01:08 <adam12> [0x1eef_]: zenspider finds your ryo.rb library interesting

01:26 <[0x1eef_]> Good to know, thanks! :) I have used it in a few projects. It can be awkward but I usually add fixes where I can.

01:28 <adam12> I don't know if I ever looked at it in any type of deep detail before today. It _is_ quite interesting.

01:34 <leah2> 2006, https://leahneukirchen.org/repos/sublanguages/sloop.rb

01:36 <adam12> leah2: wild.

01:37 <leah2> https://leahneukirchen.org/repos/sublanguages/sample/account.sloop.rb haha

01:38 <adam12> My MBA drive died in 2016 and I lost so much cool things I worked on previously but never put on Github or backed up :(

01:39 <[0x1eef_]> leah2: Nice!

01:41 <[0x1eef_]> adam12: I have always had that issue, not just for code but also for configuration files. I tried a few approaches but what I found to work best was a local git server.

01:42 <leah2> i just put anything of value online :p

01:42 <adam12> [0x1eef_]: That's a good idea. I triple check my backups now, but a big issue I had when the MBA drive died was .git folders slowing down the backups at the time (I think I was using Backblaze).

01:42 <[0x1eef_]> xD

01:42 <[0x1eef_]> adam12: Roger

01:42 <adam12> So I ignored ~/code because it was usually all on Github. But nope, lots of experiments never made it :(

01:43 <adam12> And no time to experiment anymore, really.

01:43 <adam12> That said, I'm in between contracts so maybe I'll get to do non-ticket work for a bit :P

01:43 <[0x1eef_]> My goal has been to reproduce my environment as fast as I can if I need to reinstall or get a new computer.

01:44 <adam12> That's an interesting idea. After my MBA died, I _refused_ to buy another Mac because of the keyboards. For a year+ I ran a beefy VM in my rack and just mosh'd in. Worked surprisingly well.

01:45 <[0x1eef_]> The keyboard issue definitely pushed me away as well, and now my views on big corporations have only become worse.

01:46 <adam12> I get it. I voted with my wallet for a long while, but at least for work, I need something that works, and FreeBSD/Linux wasn't getting me that, at least on a mobile device. When I was on desktop I'd only run Linux.

01:46 <adam12> bbl.

01:46 <[0x1eef_]> Makes sense.

01:47 <weaksauce> what's wrong with the keyboard?

01:48 <[0x1eef_]> A few years back it changed and became really unreliable. It couldn't survive the wear and tear of things that happen in life, so it would break easily.

01:48 <weaksauce> i broke my mechanical keyboard and used one of the mac pro keyboards i got from work and it's the shit

01:48 <weaksauce> ah maybe this is more rugged

01:49 <[0x1eef_]> It's probably been addressed by now. It's from a good few years back.

19:18 <henk> I’m writing a tool (ACME client) that needs to do mostly the same thing for multiple domains: check all its authoritative nameservers for a given record. that’s independent, so I thought I’d parallelize it, with threads or something like that. I’m still kinda new to programming. what methods should I be looking at? what’s a 'modern' way to do that?

19:19 <weaksauc_> henk how many domains

19:20 <weaksauc_> if it's not a huge number it's probably fine to do them sequentially

19:22 <bougyman> henk: HTTPx has this built in. It's become my go-to for all http stuffs now. `page1, page2, page3 = HTTPX.get("https://news.ycombinator.com/news", "https://news.ycombinator.com/news?p=2", "https://news.ycombinator.com/news?p=3")`

19:22 <bougyman> tl;dr https://github.com/HoneyryderChuck/httpx

19:32 <henk> weaksauc_: no, not huge, double-digits. and yes: it is fine sequentially but I _want_ to parallelize it. one reason is "scaling" as I hope others will use my client as well and they may have more. the other is for me to learn how to do this. and I’m not very patient (;

19:33 <henk> bougyman: cool, that would solve it for the http challenge, but for now I need it for DNS requests, not HTTP. thank you

19:37 <weaksauc_> henk https://github.com/ruby-concurrency/concurrent-ruby

19:38 <weaksauc_> havenwood might have some input on what's the bleeding edge nowadays

19:38 <weaksauc_> https://awesome-ruby.com/#-concurrency-and-parallelism

19:39 <henk> I keep reading that plain 'Thread's are not truly parallel because of the GIL and some articles say e.g. »threads in Ruby are more suitable for I/O-bound tasks than CPU-bound tasks«. what I want to do is mostly IO-bound, so they should work fine and be a fairly simple approach, is that correct?

19:50 <adam12> Could look at using Async, tho DNS is funny because resolv.rb has been very much single threaded. I am not sure how it would be affected off the top of my head.

19:52 <adam12> Actually, async-dns is a thing.

19:58 <adam12> I extracted this from my monitoring project. https://gist.github.com/adam12/9f4a7100a5fb05e2e6573c1997f02ee8

20:03 <weaksauc_> neat and easy

20:08 <henk> adam12: oh, right. I’m already using dnsruby and that can also do it, it seems: https://github.com/alexdalitz/dnsruby/wiki#asynchronous-queries thanks

20:09 <adam12> henk: That will work.

20:09 <adam12> henk: concurrent-ruby (mentioned above) has a fixed thread-pool where you could feed it queries and wait on the values. I normally wouldn't use a thread pool for network IO like this, but like I mentioned, DNS is funny.

20:23 <henk> hm, I don’t think that’s very well suited for my usecase … I need to keep sending the request until the answer I get is the one I want.

20:25 <henk> I will get an answer in any case, so I would have to re-add the query that does not give the answer I want to the queue. compared with just a thread that keeps sending the query until the result is correct, that seems more tedious to me.

20:25 <henk> but I’m a noob, so I may be wrong. what do you think?

20:46 <weaksauce> why do you have to resend the query

20:55 <henk> weaksauce: with the ACME DNS challenge you prove ownership of the domain by putting a given token in the DNS. I deploy this token to the primary authoritative nameserver which then notifies the secondaries who then request a zone transfer from the primary. all authoritative nameservers need to have gotten that update to ensure that the ACME server will get it when it asks for it. so my client is

20:55 <henk> checking whether it has been propagated to them all and only then tells the ACME server "token deployed, please validate". so as long as one of the authoritative NS responds with an error or the wrong data, i.e. not the given token, I need to keep checking it until it returns the right answer. I’m not sure I explained that very well. let me know if anything about that is still unclear.

20:55 <weaksauce> ah i see henk

22:53 <adam12> henk: Oh I think I see. You could spin up a thread to wait for the value if that's what you're looking to do. It will allow the rest of the program to not be blocked.

22:53 <adam12> You could still get that with Async.

22:53 <adam12> But you'd need to do a bit of work.

23:04 <henk> adam12: I have implemented threading now. probably still a bit 'unclean' but it seems to work fine: https://git.netwichtig.de/cgit/user/henk/code/ruby/macir.git/tree/macir.rb#n134 and https://git.netwichtig.de/cgit/user/henk/code/ruby/macir.git/tree/macir.rb#n211

23:05 <adam12> henk: LGTM

23:08 <henk> btw: any feedback to that code and in general on how to build a proper ruby program are welcome! especially negative criticism, i.e. what’s bad about it. and not just the code but everything around it, e.g. that I have no tests. ideally with links how to do this properly, of course, but the criticism alone would be enough. I’m still a ruby noob and have no clue about conventions or tooling,

23:08 <henk> really, but I do want to create high-quality software!

23:09 <henk> adam12: thank you (: