Colin Douch

@valyala I feel like those proper users must be bot-like as well, right?? There's no way a real person is making 38222 PRs

Top 4 users with the highest number of public pull requests at GitHub during September 2024 are bots: ;> _time:2024-09Z {type=PullRequestEvent} | top (actor.login); executing [_time:2024-09Z {type="PullRequestEvent"} | top by (actor.login)]...; duration: 0.993s actor.login=dependabot[bot] hits=1269903 actor.login=renovate[bot] hits=523478 actor.login=pull[bot] hits=313301 actor.login=github-actions[bot] hits=205211 actor.login=direwolf-github hits=77875 actor.login=optimus-primal[bot] hits=62709 actor.login=leandrohstein hits=38222 actor.login=rmayr hits=37559 actor.login=andrew-nixdrone hits=35601 actor.login=aws-aemilia-pdx hits=24240

Did you know that 30% of public pull requests at @github are generated by bots during the last months according to the data from gharchive.org ?

To distract my self from *life* at the moment, I spent the last few days translating Go’s text/template into Typescript: github.com/sinkingpoint/g… . Why? *Secrets*, but it’s probably useful in the @CloudflareDev ecosystem somewhere Could an AI have done it better? Probably.

Am I the only one that starts to get confused about all the z?? tools around? Zed, Zig, Zod, it's a bit silly

@gotjosh Absolutely! Would be good to catch up :)

@sinkingpoint Woho! Nice to see you back around - would you up for a coffee sometime? I'll drop you a DM.

Minor personal update: I left Cloudflare and have moved back to the UK. If any of y'all want to grab a drink and catch up then reach out!

@bwplotka Thanks! And yes, let's!

@sinkingpoint wowowow, that's a big news! Let's grab some drink soon 😍 ..and congrats!

@ArthurSilvaSens @opentelemetry I think that's the biggest problem - the data model is _fine_, it's the surrounding components that are weird to work with

My biggest struggle with @OpenTelemetry is not the data model, it is the collector's CI 💀 Every PR I've opened has 2-5 random pipelines failing and no clear steps to fix them. I would love to see improvements here.

@boristane Back in my day we had rrdgraph and text based logs and we _liked_ it. Younguns these days with their _events_ and their _context_ - frivolous nonsense! (/s)

I don't know how you guys debug anything without wide events

Feeling some real PromCon FOMO :/ Hope everyone's having fun!

I pasted the entire Grug Brained Developer blog into the system prompt and now my code editor is actually helpful

@Tugzrida Thank goodness on the lax repeating characters limitation. ABBA fans of the world might not have been able to have their preferred password 😅

lol westpac have finally brought their password requirements out of last century

So this is kind of interesting. They have two types of screens - one that displays just text, and one that has a map. These screens are controlled by the front carriage, so this happens when a carriage with the old text screens is fronted by a newer map screen carriage.

@roskilli The weird thing is that it _is_ valid JSON... The only thing I can think of is colour vs color. I guess it errs on displaying something rather than nothing

@sinkingpoint Kind of a crazy fallback, i wonder if there’s some esoteric try/catch and just inlines on failure?

Someone broke JSON parsing on the Melbourne trains 😬

@IanColdwater Well yeah... If we acknowledge how much effort goes into being a good teacher then we might have to *shudders* _pay them more_

Today in Linux esoterica: A Netlink socket is always SOCK_DGRAM, even if you _request_ SOCK_RAW. That means that if you do a `read` with a buffer smaller than a message, you'll just drop data. Ugh.

It'll be interesting to see if we get follow up security incidents from Crowdstrike after this. If they weren't already, you can guarantee that every nation state will be deep diving into it given the breadth of impact

@valyala Interesting idea... My main objection is that in table driven tests, the arguments are named which makes it easier to understand the actual logic of what's being tested. In your f-tests, I have no idea what argument is what, especially if we go beyond non trivial functions

Published an article on how VictoriaMetrics uses f-tests instead of table-driven tests in Go - valyala.medium.com/f-tests-as-a-r…

@boristane There are many, but I fear most are more post-apocalyptic than we would want

is there a future where the world exists without web browsers?

Nothing quite like an impromptu 8 hours in Sydney because my flight came in late 🙄