Blake Newman

We are hiring Frontend Engineers at @AskAttest There is lots of fun to be had, and a great place to learn and build complex applications with modern tooling. Any questions? AMA here or in DM. askattest.com/careers/openin…

@boshen_c @mileswjohnson Some feedback from me (vscode): - autofix all issues not present, would be a great addition - the version doesn't follow local install of project. Ie we see errors for new rules added to newer version in IDE before we have updated

@mileswjohnson Thank you for your feedback, we'll reconsider IDE prioritization Q1 next year.

Last month we switched to oxlint for the speed, but reverted back to eslint because oxlint's editor experience is a buggy mess. Also using type aware rules forces us to TS v7 which required unwanted changes. I'm looking forward to oxlint but it still feels incomplete.

NPM should have a trusted device system to be able to publish packages. Trusted devices can only publish after 48 hours, and there is alternatives for CI tooling. Maintainers of packages have time to react to breaches and can't be immediately exploited via phishing attacks.

@boshen_c How you getting on! 😲😜

Aim to release type-aware linting on the 28th to celebrate this two years old issue :-)

@jaffathecake Right level of abstraction for public and private functionality in our monorepo packages.

@jaffathecake I think the differentiation should be wild card vs name exports in barrel files. The former is hard for compilers to optimize, and treeshake. The latter has the ability to be statically analyzed efficiently. We use named barrel exports fairly heavily. To provide the

One thing I find barrel files useful for, is creating a single "next interaction" bundle for web app, rather than having an explosion of separate files. rollupjs.org/repl/?version=… vs rollupjs.org/repl/?version=…

@mootoday If this is the slowest part of your ci pipeline then that your problem. They should exist, as easy to bypass. Not on purpose, lots of scenarios where it can happen. Rebasing for example. If its not the slowest part of pipeline then its not a worry.

@sindresorhus Or is it the true value... Before they mask the data

It's an App Store Connect bug...

time to retire

@srsholmes A benefit of opening as a draft PR is you will be notified if you get conflicts, and rebase early. Also gives you the opportunity to ask for early reviews early in implementation, which saves time if you start going down wrong paths.

I once worked with a guy who refused do work in progress PR's. I asked him to create a WIP PR to help him fix a bug he had and he said 'No. It's not finished yet' I said 'The point is not to review it, it is for me to try and understand your bug'. Thoughts about WIP PR's?

@BuckyMaler @TheLarkInn @vlkosinov If the team doesn't get it done / refuses then they should be held accountable and it should affect there performance especially if they agree upfront. Keep to one dependency at the time and when teams are in flow allow more larger pieces of work at a time.

@BuckyMaler @TheLarkInn @vlkosinov So IMO is ownership is done wrong here, ownership in teams means being core maintainers. Teams should be empowered to contribute into other portions of application, so if updating a dep is patching a couple of things here and there then they should do that.

Having multiple versions of _ANY_ package in your monorepo is 🚩🚩🚩. I'm sure each has their reasoning but at the end of the day it is high priority tech debt.

@BuckyMaler @TheLarkInn @vlkosinov If it's a big change then coordinate and set a plan for migrating to new version, allowing teams to set new version but it must be pinned to that minor version. Teams can then migrate independently, and be hold accountable for getting the migration done.

@drosenwasser Created an issue with reproduction here: github.com/microsoft/Type…

@blakenewman Mmmmaybe - can you give me an example? What's not working/what is working in its place?

@drosenwasser I'm trying to use 'bundler' module resolution. I'm noticing that `.d.ts` files can't be resolved in this mode without it explicitly written in path. Is this correct or have I possible got a config issue somewhere.

@TheLarkInn The only legit case I can think for this is a migration, in migrating from a major version to another and allowing owners to do their migration effort. Even then it should be treated as a red flag to prioritise the effort and hold accountability

@LukeDiebold ... generally eliminates these problems as the component tree is shallow. The problem with bad composition is that it's a pain to maintain, change a prop and you need to update a long tree. Also makes it hard to extend and reuse deep component trees

@LukeDiebold I believe it's just a signal of bad component design and composition. Passing props and emitting events isn't bad, but deep component direct composition via imports causes prop drilling and event 'fizzing' to occur. Using slots, state, etc with good component composition...

Just did a #VueJS talk and did some light prop drilling. Some didn't seem to like it. So I looked at a couple of articles on prop drilling and am unconvinced it's a bad idea (at least in my experience where we rarely exceed 3 layers) Why am I wrong?

@evanyou What if this was the real stats? Just the previous count had filtered out incorrectly 😂

Whoever is causing the download spikes for Svelte and Vue please fix it… it’s making the stat useless!

Flashback to the @vuejs London 2019 event, at the 02 Cineworld. Great memories, and very proud of how it turned out. Fun fact - they had technical problems trying to remove our content, so it was on display to the 02 for some time after the event. twitter.com/VueJSLive/stat…

.... When others exploit the community. Could I have done things differently? Yes. Would I now? Probably not. I made memories and hopefully good ones. Still see so many shady organisers claiming it's for community but in reality it's profit margins.

I lost ~60k in total. I did it for the community, and to make a conference that everyone could remember. COVID hit which in the weirdest way saved me from further collapse. I don't regret doing it and I hope anyone attended remembers it for life. It makes me sad when others

surprises me is when I ran Vue London conference is that it was sooo hard to get any interest in sponsorship. How in the world did she bring so many on. FYI in two years of running it, I covered everyone's expenses. From my own personal finances to make sure 0 went without.