Stas Lashmanov

8K posts

Stas Lashmanov banner
Stas Lashmanov

Stas Lashmanov

@CyberAP

Staff Frontend Engineer @gitlab (https://t.co/q8MMrM0IEz)

Katılım Mart 2009
64 Takip Edilen318 Takipçiler
Evan You
Evan You@evanyou·
I've left most of what I want to say in the VoidZero blog post. But worth repeating: Thank you @voidzerodev team for trusting me and joining me on this wild ride. I am very proud to have assembled such a talented team and even prouder of what we have built together. Thank you all our investors for believing in my vision, in particular @caseyaylward from @Accel who led both our Seed and Series A. Thank you the @vite_js community. Vite and VoidZero wouldn’t have come this far without your trust and support. We will continue building with all of you, together, in the open. And thank you to everyone that made this happen at @Cloudflare. Looking forward to working with you all! voidzero.dev/posts/voidzero…
English
146
188
2.3K
279.6K
SerKo
SerKo@serkodev·
✨ markdown-exit is finally out! A TypeScript rewrite of markdown-it with built-in types, async rendering and more. ⚡️ Drop-in ready and works with markdown-it plugins. 🔌 #typescript #markdown #web GitHub repo link below 🏃⬇️
SerKo tweet media
English
13
43
467
94.2K
Evan You
Evan You@evanyou·
@CyberAP Do you have a repro for your case? Could be a configuration issue.
English
1
0
0
602
Evan You
Evan You@evanyou·
Great to hear! We are getting closer and closer to production ready for Rolldown-Vite. Vite 8 will be fully Rolldown-powered 🚀 - beta soon!
Nicolas Beaussart@beaussan

Like every couple of weeks, I try @rolldown_rs @vite_js on PayFit's frontend codebase, but this time, it's all green 🎉 And the stats are pretty impressive, from 2m builds down to 8s ! All of that, with nearly no config change (just a polyfill plugin). Maybe it's time to run it in prod 🎉 And it's not a "simple" vite repo, in that repo we have: - @astrodotbuild - @storybookjs with @vitest_dev addon - @vitest_dev with browser mode, and unit mode - And a couple of vite frontend apps Well done to the rolldown team!

English
7
23
499
43.9K
Stas Lashmanov
Stas Lashmanov@CyberAP·
@jaffathecake So .finished resolves exactly 1 frame before the styles return back to initial state or do we get a 1 frame jump before the styles are committed?
English
1
0
0
87
Jake Archibald
Jake Archibald@jaffathecake·
commitStyles() has existed for years, but now in Firefox, it behaves like I always thought it should. Firefox is the first to make this change, but hopefully others will follow soon. developer.mozilla.org/en-US/docs/Web
English
5
0
77
10.9K
Jake Archibald
Jake Archibald@jaffathecake·
Yay! It's the first Firefox release since I've joined the team, so let's take a look at some of the new developer-facing features in Firefox 142… 🧵
English
21
31
526
80.6K
Ryan Carniato
Ryan Carniato@RyanCarniato·
@EOlonov Actually.. hmm I didn't get `.prop` working.. maybe that is for stuff they don't hardcode specifically. In any case `:value` seems to do the attribute which is interesting.
English
1
0
0
613
Ryan Carniato
Ryan Carniato@RyanCarniato·
People say they want their framework to behave like plain HTML. Yet every major framework except(Lit) special cases "value", "checked", "selected" as properties since the setting the attribute doesn't update the input, but sets the default for form reset. Do you prefer:
English
10
3
17
6.8K
Ryan Carniato
Ryan Carniato@RyanCarniato·
The dividing point between "Frameworks" vs "Platform Purists" is the belief of whether DOM should be the source of truth for state. Almost every Web Component argument, issue with design comes back to this. You don't need to go any further than Forms to see this in action.
English
14
6
106
11.4K
Stas Lashmanov
Stas Lashmanov@CyberAP·
@TheAlexLichter @gitlab @rolldown_rs I am still verifying the 100x claim, it might be that the other processes are not tracked like sass compilation. Will experiment more, including disabling CSS assets completely.
English
1
0
7
1.3K
Alexander Lichter
Alexander Lichter@TheAlexLichter·
43x faster than webpack?! Remember how @gitlab got 2.6x faster builds and ~100x less memory consumption when using @rolldown_rs? Now, with all the native plugins enabled, they are down to 22s. That is almost 7x compared to the original 2.5 minutes of Vite, almost halving the build time compared to using only the native resolver before and 43x faster than their webpack build 🔥 MR by @CyberAP: gitlab.com/gitlab-org/git… More performance improvements: github.com/vitejs/rolldow…
Alexander Lichter tweet media
Evan You@evanyou

