Soldered RAM is just about impossible to upgrade/replace, you have to desolder it and do a bunch of other tasks. I have only ever seen one person do it on a modern laptop and it looked horrendous: https://gregdavill.com/posts/dell-xps13-ram-upgrade/

I would just completely abandon any plans to upgrade a soldered RAM laptop unless you are extremely skilled.

Unsoldered RAM is just push fit sockets that have a release clip you need to pull if the socket is already populated. The RAM can only go in one way, its super super easy. Just make sure you double check the spec of RAM you buying against the laptop specs to make sure you buy the right sort.

The object selection tools in lightroom are pretty amazing, being able to choose select subject on the menu and it just selects the ducks or people or whatever pretty much perfectly each time to make your mask is pretty bonkers. It saves so much time over trying to select an object in dark table.

I really like dark table, I actually prefer the way it stacks modules vs. lightroom, and this is just for complex object selection, I can select the sky or background or whatever pretty simply in dark table.

Oh and dark table does not support DNG files. My workflow using PureRAW outputs compressed DNG files and Darktable will not support them currently. Sure I can go other routes for my export but the smaller size of a compressed DNG is very attractive when I can be working with about 100 * 40megapixel images.

PureRAW can sort of be duplicated by dark table, but again its not quite as good, doesnt quite have the same list of lens for correction/denoise capability. I shoot wildlife a lot and high ISO is a factor of life. Its not that dark table is bad at this its just PureRAW is very good at removing that noise and sharpening.

So lightroom and PureRAW forces me to have either windows or macos, and a shitty subscription for the former.

Distro is more an alignment of philosophy between you and the distro. Something slowly updated but really stable? Debian. Something cutting edge, but with lots of guides? Arch, etc. etc.

Any of them can pretty much run any shell, DE or WM, and as that’s what you spend the most of the time interacting with, that’s a more personal touch point. The distro is really just the package manager that you regularly interact with, and thats easy enough to hide behind something like topgrade.

I have only used Sway for a few years and anything else feels bloated and slow to use to me now. I spent a long time tweaking to get it how I wanted both in terms of add ons and config, then setting the keyboard shortcuts that work for me. I even have a bunch of them configured on my actual keyboard on layers to make them even easier to activate.

Its worth the investment for me as its now transparent to my workflow. I run the same config across all my machines and its been a stable config for the longest time. Long term stability is the key for me.

Redhat 4.1 back in 97. I even purchased the CD from PC World, seems wild now to buy a CD/DVD of a distro.

First PC I installed it on was a work laptop, had to compile a bunch of kernel modules and then the kernel to get everything working but get everything working I did, Thinkpads being good for Linux even then.

Another starship user. Mostly want it to summarise useful stuff for folders pulled from git or whatever so it’s pretty plain rest of the time. I use the same on all my boxes

Borg daily to the local drive then copied across to a USB drive, then weekly to cloud storage. Script is triggered by daily runs of topgrade before I do any updates

Ive said repeatedly that its certain to be an issue with their setup I have even provided proof that its not typical.

Its just a terminal applies to pretty much every type of application. There are lightweight WM, text editors, terminals, etc. etc. for a reason.

I can run my current setup with its 2 different animated 4k wallpapers and all the over shit I have running on my old 4800u that is only slightly slower and it consumes an insane amount of CPU at idle. Just because you can does not mean you should.

Its six years old, that’s starting to get on a bit now for a processor that was never anywhere near top of the line from AMD when it was new.

I think if you are trying to bling our your desktop and not expecting it to impact performance from an older, less powerful setup then generally speaking you are going to have a bad time. You should be pitching your desktop experience based on what your hardware can handle, there are plenty of terminal options available depending on what you need, just like there are plenty of WM/DMs if you have a lower spec machine.

Having said that, it was pretty damn obvious that there something wrong with ghostty on their setup, and its misleading to say that ghostly is just bad because of that.

Its a bit slower than Alacritty for my use case, not massively enough, but enough to put me off. The extra functionality such as its TMUX stuff I just do not need. I think if you want a more fully featured terminal, particularly if you do a lot of code writing in the terminal, then I would pick Kitty.

I only really do quick remote editing in the console so its not important for me, and I do not want TMUX as I use a tiling WM. Terminal launch speed is particularly important to me because of this.

I haven’t tried foot yet, that is meant to be good for wayland and as I use Sway it might be better fit. I would need to get frustrated with Ghostty before I could be bothered to switch, which is what happened to me with Alacritty over image support, shallow as that sounds.

ryzen 5 3600

Yeah thats what I would call old now, lol.

There is 100% something wrong with your Ghostty.

Heres a shitty gif I made to show you how fast Ghostty, Alacritty and Konsole is on my PC, all way less than a second and I am running a ton of background crap including different 4k animated wallpapers. Gif

Something wrong there, Ghostty is just as quick for me. Are you using an older PC?

Both load for me in milliseconds, even with fastfetch stuck in my startup.

I recently switched from alacritty to ghostty as I wanted image support as ghostty implements the kitty protocol for it. Ghostty seems as fast as alacrity to me, but with better support. It even has a tmux type replacement, although I haven’t used it as I don’t need it with sway doing that for me.

If I am relying on it, I buy from brands I trust. No brand is going to be perfect but some are clearly going to be lower risk than randoms from aliexpress. Its as much to do with reliability, achievable duty cycle (rather than promises of duty cycle), support (especially how easy it is to get a replacement under warranty), how long they will push firmware updates for, than just security trustworthiness.

Pretty much any device is going to have a vulnerability or potential for a back door at some point but the company being transparent about the issue and fixing it promptly is worth a lot. Its the same reason I would have a Google or (premium) Samsung phone, I trust that they will support the phone for the time period they say they will, something I would not do with say Oneplus based on my past experience of them.

I buy electronics from aliexpress all the time, but nothing I rely on day to day like a router, simply because I am shit out of luck getting it replaced quickly if it goes wrong, even if I want to get a replacement. I have a cheap mikrotik hex I keep as a backup of a backup (my APs are my primary backup for my router), and this is fine for a week or so but I would not want to be out a month or more with it.

I guess you could plan in proper redundancy as I have, or may be you can afford a an outage, so may be you don’t need that. If I cannot work, I cannot earn, so I have backup internet, routers, wifi etc. planned into my install.

I think what someone else wrote about defense is depth is the real key here. I have my network divided into separate VLANs that are firewalled off from each other, so one for IoT, one for cameras, one for my TVs and other screens, one for my devices. This means if something is compromised they still have to get across the network and it simplifies my firewall rules as I am applying them to subnets rather than individual devices in a self maintained group. It makes it easier to say block external DNS queries and redirect to my pihole for my IoT and TVs but not my personal devices as I would have a good reason to go external.

May be you do not have a lot of devices, I realize I am nearer the upper end of a home network with over 50 active devices and it will be over kill if you only have a laptop and a phone on your network.

If you are the person asking the floating window question from the other day as JustAnotherKay spotted, then this is how I set a window to floating in my config;

for_window [title="www.youtube.com" app_id="firefox"] floating enable, resize set 1280 720 , opacity 1

What this is doing is selecting any firefox app that has www.youtube.com anywhere in the title and make that floating, with a set size, and remove any opacity (transparency) that might be applied to the window.

You can add move absolute position 0 0 on the end if you want to set the absolute location for the window.

You can force a window to a particular workspace by:

assign [class="discord"] workspace number $ws2

and that workspace to a particular monitor with:

workspace $ws2 output DP-1

If I wanted to do this for all firefox windows I would just remove the title= part from the selection.

How do you get the titles and other components? Using swaymsg as follows:

swaymsg -t get_tree

this will output all your open windows per monitor, for each app you want to manage you are looking for something like:

#15: con "#tech-talk🖥 | 40% Keyboards - Discord" (xdg_shell, pid: 6260, app_id: "discord")

from here you its simple to pick up what you can use for a unique select, so app_id:=“discord” in this case.

If you reload your config file, then reopen the app, it should reflect the changes you made to the config file. Logging on and off in the worst case will restart it.

Sway let’s you float and overlap windows. Also for both tiled and floating windows you can set the size, position, workspace and monitor by the app id (e.g., firefox) or part of the title (e.g., lemmy).

Yeah roll your own everything even greeter is the way forward if you want to customise.

I used to use gnome and I am heavily into customization. I gave up using gnome as they would constantly change things often for no real reason that whimsy, breaking previously working scripts, extensions and so on so I stopped using it. Its fine if you want to customize the basics like wall paper but I really wouldn’t bother for in-depth customization. Not because it isn’t possible, but because maintenance of it is a PITA.

Yeah I added bluetooth in mine, it’ll show me whats connected on hover and just launch the gnome app for bluetooth if I click it, super lazy implementation. I don’t need brightness controls so never looked at them.

EOS seems to use mako for notifications? I have never tried it.

I use swaync, which once themed and the rights bits you want, added, is ok. I wanted something more like the Gnome notification drop down that had do not disturb, media player controls, extensible menus, etc. in it.

The EOS files look a great place to start.

Do you default your apps to start on a particular workspace?

Never used hyprland but Sway you can use the mouse to move stuff around, resize windows, etc. just hold down you mod key, usually super/windows key. If you have a bar setup correctly you can even click between workspaces or have a task list like on windows that you can click on. Alt Tab needs some re-imagining as its now three dimensional, but that’s easy to tweak to how you want it with something like swayr. You can even add a start button equivalent if you wish.

I use Sway on Tumbleweed, before that Sway on Ubuntu. I have six main workspaces defined, odd numbered workspaces on my left monitor and evens on my right monitor. Both monitors are 32"@4k so a ton of real estate, I can easy fit in four large tiles per monitor, eight is a stretch but if you use the option to make windows full screen then you can run stuff in the background and then flip between things that are running in the background.

I use the layman add on to predefined layouts for my different workspaces, then bind apps on start up using my config to a particular workspace. I can still move them around, but automating as much as possible with a tiling windows manager is the secret IMO. Having everything just work and appear where I want with zero faffing around speeds up my workflow enormously. On Windows I use power-toys to provide a noddy version of tiling, but everything has to be done manually and its a complete PITA over a work day where I am opening and closing stuff.

As an example, I have my third workspace as my main coding workspace. Its divided into 3/4 and 1/4. The larger part I lock VS Code to it, the smaller part is usually a Firefox tab for reviewing documentation. My second workspace is my social workspace, that’s divided into four long quarters, one for music, one for discord, one for signal, one for mail. All of this, including binding the apps to the workspace, are fully automatic.

I use the keyboard for most things. I use QMK based keyboards (configured using Vial), so I can bind multi modifier shortcuts to just two keys either on a separate layer (activating the layer is one of the two keys) or a chord. Reducing the number of keys you press really helps the ergonomics of activating them, especially if you move them to the home row and away from the pinky finger hell hole that is where the modifiers are on most standard keyboards.

I think the biggest problem is that it requires work to get the right add ons and make it work the way you want to work, but get it right and the WM becomes transparent to how you work.