this still happens, When i search Deep Links i only find this:
set it to Not Eligible didnt help. Did i do sth wrong or i didnt find the right setting. @Vendicated
They removed the deep links experiment
Same issue in wayland compositors.
Tested with Arch & NixOS.
We should have at least an option to bring official discord tray icon.
For now, patching Discord with Vencord is the best option if you have this issue.
43f59ce use Vencord's addPatch instead of manually addi... - Vendicated
we should move most patches from vesktop to the vencord codebase
VesktopFixes core plugin
not rn tho
either that or we need some way to hotfix vesktop renderer
just so we can fix broken patches without needing an entirely new vesktop release
RCE support for Vesktop
Remote update patchs
@vernal lintel re:
don't think this makes sense to be fixed upstream since we just repack the AppImage
if this was about my comment it was because we made desktop file changes upstream that then trickled down to flatpak, and we hadn't had any complaints about it since we did that so i wasn't sure if it was still needed
i cant quite remember but it was like
the name of something
ill see if i can find the commit that comment is referencing
i really hope the fucking flathub autoupdater works now finally
after my previous fixes
it did this flathub/dev.vencord.Vesktop#49
but thats 1.5.5
sorry i would open that link but windows explorer decided to crash so now i cant open links
i really hope it prs 1.5.6 soon
should we ask the flathub maintainers if they can add our webhook now
or should we add a workflow that does it manually trol
no response or no?
also i think they have security that prevents this
no response
lost in the wind
i bumped it and same result
should I open an issue on the flathub repo
not a bad idea tbh
yeye
Is this still an issue?
@vernal lintel no (at least for me)
btw not sure if I'm going crazy
but after upgrading to vesktop 1.5.6 text all over the app seems a bit blurry...?
probably some electron shenanigan
nevermind
I have a window rule applied to vesktop
so maybe it does appear
I had a similar issue with ruffle, the devs fixed it there
probably discord issue, see #๐ฅ-vesktop-support-๐ฅ
VICTORY
for some reason it made two prs
Vesktop: Update dev.vencord.Vesktop.metainfo.xml to 1.5.6
Vesktop: Update Vesktop.AppImage to 1.5.6
๐ค This pull request was automatically generated by flathub-infra/flatpak-external-data-checker. P...
weird
oh wait
it's cause I made a commit
one of them is outdated
ig it automatically made a second pr cause it noticed the first one was outdated
flatpak install --user https://dl.flathub.org/build-repo/176756/dev.vencord.Vesktop.flatpakref
time to test
look at this
the resolution is undefined when it refused to load
when hardware encoding is enabled
i wonder if thats related
sometimes it fails somtimes it takes a whie
weird as hell
[RTCConnection(111, stream)] Go Live Media sink wants: {"116989":100,"resolutionWants":{"116989":{"width":0,"height":0}}}
so the failure to load is cause its timing out
so why does it take forever to load
almost every time the hardware encoded stream fails that log happens
so for some reason the RTCControlSocket is closing which then triggers a retry timer
but that timer is too long and the stream times out
so why is the socket closing on hardware encoded streams
encoded webcam streams load just fine
@vernal lintel https://github.com/Homebrew/homebrew-cask/commit/8828875750c633c588b196b27b7bab15dd299fa2 brew automatically updated it
so theres your answer hehe
yay!!
index.ts: Lines 39-51
const settings = definePluginSettings({
// This needs to be all in one setting because to enable any of these, we need to make Discord use their desktop context
// menu handler instead of the web one, which breaks the other menus that aren't enabled
addBack: {
type: OptionType.BOOLEAN,
description: "Add back the Discord context menus for images, links and the chat input bar",
default: false,
restartNeeded: true,
// Web slate menu has proper spellcheck suggestions and image context menu is also pretty good,
// so disable this by default. Vesktop just doesn't, so we force enable it there
hidden: IS_VESKTOP,
}
});
thx for clarifying :p
that seemed like a bug because there is "Settings" without any settings
maybe just hide the cog wheel instead of the menu inside the cog wheel
likely this
interesting; is there a way a patch could be made for vesktop to disable this new flag?
