As a by the by... It looks like the basic small set of things I populate /dev with are what come from the bsdinstaller which puts those into place. So that's the origin for the files for me. It's what the bsdinstaller puts there.

does anyone else use lynx and notice there is a 30 second lag when requesting every site (http or https), yet curl gets the same site in milliseconds?

i'm using lynx from pkg fwiw

when I do lynx google.com it loads http, then https same as lynx cnn.com and constantly needs cookie approval. using elinks is instant. I'm not a lynx user however. but after just installing it to test, does seem much slower than say elinks

thanks, something odd then with my setup

the pause happens in elinks too

realdeimos: 30 seconds sounds like a dns timeout. Maybe it's using its own resolvers and the first one fails?

how can I grep two files in one command ? I used to do grep blah file && grep blah file2

but I just found out it has a major pitfall

I guess it's best to do grep blah file file2 ?

the hack I usually use is grep1 | grep2

how's that work ?

last1 yes you can grep multiple files with the 2nd syntax you provided

antranigv: tested out the changes in your test branch. worked like a charm!

quick question, id like to print some of the freebsd books but the pdfs have grey boxes for the code, trying to be nice to the copy service, what would be the easiest way to replace that with idk, just a border?

i was thinking some quick perl on the html then piping that in pandoc but if theres a better solution thatd be nice

drobban thank you! ^^ I will be doing more changes in the coming weeks.

nice! =)

dautor8518050: this is my first thought too, so I ran ktrace on it, and it is using the dhcp assigned dns servers. later today I plan to dig through ktrace/kdump some more.

last1, to grep one pattern from multiple files simply list multiple files on the end of the grep like "grep pattern file1 file2 file3"

When grep has multiple files then it prints the file name of the match at the start of the line. To avoid that use grep -h. To get just the list of matching files use grep -l.

does anyone use Kea DHCP? i'm having an issue where DHCPv6 works fine for one subnet, but for another it won't allocate any addresses with a vague error message. the config is pretty basic: le-fay.org/tmp/30d/R14W8s.txt

lw, is the server local to both prefixes, or is DHCPv6-relay in the mix?

CrtxReavr: both subnets are relayed

And what's doing the relaying?

a mikrotik switch

the relay config is also pretty straightforward: le-fay.org/tmp/30d/7Kq8h7.txt

So everything fails for 2001:8b0:aab5:104::/64 ?

well, it seems like there's only one client in this subnet doing DHCP... so perhaps it's something specific to the client, but i can't tell from the error

So. . .. full disclosure, I've not done anything with kea, though I have done a lot with ISC dhcpd6 & dhcp6-relay.

In dhcpd6 though, all prefixes had to be "children" of a shared network.

aha, maybe this is the problem, vlan104 is missing managed-address-configuration in the RAs.

there we go: lease for address 2001:8b0:aab5:104:1:: and iaid=1 has been allocated for 4000 seconds

CrtxReavr: thanks, that prompted me to give the config a second (or fifth) look :-)

Well, there you go then.

I need to get around to playing with kea at some point.

i will say this would have been a lot easier to fix if the error had been more clear, it wasn't that it couldn't allocate an address, the client (i assume) just wasn't asking for one

lw, is that static reservation for the host in question?

no, that's a different host

now to set up ddns... although i've noticed none of the DHCPv6 clients seem to send the hostname option which is a bit annoying

is bmake the default make on freebsd as opposed to pmake which the man page says to read up about?

johnjaye: yes, but bmake is derived from pmake so the syntax is very similar

ah ok, thanks

is there no way to configure an interface to use router advertisements for default route but not configure SLAAC addresses?

I haven't gotten around to configuring IPv6 yet, but I do remember it was somewhat of trial-and-error pain to figure out how to configure the clients with client-class/vendor-class/hostname/etc.

I mean for IPv4. It's different with IPv6.

I think it defaults to DUID, but I can't remember off the top of my head.

CrtxReavr: The somewhat disappointing aspect of kea, when compared to isc-dhcp, is that they've subscription-gated functionality, so you need to pay for some additional features.

Title: 16. Hook Libraries — Kea 2.5.5 documentation

The ones with Availability listed as "ISC premium library" are paid features.

Also, their dashboard is ok but I found it a bit weird: github.com/isc-projects/stork

Title: GitHub - isc-projects/stork: Stork is a dashboard for BIND 9 and Kea DHCP. It is under active development, with monthly releases. The initial priority is to provide monitoring and troubleshooting support.

that's disappointing

Overall though I still like kea over isc-dhcp so far.

i intensely dislike the json-based config but i know that's the trendy enterprise thing nowadays

It's easier to process using jq or similar tools than a lot of niche-oriented formats. That's where the value is, that it's widely used and supported with plenty of tooling.

lw: at least it's not yaml

also not like the bind config is very likable either

_xor: i understand the reasoning. i just don't like it

Heh, as a user I like YAML, but I can definitely see where parsing it is...a nightmare.

i edit DHCP config infrequently enough i can live with it, but the day nginx switches to json i'm going back to Apache...

I don't like programs that use for configuration yaml, json, xml, ini, or any other human hostile configuration format.

except for xml non of those are really human hostile

well they do not have to be

especially yaml

and ini is just key val pairs

yaml and ini are probably less bad than xml and json

nicely tidied json is not so bad

it does make sense to the eye

JSON is pretty awful, unless you use the non-standard variant that supports comments and trailing commas (Kea still complains about trailing commas even though it supports comments)

JSONc

Personally, I treat human-readable and machine-processable (including scripting) as two separate categories, but most of all, I care more about usage and support than language-specific annoyances beyond a baseline (meaning things like non-ambiguous expressions, etc).

I'd prefer to use XML or JSON or YAML if it means that it's like that across a bunch of codebases/services/whatever than if it's mix-and-match with a bunch of different languages.

Yeah, that's a JSON vs JSONC vs JSON5 thing.

Also, YAML is technically a superset of JSON.

Isn't it the other way around? json is part of yaml but yaml is not part of json?

Chomp operators in YAML make files eaiser to read, IMO.

my-property: [a, b, c, d]

my-property:

- a

- b

- c

- d

i wish we could have just standardised named's config format, that's so good it became basically the standard Unix configuration format until json took over

Oops, my bad.

Thought it was show up as a single message, doh.

I wrote a thing that lets me write out my configuration in lua and it'll convert it to json or some other format

kevans: LOL I was literally just thinking that.

"Then there are people like kevans who I'm sure use lua or whatnot."

kevans: Actually, you might be interested in this...

Title: GitHub - mikefarah/yq: yq is a portable command-line YAML, JSON, XML, CSV, TOML and properties processor

does json, ucl, and yml by default

interesting, though I try hard not to have to parse these things

I've been using it for quite a while now and it's been a huge boon. They've recently added Lua support, so it can transform between Lua and YAML, JSON, CSV, ENV, Properties, XML, etc.

I try to stay away from it, but I also try to write POSIX-compatible scripts, so yq has helped tremendously.

Title: Working with LUA - yq

HCL is another language that I can appreciate, though it's not nearly as common.

rwp: YAML can represent JSON docs without losing information or altering content, whereas the reverse isn't always true.

For very small configs I am not unhappy to use yaml for it but I don't like editing json of any size. I am okay with using json for a data interchange that one never looks at and I feel the same about xml too.

I tend to use msgpack or protobuf or capnproto for machine-produced/machine-consumed data that needs to be serialized/deserialized.