Or use a while-read loop anyway?
while read -r foo; do
cmd "${foo}"
done < <(ls -1)
(But don’t script ls)
The find example should just use -exec grep "USB" {} + also, though -exec generally is the correct way to do this.
iptables-->nftables
And if you really want the iptables UX, iptables-nft is also an option (at least on Debian). While I prefer firewalld most of the time on a server, my boss really wants to stick with the same tools he’s used for 20yr; so iptables-nft it is.
I’ve somewhat recently taken TrueNAS to avoid the headache of DKMS for ZFS support.
While it is nice to know I basically don’t have to worry about that, the rest of it is pretty painful – I can’t even install a package without manually running it in docker.
We’ve in a CentOS/Rocky shop at work until recently, moving to Debian.
If I has to redo my NAS, is take Debian or something and call it a day. The hassle of TrueNAS and similar distros do not really seem worth it, if you want more control.
Dunno if you saw, office.com now refers to Office (the “launcher” ?) as:
the Microsoft 365 Copilot App (formerly Office)
Can’t wait for Copilot OS (formerly Windows)
I recently had to port a bash script to POSIX because my boss demands we still use at, and wants exec to look like at -f.
99% of my time is spent in bash, so in conjunction with the POSIX spec documentation, Copilot was useful in finding a solution for >(and ...) redirection (logging related), as my bash solution was not compliant.
It was helpful, but I still took the time to locate examples in the spec, and validate every single answer I got.
Outside of this, it has basically never been more productive than doing my own searches, research & testing (sysadmin & “devops”).
I’ve recently switched from np++ to Sublime for some non-standard issues – I would say that could be closer in performance & extensibility to Vim/Emacs; though limited to GUI and non-FOSS of course.
If we’re including shell scripts/functions as “terminal commands”:
bm()
mkcd()
mkdir && cdbashlib
gitclone.sh
git clone wrapper without needing to give a full URImd2d.sh
pandoc wrapperdocx format, so I use this to avoid Office as much as possibleIf we just mean terminal applications:
hyperfinejq and yq
xmlstarletshellcheck
A couple of bash specific items I’m using quite often these days:
mapfileprintf '%(datefmt)T
%T doesn’t expand to true ISO-8601 compliance(( expr ))
shell built-in command
After looking into it a bit more, its at least a builtin for bash but is otherwise not POSIX. I guess nohup ... & would be the POSIX compliant equivalent, though still not a builtin.
Its my understanding that & backgrounds, not necessarily detaches, a process – if the parent process closes, I think the background tasks would still be wait()ed on, if only using &.
There are times when dealing with annoying field separators that awk is a more convenient tool – though, I’m also now at the stage that I want to do as much with bash-builtins as I possibly can.
You can get rid of those greps, btw:
ps aux | awk '/zoom/ && $0 !~ /awk/ {print $2}'
Or just use pgrep.
I both love and hate awk – on the one hand, it provides the same/similar functionality of similar tools (sed, grep, cut, etc); but it is a bit of a bear and can be pretty slow.
If you need more “complex” tasks done what would be cumbersome with the rest of the standard tooling, and performance is a non-issue, awk/gawk can probably get it done.
Though, I too am trying to use it as little as possible in scripts. I think multiple subshells/pipes is still better than awk in some cases. Syntax also leaves a lot to be desired…
Shouldn’t you end with & disown to fully detach?
From my experience, that sounds like an added bonus.
PuTTY
If you find yourself back on Windows, 10 & 11 both come with openssh natively. Combining that with WSL even gets you X11 Forwarding, if that’s a useful feature.
Titling choice probably fueled by their otherwise unrelated ongoing controversy, if I had to guess.
More blood for the blood god.
Thanks for posting the blogpost – when I checked this thread originally (and the article), I seemed to have missed the focus on DHH. Admittedly, I just don’t know much about him – though, I’m starting to get an idea why this blew up so much.
As I added in another comment, I misunderstood the DHH element of the discourse as I, admittedly, don’t know much of anything about him – I’ve heard some references here and there, but that’s about it.
Taking a stand against things like this causes change for the better in the long run.
That’s also fine, and I generally agree. My concern basically boils down to killing momentum by sinking a company with (probably?) sane views on right-to-repair & libre as topics.
If the goal of a boycott is to starve the company until it goes under, because they made a move we don’t like – then that I don’t really like in this context. If the goal is to force their hand towards at least transparency, or maybe force NP to step down; then I’d support that.
The main problem is […]
I’ll admit, I only vaguely know of DHH by name and Rails, vaguely remember the Omarchy announcement, and that’s about it. I seem to recall Prime referencing DHH’s controversial opinions, but I can’t say I’ve gone any deeper than that.
If the discourse really is primarily focused on DHH/Omarchy, then I guess I just misunderstood this post/title & the article…or just don’t have the full context regardless.
I’ll admit I’m not up to date on the hyprland/vaxry lore – but I don’t understand the level of outrage based on this article…
I’m also not sure why the sponsorship of a software project is necessarily being treated as a 100% endorsement of both the maintainers and their alleged views.
I’m also not sure if infighting and purity testing will help the movement(s) right now. Once it’s the norm, sure, but it’s still a relatively fringe movement within the industry.
Edit (2025-10-15@20:14): At the time of writing my comment, I was both unaware (and uninformed) on the DHH side of this topic. While I still think the level of outrage is maybe a melodramatic, the push back seems more warranted than it initially seemed to me. I still don’t know much about DHH beyond Rails (and even then, not much); but from what I’ve seen since my comment, the response is more understandable.
Several more states have introduced/passed similar legislation already, so it’s becoming more of a problem; or see Brazil’s law.