^ @simple willow
just launch Vesktop with the disable flag for it??
what would that be? ๐
dunno
found it
--disable-features=FontationsFontBackend
I have not compiled and tested Vesktop with this new change, but it should work just fine.
I have tested the flag and I can confirm it works.
Restricting this to Linux-only as I don't know if this problem is present in other platforms.
where'd you go to find this
https://github.com/googlefonts/fontations/issues/1359 has https://issues.skia.org/issues/396360753, and the latter has https://issues.chromium.org/issues/375219950 (text is "issue 375219950" in the Skia page)
and in the chromium issue there is this: https://issues.chromium.org/issues/375219950#comment29
insane
i wonder if the electron people would be up for backporting the fix
ekse we'd need to wait for electron 36
// FIXME: Remove when bumping to Electron >36
Linked issues have people on Linux, MacOS & Windows
disabledFeatures.push("FontationsFontBackend");
@Covkie Done. Thanks!
we'll probably need to wait
u dont need all that yap in the comment
not platform limiting means its obvious it affects all platforms
and it IS fixed in 36 cause that electron version has the chromium version with the fix
@vernal lintel pls take a look at https://github.com/Vencord/Vesktop/pull/1149 to make a new hotfix release ASAP. thanks!

no that won't happen
if it really has to be a hotfix then it should be added to vencord instead. Making an entirely new release for that is insane
but so far only 2 people complained
you can just pass the flag yourself
I didn't know it was possible to add new runtime flags to Vencord while also affecting Vesktop
I thought they were separate
this is already fixed on electron 36, but it's still in beta
yep, did that already
Also experiencing this on Ubuntu 24.04 latest patch as of this moment with nvidia 550 drivers (official & installed via ubuntu-drivers) and with 3080ti & quadro p600. I thought it was a driver issue with the v4l2 usb attached to my dslr "webcam" then noticed it displays properly in other applications and the official version of discord. Hopefully it can be resolved soon as this is a major downside to this application now...
@humble mortar dissertate
i mean theyre right
its a PR that should exist if vesktop ever tags another version on 35
but its not high priority
next tag 2027
@vernal lintel what happened to vesktop troubleshooting wiki
lazy
nope
confirmed, @vernal lintel actively makes the linux experience worse
that's why he invented vesktop

you will push flag that fixes an edge case as a single commit tag
@vernal lintel I'm hostile forking vesktop to merge all the shit you won't (evil)
yop
Content
Discord has updated their screen sharing to support the new macOS ScreenCaptureKit API a while ago. It would be better to use this than the custom share screen used in Vesktop.
Discord's is the one that's custom. Vesktop's is just chromium's picker.
No, we will not use Discord's proprietary implementation, use the Discord app for that
Content
An assertion that was added here causes vesktop to not start.
Output:
APPIMAGE env is not defined, current application is not an AppImage
Vesktop v1.5.6
checkForUpdatesAndNotify called, downloadPromise is null
[4132:0415/194107.072039:ERROR:wayland_event_watcher.cc(47)] libwayland: xdg_surface#35: error 5: the effective window geometry is empty
^[[Afish: Job 1, 'vesktop' term...
not an issue in our code, nothing we can fix. it's an issue with any of electron, chromium, or (most likely) your system
Content
Hey I noticed Vesktop is having graphical Window issues and I had the same issue on my software and wanted to report how to fix it if you did the same workaround for KDE
I have a issue report and documented it on my website news post
https://gitlab.com/linuxbombay/crunchyroll/binaries/1.0.4-4/-/issues/1
https://coreybruce.netlify.app/news/software-bug-fix-for-gnome-12-4-2025.html
On package.json for example if you add -ozone-platform-hint=auto after electron . it will c...
Not a Vesktop issue, Upstream issue that's now fixed: https://github.com/electron/electron/pull/46624
Not a Vesktop issue, Upstream issue that's now fixed: electron/electron#46624
Ah wow I didn't know that the issue was this big of an issue, glad it was fixed and thanks for informing me! ๐
I can also tell that my voice through the "Mic Test" sounds like absolute garbage. I'm unsure whether the noise suppression applied there is also applied to regular calls.
I have noise suppression off since I use external rnnoise anyway. Echo cancellation still on.
what a cute icon
true
After some long trial and error (while also discovering the Patch Helper thing in setting kinda late .-.) I did the patch thingy! It was fun to do, however I really couldn't think of a better way to apply the change other than saving the value in the
Settings.storeobject whenever it is toggled and then check the value on the startup over atsrc/main/index.tslike before. If you can think of one you can tell me and I'll check it out. :P
The switch doesn't seems to do anything, even...
not sure how i got it into this state.
I ran Vesktop through the terminal this morning and global PTT worked just fine. didn't even log out, but when i tried it just now, it didn't work. Deleted the shortcut from KDE System Settings but nothing changed. There's now no trace of Ctrl+` in my ~/.config, but it still prints this error when i press Ctrl+`
๎ฐ git describe --always
12561e3
๎ฐ npm run start
npm warn Unknown project config "node-linker". This will stop working in the...
Content
The README is a bit outdated:
Might as well link to https://vencord.dev
Uh... how? OpenASAR? Please specify and include a link so I know what I'm installing...
Th...
Also, what other specific mods do you guys do? Newer electron? HW-accel? It would be nice to know this in the readme
Just did an install from Flatpak and the issue seems to be resolved now
Content
Operating System
Linux Mint 22.1 Cinnamon
Linux Only ~ Desktop Environment
X11
Package Type
Flatpak
What happens when the bug or crash occurs?
The following error message appears. Closing the error message sometimes loads the full image and other times Vencord crashes to a blank window. Either way the client has to be restarted to access Vencord again.
A JavaScript error occurred in the main process
Uncaught Exception:
Error: ipc: Disallowed host cdn....
...how so? On regular Discord I'm able to click on the same image and zoom in to full size without any issues.
now try on discord in your web browser. use our support channel like the issue template tells you
Causes computer audio to be muted when screensharing with audio
c7d830c Revert loopbackWithMute to loopback (#1155) - Glitchtest51
tbf the form is quite unclear
since the random that vibecoded some next js app is a "developer" and to his eyes can open an issue
I am also running into this issue on KDE/Wayland while using the flatpak. Vesktop goes completely blank, and I have to manually restart it and rejoin the call. It's inconsistent, but it usually happens around 30 minutes into streaming.
I have this exact problem as well
Me and my friend both share our screens while playing games, he's using Fedora 41 with Nvidia 4070Ti Super and AMD CPU (can't remember the model) and he has no issues at all, however, for me it freezes a few minutes into sha...
@vernal lintel another yapper in vesktop support about vesktop exploding cause no loadurl retrying..
you will merge the fix
vgif hammer @humble mortar
NO STOP HITTING HER
Electron kindly spams this for us
@vernal lintel ogrb
still do like ```
Failed to load canary.discord.com: ERR_INTERNET_DISCONNECTED
long ass message
is there an error map
love
logs include the real description
just do this
no one who needs splash to tell them the problem will be looking at logs
evil
i love the word evil
wonderful
honestly id rather just Failed to load Discord: description
if it fails with a uri payloud the actual error description could leak off the splash
ill do this
ok i pushed
OOPS
i may be stupid
i forgot to save
Content
After inspecting an image, I tried zooming it in directly from the Discord client (I didn't open it inside a browser tab) and then I got an error and the renderer crashed. From what I can tell, this new feature (zooming in directly from the Discord client) came with Discord's new UI update.
This is the what the error says:
`Error
A JavaScript error occurred in the main process
Uncaught Exception:
Error: ipc: Disallowed host cdn.discordapp.com
at pc (VCDMain:23:571)
at IpcMainImpl...
Shouldn't this be reopened? Since it's a prevalent issue.
