<eyJhb>
Any good PDF editor? LibreOffice screws with the format :(
<etu>
I usually use gimp :p
<etu>
It's not made for that :D
<etu>
But like, filling a form that isn't a form to then print it. Then gimp is fine :p
<ekleog>
xournal is pretty nice too
<eyJhb>
etu: I just need to edit a t => n, and put in my signature :p
<eyJhb>
etu: the quality when I open the PDF is sooo bad in gimp
<eyJhb>
ekleog: isn't it a note taking tool?
<ekleog>
eyJhb: it is, but it also can take notes onto a pdf
<ekleog>
then, if it's still select-able, you can just vim the pdf, I guess
<joepie91>
eyJhb: huh. I've been using libreoffice for this without any issues
<joepie91>
specifically the form editing usecase
<joepie91>
you sure it's opening in Draw and not Writer?
<eyJhb>
joepie91: it is screwing with the text going over the sides of the pages, might be a font issue
<eyJhb>
And yeah, I use draw normally :D But I just set up the resolution for Gimp, and it worked great
<eyJhb>
gimp+pdftk => wonders
<avn>
eyJhb: you need edit pdf, or just create? if create only -- pandoc is a first candidate ;)
<eyJhb>
avn: edit ;) It was a contract
<joepie91>
eyJhb: on which note, I can strongly recommend having a PNG of your signature :P
<joepie91>
transparent
<eyJhb>
joepie91: already have it :D
<joepie91>
even docusign accepts those nowadays!
<joepie91>
ah ok :P
<eyJhb>
Use it for my hourly based salaries, which I need to submit 3 of each month :D
<avn>
well, then use docusign or dochub ;)
<eyJhb>
But not transparent, but that is just me bing lazy
<eyJhb>
`bash: /run/current-system/sw/bin/du: Argument list too long` :(
<eyJhb>
I want to find the packages in my /nix/store that is biggest
rardiol has joined #nixos-chat
<eyJhb>
`80106 store paths deleted, 56463.94 MiB freed` sweet jesus :D
<eyJhb>
joepie91: do you prefer to put functions that require are required by another function BEFORE or AFTER that function? E.g. bar depends on foo, would you place func foo then func bar, or func bar then func foo?
<eyJhb>
*tagging you because I have a feeling you have someting on that subject*
<joepie91>
eyJhb: I don't have a strong opinion on it, so long as you do it consistently. I usually try to place dependencies before their dependents, and exports go at the end of the module
<joepie91>
just don't do the thing where you put them in random order :)
<eyJhb>
I like both of the methods, but it is only now I am thinking of it
<joepie91>
there's benefits of either approach; dependencies-first means you can clearly answer "what parts do I have by this point to build bigger things out of", dependents-first means you can clearly answer "what's the highest-level thing this module does"
<joepie91>
I find that an exports-last convention also answers that second question reasonably well, so that's why I pick that approach
<eyJhb>
I think I like it the most by placing dependencies under them, as if I need another one, I just add it after.. And makes the code easier to read, as if I have the main function, I can look at the other functions if I have no clue what they do by the name, and am not forced to it :D
<joepie91>
especially since I try to keep modules small, in terms of complexity and responsibilities
<joepie91>
but the difference is admittedly small
<eyJhb>
joepie91: when you say "exports-last", do you mean public things last?
<eyJhb>
So all functions accesible as a class etc. goes last in the code?
<joepie91>
eyJhb: I think so, yes, depending on what you mean with "public"
<joepie91>
which language are we talking?
<eyJhb>
joepie91: Golang atm. so everything with a capital letter to begin with :p
<eyJhb>
I usually code Golang... Loveit
<joepie91>
ah, I don't know much about the typical code structure
<joepie91>
and how classes are involved for example
<joepie91>
but yes, at the end goes the stuff that you import into another module
<eyJhb>
It wouldn't surprise me, if there were some Golang guidelines for this actually
<eyJhb>
`It's generally most common to put your most important functions first. Thus, all your helpers would be at the bottom.`
<eyJhb>
That is somewhat what happens over time
<eyJhb>
`var foundBuildImage, buildImageNeedsRebuild, foundGenerateImage, generateImageNeedsRebuild bool`. Seems too specific
<__monty__>
I prefer dependencies first, because that convention works across languages. Python for example *requires* dependencies first.
<joepie91>
eyJhb: that sounds like a terrible guideline to me
<__monty__>
However, I do prefer `where` to `let in` for tiny helpers that don't really matter, talking about haskell syntax.
<joepie91>
eyJhb: not just because of the term "helpers" (which is generally a red flag for bad modularity), but also because it's really ill-defined what constitutes "most important functions"
<joepie91>
it's notoriously difficult to predict upfront what parts of the code are going to need maintenance later, so how would anyone expect that guideline to be usefully implemented?
<joepie91>
at least with a consistent layout, you can reasonably predict where something is going to be, you don't also need to account for what you believe the original author might have used as a metric of importance...
<eyJhb>
joepie91: well, I think in most cases it becomes somewhat "obvious", but maybe notin all cases. e.g. currently I have some code to maintain my challenges, which should rebuild images, so the dependency list is doChallengeActions -> buildDockerImage -> createTarball -> readDir
<joepie91>
eyJhb: those are the most dangerous bits, the bits that seem 'obvious' but nevertheless are seen differently by everybody :)
<eyJhb>
So basically, doChallengeAction would be something that can be called from outside, whereas the other functions wouldn't be exported at all
<eyJhb>
That's true joepie91
__red__ is now known as redredred
<tilpner>
Hey, I know this isn't really the place, but I don't know a better place to ask
<tilpner>
I'm having a new networking problem, where many connections are not accepted from my home network, but some are
<samueldr>
(I know you don't want to do it) have you tried turning it off and on again?
<tilpner>
curl http://54.217.220.47 (nixos.org) is fine from outside, but gets stuck connecting from inside
<tilpner>
I've only nuked and restarted kresd before I checked if DNS was the cause
<tilpner>
I will restart things as a last measure, but I'd also like to understand what's happening, and restarting things is likely to hide that
<tilpner>
So if anyone has ideas on how to gather more information on why these connections aren't accepted, please do tell :)
* tilpner
powercycles router
tilpner has quit [Quit: tilpner]
tilpner has joined #nixos-chat
<tilpner>
Well, that didn't help
<tilpner>
And since it's happening on two NixOS machines, I'm not going to restart them yet
<gchristensen>
restart one but not the other?
<tilpner>
It's even happening on my Raspbian Pi 1b
<tilpner>
So it's not NixOS-specific
<tilpner>
(And on my Android phone)
<tilpner>
github.com works, nixos.org doesn't
<gchristensen>
differences in routes?
<tilpner>
How do I check that?
<tilpner>
I would look it up, but the only search engine available to me is bing
<gchristensen>
traceroute
<gchristensen>
traceroute nixos.org inside and outside your network
<tilpner>
By outside, I mean from a Hetzner cloud instance
<tilpner>
And the parts that aren't just "* * *" are different
<tilpner>
They don't share a single route element so far
<gchristensen>
not surprised
* tilpner
restarts laptop, but that can't do anything
psyanticy has joined #nixos-chat
<tilpner>
I might just call my ISP :c
<cransom>
do you have a switch at home or is it all plugged into the router?
<tilpner>
cransom: All plugged into a fritzbox (router)
<cransom>
mtr to nixos.org and github may be interesting.
<cransom>
(working on non working examples)
<tilpner>
Huh, cache.nixos.org still works, just fetched mtr
<gchristensen>
cache.nixos.org and nixos.org are on totally different networks
<tilpner>
Oh, I just assumed nixos.org was also fastly-hosted
<tilpner>
I let them run for a while, I don't think those replies are coming
<cransom>
nah, they are likely dropping those packets. only other thing i'd do is swap an ip one of the machines to see if the behavior follows.
<tilpner>
What do you mean, swap them?
<tilpner>
It happens on all devices connected to the router somehow
<cransom>
oh, i thought some worked.
