<steveeJ>
a co-worker pointed out to me that the sha256 strings in nix are invalid due to their length. is there any documentation on why they are only 52 characters long?
Jackneill has quit [Remote host closed the connection]
<__monty__>
With a specific non-standard set of characters, hence the link.
tilpner has joined #nixos-chat
<colemickens>
fun one: I'm working on trying to get wg working so my phone can access some services on my wg host. If I add a post-route masquerade rule, then I can access other LAN hosts and the internet from my phone (tested with only allowing VPN connections in Android 10). But I can't access svc on the WG host.
joepie91 has joined #nixos-chat
<colemickens>
Except for when I reload my system configuration ;)
<colemickens>
And then somehow it works for a second. If I start the page load on my phone, and switch-to-configuration, it works for a request or two, but then goes back to not working.
<__monty__>
Ah, the joys of networking...
<infinisil>
Hehe
<colemickens>
ah, it seems I forgot that I enabled/configured networking.nat.* last night. I guess that is covering the 90% that works.
<colemickens>
I'm half tempted to just throw this in a VM or something.
<colemickens>
from nat.nix: "# Allow connections to ${loopbackip}:${toString fwd.sourcePort} from the host itself" so maybe I need to configure it better or there's a bug.
<steveeJ>
__monty__: so it's base32 by default?
<steveeJ>
__monty__: also, do you know why nix uses a non-standard set of characters? I see the point my colleague makes that the sha256 string doesn't look like one and it's confusing
<__monty__>
steveeJ: The base32 encoding is *much* shorter. And it's non-standard because it stems from before base32 was standardized afaik.
<pie_>
i was lookng at some github or discourse thread the other day and its kind of funny
<pie_>
i think the chosen characterset was so that you wouldnt have curse words in them and its somewhere in eelcos thesis
<pie_>
which i think is kind of dumb but maybe it was a valid problem xD
<__monty__>
Well that was just a little extra bit probably? Surely the main goal would've been making typos invalid hashes? I.e., no o, O, 0, 1, I, l typos resulting in valid hashes?
<pie_>
well, im guessing from the above you can check, i didnt bother
<pie_>
__monty__: maybe that makes too much sense :P
<__monty__>
In the current PC climate I could imagine companies losing their shit over swear words in their production systems.
<__monty__>
So no consideration for accidental transposition? That's silly and counterproductive imo.
pie_ has joined #nixos-chat
<infinisil>
I'm not sure why tiny mistakes in hashes would matter
<infinisil>
Who types hashes manually anyways
<__monty__>
infinisil: I regularly type prefixes of hashes to find stuff in the store.
<__monty__>
Also, it's not like it costs effort to make it impossible.
<__monty__>
And the whole point of a hash is to identify something *uniquely* if visual collisions are possible the number of possible attacks increases a lot.
Synthetica has quit [Quit: Connection closed for inactivity]
psyanticy has joined #nixos-chat
atlas_ is now known as aatlas
aatlas is now known as atlas_
rardiol has quit [Ping timeout: 265 seconds]
rardiol has joined #nixos-chat
waleee-cl has joined #nixos-chat
pie_ has quit [Ping timeout: 265 seconds]
__monty__ has quit [Quit: leaving]
drakonis has joined #nixos-chat
thePirateKing has quit [Quit: Leaving]
endformationage has joined #nixos-chat
drakonis has quit [Quit: WeeChat 2.6]
rardiol has quit [Ping timeout: 260 seconds]
drakonis has joined #nixos-chat
waleee-cl has quit [Quit: Connection closed for inactivity]
drakonis has quit [Ping timeout: 268 seconds]
drakonis has joined #nixos-chat
MichaelRaskin has joined #nixos-chat
psyanticy has quit [Quit: Connection closed for inactivity]
drakonis has quit [Ping timeout: 265 seconds]
pie_ has joined #nixos-chat
<worldofpeace>
âšī¸ docbook isn't feeling very fun to write for nixos (at least for me). it feels like one of those things an editor is created for so you don't have to care to write it.
<pie_>
...are you trying to write docbook by hand
drakonis has joined #nixos-chat
<gchristensen>
đ worldofpeace
<worldofpeace>
I ain't hip, but I know my head hurts
<worldofpeace>
I might have to go to the asciidoctor
<gchristensen>
can I ask grep to skip the first 80GB of a file?
<worldofpeace>
had other sections, but I don't think docbook is good for the health :)
<infinisil>
worldofpeace: Would you prefer markdown over asciidoc?
<worldofpeace>
infinisil: tbh, at this point yes. I can produce and maintain it much faster. I think the most important fact should be, what will developers and documentation writers be able to produce the best documentation most quickly with.
<infinisil>
I feel like even though markdown isn't the best, the fact that everybody already knows it outweighs the downsides
<gchristensen>
+1
<infinisil>
s/the/but the
<infinisil>
Wait no
<worldofpeace>
I did docs in elixir, and it's quick and easy (was markdown).
<samueldr>
I don't find markdown to be acceptable for anything, it's a poor transcription of HTML idiosyncracies in an underdefined soup of characters :(
<infinisil>
samueldr: Yet it works for so many projects
<infinisil>
I wouldn't be opposed to asciidoc or rst though
<samueldr>
though asciidoc (through asciidoctor) I found pretty good as I used it for the Mobile NixOS docs WIP
<infinisil>
samueldr: Can you link to the source to have an example?
<worldofpeace>
samueldr: link? I never actually got to use it before
<samueldr>
though I had to build a bunch of stuff overtop asciidoctor to get all I wanted
<samueldr>
the parsing of asciidoctor is "lossless", so you can manipulate the AST for fun stuff in the future, just as you could with docbook
<qyliss>
Nobody knows Markdown
<samueldr>
e.g. you could add a step in the build where the lib.* functions examples are all tested
<worldofpeace>
ooh, that is a bit cool
<qyliss>
They think they do
<qyliss>
And then you get horrible broken docs
<worldofpeace>
qyliss: I think you've mentioned this before, I did like the points you raised on discourse
<qyliss>
I think I've come round to Asciidoc though
<qyliss>
It seems like the most sensible way to do plain-text-ish
<gchristensen>
I don't like that asciidoc is docbook
<qyliss>
is it?
<samueldr>
hm?
<samueldr>
qyliss: after *working* with ascii_doctor_, I can say asciidoc is a good enough foundation, but it's a bit "bring your own tooling" for generating the docs if what they give you is not exactly what you want
<gchristensen>
yeah it semantically matches docbookb
<samueldr>
though, it's extensible enough, probably through it doing the usual ruby things, that you can probably shape it like you want
<worldofpeace>
I think it's because it feels soo ascii it's quick for people to get their hands into. not to mention the ubiquity, people format their comments in markdown on github. I want something I can think in without feeling like a machine
<worldofpeace>
gchristensen: isn't domen using asciidoc with nix stuff?
rardiol has joined #nixos-chat
<qyliss>
gchristensen: what sort of semantics?
<qyliss>
I'm not sure I understand
<gchristensen>
I mean, asciidoc is literally designed to be a shorthand to author docbook