Content
It would be nice if the "about" page from the tray menu could show the current version.
temporary fix found on Vencord's Discord server (thanks for whoever posted this)
add this to your QuickCSS
.imageWrapper_af017a > a {
pointer-events: none;
}
@vernal lintel
like arch repos?
extra yeah
interesting
i wonder why the system electron choice
official discord packages and stuff don't do that either, do they?
no they dont
i mean it doesn't matter that much but electron does have issues pretty often like the crashing in the past (v31 iirc) and the weird border rn on v35
yeah
and the recent hw acceleration flag changes
I'd be okay with the package bumping minor electron versions in the build and shipping that bundled
but system electron is too volatile
in fact I was thinking of doing this because of the border issue
ah did they fix it?
wonder if that guy even uses Discord
or if he's just interested in packaging it due to popularity
cool
we'll see what they say about joining the server to talk about it
ughhh still no response from discord about the hardware encoded streams refusing to load issue
I don't think anyone has ever successfully reported a bug via discord support
there was the bitrate issue that was reported and fixed
was @hybrid fjord I think
on the other hand, if they use independent electron version they can push electron fixes independent from us
which also has its upsides
I'm not against using system electron
it's only bad when it creates issues for us (aka people complaining to us over something we don't control)
which is why i think the package bundling electron that can be changed when the package builds is better
so it can be reverted or bumped
The switch doesn't seems to do anything, even I turned the hardware acceleration switch off the screen artifact still appear
If disabling Hardware Acceleration under Vesktop Settings doesn't fix the artifacts, then this PR won't do so too. For some people the cause of artifacts when screen sharing is video hardware acceleration, this PR is made so those people can disable only video hardware acceleration while not disabling all hardware acceleration.
However video hardware accelerat...
@charred geyser
@vernal lintel this is their acc
v+ prog 1363481780979503127
Is your new account for privacy or because you don't use discord?
Hey
I have an account using my personal email, but created this one now since this is arch business
Not a regular user any more tho, no
Happy to hear your input on vesktop in official repos
draft PKGBUILD, for reference: https://gitlab.archlinux.org/-/snippets/3774
Is this getting merged yet? This is a must have for dual screens with different resolutions due to how Wayland sees multiple monitors as one :(
I'd prefer someone who is a regular user to maintain it simply because of how things might break and the like
Sure, I see that PoV
Just generally, what breakage are you expecting? Electron packages are in my opinion very stable, as long as one uses the same major as upstream. But I can see how vesktop might be more complex
The ambition is to reduce issues for you, not the opposite ๐
can you read over the pkgbuilds in https://github.com/vencord/pkgbuilds
why start fresh instead of using these existing packages that we support
I'd also prefer NOT to use system electron. (you can read why if you scroll up)
youre also not using the right desktop icon
both are fine honestly, doesn't matter much ๐
what kind of delay is expected between us publishing a new release and it getting updated in the official repos?
I'm also curious if we can setup an action to publish the electron builder .pacman target instead of using the pkgbuild
vesktop usually doesn't have that frequent updates but things are a bit volatile due to depending on discord
there's always the possibility that things could break badly if discord makes major changes and we need to push an update quickly
yop
I had a quick skim of the package and it seems to make sure it uses the same electron major as upstream
minors don't necessarily have to match, it's probably even beneficial to use independent minor to get new electron fixes quicker
like the window border jank on gnome rn
it uses system electron yeahg
system electron on arch is usually latest minor for the packaged major versions
electron is just a link to the latest electronXX major
we should copy over some of these dependencies from the discord package
same optionals and libnotify (same story as official discord app, notifications with no libnotify will crash the app)
those will be brought in if using system electron
look at vencord/pkgbuilds/vesktop(-electron)
libnotify won't
yes look at the packages ive already written ๐ญ๐ญ๐ญ๐ญ
ive been through this whole rigmarole
would be better to use the desktop build generated by electron builder instead of manually maintaining a separate one
Generally pretty quick, I check daily for updates. But of course, it adds a middleman
how difficult would be to have myself as a co maintainer as youre not super active on discord?