<tilpner>
Only my Hetzner Cloud VPS works, and when I switch to mobile data on my phone
* colemickens
giggles at Mirantis.
<tilpner>
I blame this on Vodafone
* colemickens
maybe the usb type c haters should admit they're upset their iphones have lightning ports
* colemickens
ducks
* gchristensen
is
<__monty__>
You have to admit that seeing a port and knowing what it does is nice though? Is it thunderbolt, does it allow charging, is it reversible?
<samueldr>
hm?
<samueldr>
type-c is neither of all that
<samueldr>
like, you can't know what it does, except that it will accept a type-c plug
<samueldr>
but is the type-c plug actually reversible? will it accept the alternate modes? will the device have alternate modes?
<samueldr>
will power delivery work?
<__monty__>
samueldr: Read my statement again. That's exactly what I said.
<__monty__>
Using 4 fewer lines even : )
<samueldr>
didn't contradict, exactly, probably said a similar thing, but I wasn't sure because of the tone of the first sentence
<samueldr>
"You have to admit that seeing a port and knowing what it does is nice though"
<samueldr>
though, the reversible issue is not type-c, but implementors doing it wrong
<samueldr>
(except for debug cables, SBU1 / SBU2 are not reversible per se)
<colemickens>
I would certainly be annoyed if I plugged an expensive nice TB3 dock into a new Surface Laptop 15 and have it not work.
<samueldr>
it might, and AFAICT there is no way for the devices to explain why
<samueldr>
is it because you're using another cable than the "right one"
<samueldr>
like, I have this type-c dock (not tb3) which *has* to be used with a short cable
<samueldr>
but it's not documented!
<samueldr>
and when you're not using it, it just fails in weird ways
<samueldr>
(which is dumb, because the docking station was desired to make it a single-cable solution that could be stowed far from the computer!)
<__monty__>
The reversible thing isn't the standard's fault but let's be honest, did we really expect everyone to do it *Right* especially for usb-C to type A adapters?
<__monty__>
The cable thing is even worse than the port thing.
<samueldr>
now that's cynicism for cynicism's sake
<samueldr>
we did expect everyone to do it right, and the USB IF did
<samueldr>
and it shouldn't be hard to do right
<samueldr>
especially for A to C adapters
<__monty__>
That sounds like mindless optimism tbh. People aren't perfect. Especially when money's involved.
psyanticy has quit [Quit: Connection closed for inactivity]
<__monty__>
That's for a license to be a registrar?
<__monty__>
Afaik they buy the domains at auction?
<joepie91>
<monty> Why do we all put up with leasing domain names temporarily?
<joepie91>
because noone has - reliably - figured out decentralized namespacing yet
<joepie91>
and for whatever central company manages them, there's an incentive in asking money
<__monty__>
But so much about the internet is open. Why not this?
<tilpner>
__monty__: I think you have to pay a fee to be allowed to buy IPv4 addresses
<tilpner>
(I don't really know what I'm talking about, I only vaguely remember this)
<tilpner>
So is it really that open?
<__monty__>
I'll ask again here because #nixos didn't offer advice. Since updating my channel on a mac, mpv is crashing with an error about a dynamic lib. What's the next step here? Submit a nixpkgs PR marking the package as broken?
<tilpner>
__monty__: A bisect might help
<gchristensen>
open an issue describing the problem first
<__monty__>
I'll open the issue now, can't bisect until tomorrow though.
<joepie91>
__monty__: because, like I said, decentralized namespacing is an unsolved problem
<joepie91>
you can only make things 'open' in the sense of decentralization if you actually have a model that makes it work
<joepie91>
reliabl
<joepie91>
reliably*
<__monty__>
It's only gotten attention in the last couple of years. It's not solved because people weren't interested.
<joepie91>
for a lot of other things that's easy if you're going for a federated network like the internet, but namespacing is not one of those things
<joepie91>
same reason that virtually every other decentralized system still relies on DNS
<joepie91>
__monty__: that's nonsense
<__monty__>
For *a* definition of solved.
