Programming

Bootstrap?

Well for one thing management may only want to look at a few employees on a given day. Can you query all the repos for one employee in one query? Or does it need to be 600 queries?

I had to do my own onboarding because of the rona lol

He’s talking about the CSS framework. I think.

They want, all commits by a user across all of the repositories. From what I’m seeing in the api - but I do need to do some more digging - is that you can grab all commits from a repository, and it gives you one massive blob back that you need to filter. So i need to go repo by repo to get all commits, then filter the results for the front end. That was the plan me and my lead came up with but I hated it.

It’s funny we can accomplish everything we need to do with simple git commands.

lol at bootstrap v bootstrap

1 Like

Report a random number of commits for each dev in some realistic range.

Commits is a great way to get the temperature of who your strongest devs are (unless it’s a dev that always needs to change a lot of lines of code for some reason).

But again - you never let on that you’re looking at that or it makes things much worse. Do the other devs know you’re working on this?

My understanding is no one ever knows what anyones working on here outside of your own team. And my team has stopped doing daily standups because my boss disappeared for a month. But the info and the issue are public.

What’s a good number of commits?

  • I need a framework for a quick front-end project for my company that I’m in charge of despite no front end experience.
  • Bootstrap?
  • I know, right? I even had to do my own on-boarding at this start-up!
  • Bootstrap?
  • It is. The founder made it to breakeven living on ramen, so now he owns all the equity!
  • Bootstrap?
  • Oh no, he came from money. Huge Trumper too

The fuck? Fuck off.

I can’t tell if this stopped being a funny little misunderstanding before, during, or after bobman’s last post.

I mentioned bootstrapping to refer to my onboarding in one of my previous posts and in the flurry of responses/posts misunderstood his (one single word) post.

Not only is the above post completely off the mark on basically every level, it’s dickish and unnecessary.

More than most of the other devs.

OK I have ~1000 users subscribed to an event we’ll call ‘foo’. The list that foo is responsible for changes a lot, maybe a few times a second, maybe once every few seconds. I want my users to see a relatively fresh state of foo. But updating and emitting all of foo’s list every time it changes would cause a massive flood of cpu/network traffic.

So I think I have 2 options:

When foo is updated, see if the current time is less than the last time (held in server memory) by 4 or more seconds, if so, emit foo.

When foo is updated, mark it as updated. Every 4 seconds, foo gets that number, sees if its has been updated since last tick, if so emit foo and mark it as updated.

Which one is better? Am I missing anything?

Run a timer and emit foo if it has been updated since the last time the timer ran. If you go with the first version you could end up with arbitrarily long intervals between emitting, as you might receive an update at 3.999 seconds and then not get another one for another 5 seconds or whatever.

Could you just publish foo every 2 seconds or something? How many different foos are active on the system at one time?

I think a modified form of option 1 is better:

When foo is updated:

  • If you’re outside the lockout period since the last publication, publish.
  • If you’re inside the lockout period and no publication is scheduled, schedule publication for the end of the lockout period.
  • If you’re inside the lockout period and there’s already a scheduled publication, do nothing.

That fixes Chris’s concern about updates being left for arbitrary lengths of time, and also avoids the opposite (albeit smaller) problem where you receive an update at 4.001 seconds and hold it for an additional 3.999 seconds even if you haven’t published for a while.

What is the fastest you can publish updates without affecting performance? Why not just publish that fast and skip all the counting and checks?