I'm not really familiar with official maintainer guidelines and the like
do you mean the pacman target?
referring to the .desktop file
Depends on what you're comparing to x) It's an application process, generally requires a few years of packaging/involvment
That might a good idea, haven't looked into that
youd probably need an afterPack script to do that on the directory build
like what I'm currently doing for the icons
the dir target for some reason doesn't generate a desktop file, it's dumb
there's a pacman target that generates a tar.xz pacman package, that should contain one I think
for the flatpak we just extract the desktop file from the appimage
yeah I was gonna say just make it build the .pacman then extract it
it'd increase build times but since its only a build once distribute to everyone it doesnt really matter
so no easy way to have me just for the one package?
hm
could we maintain the official extra package on a clean branch in vencord/pkgbuilds then?
The source of truth for the PKGBUILD would be https://gitlab.archlinux.org/archlinux/packaging/packages/vesktop#
you're acting like we need to change it up all the time :P
just contribute to it now to make sure it's all good without issues and that should be all that's needed
we can always get in touch again later if there are any changes needed
oh I completely forgot I can send patches via the gitlab
yeah I'm overcomplicating it
oopsies
Just to circle back a bit on the discussion, the reason for adding to extra as I see it is that neither you nor the users would have to deal with build issues ๐ If you'd prefer to continue via the AUR, I'm not going to overrule that
I gotta log off pretty soon, I don't think there is any urgency to do anything really, so we can iterate on the PKGBUILD and also make sure you get accounts on gitlab.archlinux.org before doing anything
You'll need that for issues/patches
yeah i already have one
thanks!
I'll look at doing the .pacman target extraction stuff
Thank you! Good discussion all around ^^
adding it to extra is definitely a good idea, thank you for taking initiative
it just needs to be done right, I still have ptsd from when one of the (unofficial) aur packages bumped to a broken electron major and we got tons of reports over that ๐ถ
Totally get that and empathize, AUR can be a bit chaotic at times ๐
i don't think it's fixed
just reported and stalled
it's still occurs
@charred geyser
The switch doesn't seems to do anything, even I turned the hardware acceleration switch off the screen artifact still appear
If disabling Hardware Acceleration under Vesktop Settings doesn't fix the artifacts, then this PR won't do so too. For some people the cause of artifacts when screen sharing is video hardware acceleration, this PR is made so those people can disable only video hardware acceleration while not disabling all hardware acceleration.
However video hardware...
Yeah I meant it that way, btw I don't know how github work does the pr you mention push in to the official repo? I am using the official repo while commenting. switching off hardware acceleration under vesktop setting does fix the artifact, however under video section the switch doesn't fix the artifact. so my workaround right no is to disable hardware acceleration under vesktop setting which make my whole vesktop laggy.
In a pull request the code isn't on the main branch or repo and wil...
--disable-features=AcceleratedVideoEncoder
thank you I will try the argument
@vernal lintel you nuked issues?? ๐ญ
ye i disabled them
insane
tracking issue for more customization gootbye
people should post in support

gonna open issues by PRing a space in the readme
we can turn them on again in the future
thats what they all say...
next youll archive the repo..
vesktop going closed source!!!
TRUE
lmfao
we just need to make vesktop wiki a reality
force me to tomorrow @humble mortar
im on vacation tomorrow
and every other day that ill need to help make a wiki
sorry!
how the hell do i fork on the arch packages gitlab
husk can i not do it cause the repo is empty
evill
@charred geyser
ill force you oomf
Thanks for the patches! Busy with work mostly now during the weekdays, but Iโll get to it eventually ๐
@vernal lintel do wiki
@humble mortar
โ ๏ธ Arcane is banned.
me if you care
looks hideous...
did the blurfonts stuff get fixed in electron yet
oh my god ๐ญ
this is that gnome window border bug
its fixed in electron 36 unless they backport the patch
is the fix just a chromium bump?
ya
doubt they're gonna backport that unless google backports
does google backport to old versions?