<joepie91>
research on this topic far, far predates the current blockchain-y decentralization hype
<joepie91>
it's not like people just decided in the past few years that "hey we should decentralize things"
avn has quit [Remote host closed the connection]
<joepie91>
there's been many many hours of effort and thought invested in building decentralized systems far before Bitcoin was even a vague thought in someone's head
<joepie91>
and that includes namespacing
<__monty__>
I think you're talking about an idealized decentralized system.
<__monty__>
Domains could easily have been decentralized to the national level, ccTLDs managed by country's governments.
<colemickens>
onion?
<__monty__>
Instead only a couple were pushed hard, in the hands of a US company.
<joepie91>
nevertheless it works basically how you're describing it
<joepie91>
with a handful of (highly questionable) exceptions, ccTLDs are managed by their respective countries
<joepie91>
though that job is often outsourced by said countries
<__monty__>
Now yes, and yet ".com" is still valued more.
<joepie91>
that doesn't really change anything about the organizational structure of ccTLDs
<__monty__>
It does highlight that the damage was already and continues to be done.
<gchristensen>
who pushed .com to be most wanted?
<samueldr>
the bubble
<__monty__>
Not sure, whatever US registrar was managing it?
<gchristensen>
I doubt it
<samueldr>
it wasn't called the .net bubble
<samueldr>
but also, at first, it was the only "any kind of use" TLD
<samueldr>
.net and .org had specific uses
<samueldr>
the original ones were .com, .net, .org, .edu, .gov, .mil and .arpa... some were obviously not general purposes :)
<gchristensen>
colemickens: if I find myself asking that question, it is time
<joepie91>
heh
<samueldr>
that's when you rewrite it as an expect script?
<joepie91>
grumble grumble some people
<joepie91>
just had to hand out a ban in another channel, someone whose only mode of communication appeared to be "complaining about a thing"
<joepie91>
literally everything sucked, everything else was better, anyone responding to them clearly didn't understand a thing and couldn't help them
<samueldr>
you don't understand
<samueldr>
this discussion sucks
<joepie91>
genuinely, how do these people expect to get anything useful out of their participation in a channel
<samueldr>
(sorry, it was too tempting)
<samueldr>
the previous discussion was better
<joepie91>
lol
<samueldr>
though such things wouldn't penetrate their thick mind, I guess
<cransom>
this software is popular, why aren't you spending every waking moment making it better for likely no compensation?
<gchristensen>
I do =(
<joepie91>
cransom: oh man, if it were that, it would've been a big improvement over the actual discussion
<joepie91>
the actual discussion was more like them complaining about literally everything they didn't understand (as in, not even remotely understand)
red[evilred] has joined #nixos-chat
<red[evilred]>
Sometimes people can’t see past their own point of view
<joepie91>
they initially believed that the 'await' keyword in JS was a replacement for semicolons, for example
<joepie91>
which, like, if you genuinely don't know, fine. but then don't start ranting about how awful and ridiculous 'await' is on the basis of that (lack of) knowledge
<joepie91>
and several iterations of that about various aspects of JS and the libraries they were trying to use and basically everything
<joepie91>
with some "in Java everything is better" thrown in for good measure
<red[evilred]>
I’m opinionated, but not an ass :)
<red[evilred]>
But yeah, I have little time for constant negativity
<red[evilred]>
Also, learn to read a room
<worldofpeace>
lol, it also can be hard to do that when it's virtual. But silence is a good indicator or discussion happening that you're not a part of.
<samueldr>
aren't we all... only... reading... the room? (hmm... that's leaving users with sight issues off though)
<gchristensen>
and people who have difficulty on social cues
<gchristensen>
I find usually people who have difficulty with social cues do well with a quick PM, though
<samueldr>
(that was a poor joke attempt on the fact that it's written text)
<samueldr>
(which can be meta-interpreted as not being able to read the room['s subject])
<worldofpeace>
I found it funny samueldr :D
<samueldr>
poor != unfunny!
<worldofpeace>
For whatever reason a person may have issues with social cues, I have to agree they generally do better 1:1 (sometimes)