or do they only have exactly one latest version
they back port some things
idk if slightly blurry fonts is a priority for them ๐ญ
i am so thankful for github and friends so that I don't have to use this atrocious interface https://chromium.googlesource.com/chromium/src/+log/136.0.7103.25..136.0.7103.33?n=10000&pretty=fuller
i feel bad for google engineers
what do companies like google even gain from using their own git server
https://skia-review.googlesource.com/c/skia/+/952636 this is the blurry fonts fix btw
control
most have github mirrors anyway so like
is that true????
holy fuck
95% of google slop is in a mono repo
how tf do you even organize dependencies like that
this is why you have to use some google program to clone code instead of git clone
hop on cloning from the github mirrors
Typical use involves Clients in the Cloud (CitC). This system utilizes cloud backend and a local FUSE filesystem to create an illusion of changes overlaid on top of a full repository. This approach enables seamless browsing and use of standard Unix tools without explicit synchronization operations, thus keeping the local copy very small (average size of a local copy is less than ten files). All file writes are mapped to snapshots thus permitting restoration of the previous states of the code without explicit snapshotting. Due to the always-connected operation, CitC allows easy switching of the computers as well as sharing the modified code with other developers, the automated build system and testing tools.[7] As a result, the majority of Google developers practices trunk-based development with no personal branches; the branches are mostly used for releases.[8]
๐ญ
oh wow using a FUSE overlay is actually smart wtf
cloning the whole thing would be terrifying
WHAT THE FUCK
as of 2016, the repository was storing 86 terabytes of data comprising two billion lines of code in nine million files (two orders of magnitude more than in the Linux kernel repository). 25 thousand developers contributed 16 thousand changes daily, with additional 24 thousand commit operations by bots. Read requests each day are measured in billions.[6]
2016
google literally reinvented old svn style takeout systems ontop of git
genuine insanity and not the good kind
Soon Vencord will have these stats too
D
i've changed the way keybinds are recorded in the settings page. now it records things through venbind instead of electron (which does mean it's now even more broken on macos since there is no mac support in venbind but if anyone were to make a pr for venbind for mac support it should be slightly easier to fix stuff (@Razboy20))
this is also related to capturing mouse button stuff
also rewrote the whole ui lol
SOMEONE PLEASE FIX MY UGLY ASS REACT CODE IDK REACT OR UI DESIGN
Thanks for all the detailed reports. I will look into this soon!
I will clean-up and partially rewrite some parts of venmic now that rohrkabel v9 has been released and will make sure to take down-mixing (or rather proper multi-channel mapping) into account
I have decided to not pursue this further.
In case the system provided GLIBC(XX) is not recent enough, one should use Flatpak as that is exactly what it's made for.
Especially since rohrkabel has been updated to C++23, the static link approach would be even more cumbersome, as we would also have to pull in a more recent GLIBC (instead of just GLIBCXX) iirc.
Sorry for the inactivity on venmic - have a lot to do currently :) I've just updated rohrkabel to support coroutines (async/await) and made some general improvements.
Will work on a bigger venmic update soon :P
the thankless work continues..
As always
@pallaswept Could you post the output of pw-cli ls Node and pw-cli ls Port so that I can see how the channel map looks like?
Sure! This is with the machine idle (your timing was excellent) pwcli.txt
Do you need one with vesktop running? In chat and screensharing?
Sure! This is with the machine idle (your timing was excellent) pwcli.txt
Do you need one with vesktop running? In chat and screensharing?
Thanks! Idle suffices, could you maybe also send the output of: pw-dump | grep channel?
Compiled this change and it doesn't seem to work on windows, apologies for the lack of input but I have 0 knowledge when it comes to electron's screwery.
Hope this helps pwdump.txt
That's perfect thanks :)
This allows the icon to be properly shown
This is already what the WM class is, so adding this doesn't really do anything
Are you using the flatpak? Try this PR instead:
flatpak install --user https://dl.flathub.org/build-repo/176750/dev.vencord.Vesktop.flatpakref
@vernal lintel the DE might just be cooked and need the hint anyway
Using original Vesktop, without this change, open a terminal and run
xprop WM_CLASS
then click on the Vesktop window. What is the output of the command?
WM_CLASS(STRING) = "vesktop", "vesktop"

Using original Vesktop, without this change, open a terminal and run
I want to add that with or without this change the window class is always going to be the same vesktop.
StartupWMClass is used by the DE to know the name of the window class of the respective app, in the case of Cinnamon, KDE and others it is used to allow pinning apps to the dock, what you said makes me think that this would change the name of the window class when that isn't the case.
KDE still lets you pin t...
Is there really no way to get screenshare in stereo? I work with a friend on audio stuff and we sometimes need stereo to work. I still see no plugin or fix about that. By "completed" does it mean there's a fix or that it can't be fixed? :(
Is there really no way to get screenshare in stereo? I work with a friend on audio stuff and we sometimes need stereo to work. I still see no plugin or fix about that. By "completed" does it mean there's a fix or that it can't be fixed? :(
It seems to be down mixed on discords side
electron builder already does this?
package.json: Line 67
"productName": "Vesktop",
As name, but allows you to specify a product name for your executable which contains spaces and other special characters not allowed in the name property
????
Version: 22.4.1 Electron Version: 6.1.9 Electron Type (current, beta, nightly): current Target: linux (all) My OSS project (github.com/beekeeper-studio/beekeeper-studio) is not showing an app icon ...
closed by stalebot
electron builder is so cooked
LinuxTargetHelper.ts: Lines 136-141
// https://askubuntu.com/questions/367396/what-represent-the-startupwmclass-field-of-a-desktop-file
// must be set to package.json name (because it is Electron set WM_CLASS)
// to get WM_CLASS of running window: xprop WM_CLASS
// StartupWMClass doesn't work for unicode
// https://github.com/electron/electron/blob/2-0-x/atom/browser/native_window_views.cc#L226
StartupWMClass: appInfo.productName,
literally not true
dumb ass default tho
In what scenario would you want the wm class to be the product name instead of the binary name
never
we should pr fix to it
the wmclass and binary name can be different
in fact by spec they should be
binary name of vesktop and class of dev.vencord.Vesktop
but nobody cares about the spec
hop on doing wtfyw
but startupwm class has to match the desktop file name
either that or binary name idk this stuff is confusing
but product name is the worst possible default lmao
yop and the .desktop doesnt need to match the binary name
dev.vencord.Vesktop.desktop can launch vesktop just fine
productName is supposed to be dev.vencord.Vesktop I assume I think
name would then be Vesktop
so by spec its correct
but by convention ๐ถ
i think they'd reject that pr on grounds that it'd break people who are actually following spec
how do you even change the wmclass that vesktop uses
does electron even let you
no
product name is human readable name
i think so anyway lmao
It might brick Vesktop considering all the desktop functionality would be reimplemented in that and work fundamentally different than how it does with Vesktop
Aight
@brazen kite btw can you try your mouse buttons with the latest commit in the pr
lets the user use existing signins in their browser for linking.
// Transparency fixes, window is resizable when fullscreen/maximized
this.mainWindow.on('enter-html-full-screen', () => {
this.mainWindow.setResizable(false)
})
this.mainWindow.on('leave-html-full-screen', () => {
this.mainWindow.setResizable(!this.mainWindow.isMaximized())
})
this.mainWindow.on('enter-full-screen', () => {
this.mainWindow.setResizable(false)
})
this.mainWindow.on('leave-full-screen', () => {
this.mainWindow.setResizable(!this.mainWindow.isMaximized())
})
this.mainWindow.on('maximize', () => {
this.mainWindow.setResizable(false)
})
this.mainWindow.on('unmaximize', () => {
this.mainWindow.setResizable(true)
})
this.mainWindow.on('will-move', (e) => {
if (this.mainWindow.isMaximized()) {
this.mainWindow.setResizable(true)
this.mainWindow.unmaximize()
e.preventDefault()
}
})
some foolproof improvements for transparent window handling, enjoy
i couldnt fix the window snapping or the title bar dblclick to fullscreen
that'd require native code, cba
?
enabling transparency = buggy
this code fixes a lot of said bugs
well, works around them
but same thing
not an electron thing
transparency stuff works fine on linux
its a chromiumg thing
no, no it doesnt
ive literally not had any issues๐ญ
i mean if chromium itself is never in a transparent state and it never can be it's really an electron issue
its a chromium issue
Is vesktop taking forever to load a known issue
Recently it's been taking forever >30s to reload the page ctrl-r or from scratch
works on my machine
(rpm)
ClientTheme issue maybe?
some people were saying it causes that
No, I don't have it enabled.
you will debug why it happens
ur internet might just be cooked
or its not caching the page correctly for some reason
aren't you on nixos
maybe no perms write to cache
insane
remember: if you have an issue on NixOs the issue is NixOs
it's not a nixos issue cuz it works for me 
nixos is known to have one singular config file that everyone uses ๐ฅธ
nop but we both use the same vesktop package
sorry it took so long ๐. i reinstalled windows and couldnt be bothered to set up pnpm until now
it doesnt work yet
ok so it just broke all together, actually...
- switching tabs or leaving out of the keybinds menu resets the keybind u set
- in the menu while u have it set, trying to use the keybind doesnt work
also, just to reiterate, none of the buttons on my mouse are detected
yeah i wasn't entirely sure whether it would work or not
since then i have found out mouse buttons are an entirely separate event in libuiohook :P
huh weird how when recording keybinds it works just fine but not when saving them can you see if there is anything in the console logs??
that's intentional. discord does the same thing
except it also has a warning about it :P
WebRtcAllowInputVolumeAdjustment is unreliable on windows...
You can patch getUserMedia at runtime like in #832 to actually disable agc.
Someone pleas push this. I would like to change tray icon. And that would be perfect.
Been using this fork for the past few weeks. works fine. <3
what's the context? i was honestly considering switching to using discord in firefox because electron font rendering is so bad lol
wait huh
Vesktop and Discord's official client use the same electron version now (it seems), but the latter somehow has font rendering more screwed lol
I assumed the issue was the same
the flatpak version is fine (just greyscale but still readable) but the portable is not (feels like I have eyesight problem)... ??
so I imagine this issue doesn't have many people complaining because it somehow depends on the environment (??) and I also imagine it's much worse at 1080p 1x scale (which I use
)
cause its a bug and already fixed in electron
Recent commits have made it dysfunctional, where you can't even save keybinds. Last known working commit: 12561e3
I don't even know how I missed that. fixed!
for what it's worth... I just tested it on linux with the latest PR (thank you for fixing the issue with the keybinds not saving) and it worked
I can try windows on the PR later
Is this getting a fix? I'll have to switch back to Discord, because this is killing me
This PR addresses #161 by implementing a proper switch to disable automatic gain control.
- PR #1139 attempted to solve this using WebRtcAllowInputVolumeAdjustment, but this approach is unreliable on Windows
- PR #832 used a working approach (patching getUserMedia) but was hardcoded without UI controls
So this implementation creates a proper switch that actually disables agc on runtime. the flag used by 1139 is not reliable and does not work on windows, not sure why, ask chromium. I ...
@trigger337 since someone like you actually came and commented on this issue, I took the liberty of implementing my solution on #1160 , hopefully it'll be reviewed soon. This PR seems pretty cold, the author hasn't touched it since they created it and I would also love a fix for this as soon as possible...
hey @pap-stavros i put this up as a quick fix in case others ran into the same issue i did, i dont use windows myself, so if anyone else wants to weigh in, please feel free
@atar4xis hey thanks, your contribution was very helpful in leading me to the right direction as i don't really know react myself.
This should be split into its own file in the preload and it should have a liscence header referencing https://github.com/joeywatts/disable-autogain-control-extension/blob/master/disableAutogain.js
Thanks for the clarification, doing that now
@Vendicated any updates on the status of this PR? Would be great to have this merged in some form before the two-year mark rolls around.
can someone here clone my fork and update the venbind version to 0.1.6 and see if everything works properly on x11???
twitter eleven
Is there a continuance of this PR somewhere or a way to track the issue?