GitLab is reporting up to 100x less memory used and 2.6x faster builds (2.5m to <40s) with rolldown-vite, without even fully enabling native plugins 🤯 gitlab.com/gitlab-org/git…

English
2
16
245
36.7K
ViteConf
ViteConf@ViteConf·
🚨 GUESS THE SPEAKER! CLUE #2 🚨 🎟️ Your chance for a FREE ticket to ViteConf. Guess to enter the raffle. 👤 A new speaker that’s not on viteconf.amsterdam 🔎 One more clue tomorrow. Big reveal on Monday. 👇 Drop your guess below!
ViteConf tweet media
English
4
7
17
13K
Ryan Carniato
Ryan Carniato@RyanCarniato·
People claim JS frameworks come and go. Sure many are created, but like most software libraries most are never adopted in a serious way anyway. Have you in the last 10 years ever found yourself stuck on a framework no longer actively maintained?
English
27
3
30
15.8K
Stas Lashmanov
Stas Lashmanov@CyberAP·
@evanyou Vanilla JS most definitely can scale with the project. But only if you have well designed architecture, excellent documentation, knowledge sharing across the whole team, automated protection rules, etc. And at this point you're just building your own framework.
English
2
0
6
794
Evan You
Evan You@evanyou·
People have been saying this before all the mainstream frameworks came out (more than a decade ago) Any sweeping statements are likely wrong. - It’s a *very* wide spectrum for what people consider “web apps” - frameworks can help you build most of them, but can be an overkill for simple cases - vanilla js can handle simple cases, but a maintenance nightmare for complex ones. People talk past each other because they only refer to the part of the spectrum they build for when they use the word “web apps”
Lewis Campbell@LewisCTech

Vanilla JS is by far the most sensible thing to build your web app in. The mild discomfort of manual DOM manipulation absolutely pales in comparison to the endless grind of transitive dependency and tooling updates. I will die on this hill. The front-end emperor has no clothes.

English
26
33
627
82.4K
Stas Lashmanov
Stas Lashmanov@CyberAP·
@LeaVerou A modern revision of the document.write for streaming. That can accept readable streams (with backpressure handling) and also start at any element without the need for creating a new document.
English
0
0
1
105
Lea Verou, PhD
Lea Verou, PhD@LeaVerou·
🔥 I have news! After a 1y hiatus, I’m once again leading this year’s #StateOfHTML survey, thanks to generous funding by Google. ❤️ 🙋🏽‍♀️ Is there is an HTML feature or Web API you're just dying to see progress on? Suggest it and it could be included! Browsers fund these surveys to see what matters to devs — so if others care too, it can really help. E.g. these surveys are the reason we have CSS Nesting. ⌛️ Don’t wait too long to suggest your ideas though; the survey is launching later this month, so time is of the essence! lea.verou.me/blog/2025/desi…
English
6
8
55
11.3K
Stas Lashmanov
Stas Lashmanov@CyberAP·
@NickADobos @diabrowser Obvious next step: 'fix the website for me' action. When the form won't send, when the buttons won't click, etc. Would be even more useful for mobile browsers that lock you out of inspector.
English
0
0
1
36
Nick Dobos
Nick Dobos@NickADobos·
Last week @diabrowser launched in beta Here's a preview of what's next A new prototype we're testing & thinking about Should we add this to Dia?
English
139
74
1.2K
236.5K
Andy Allen
Andy Allen@asallen·
Our !Camera launch poster is of course an homage to the famous 1999 iMac "Yum" poster. At a time when everyone else was obsessed with power & efficiency, they made something more fulfilling.
Andy Allen tweet mediaAndy Allen tweet media
English
13
16
582
38.1K
Evan You
Evan You@evanyou·
Btw, with Vite we've seen many cases where people run into the limitation of unbundled ESM even during development. People are building seriously large apps on the web platform. And before you complain about "bloated JavaScript" - we need to acknowledge the difference between websites and web apps. Websites should load fast and have minimal JS, but apps can have arbitrarily large scope and complexity, in many cases surpassing what the web was originally designed for. You can't just tell people building inherently complex apps to "ship less JavaScript." Tooling is what makes such apps possible.
English
19
32
358
74K
Boshen
Boshen@boshen_c·
For the Oxlint 1.0 announcement, I'd like to show more real world performance numbers on large repos. If Oxlint has been satisfactory, would you be keen to send me the numbers Oxlint reports at the end? If you have never tried Oxlint but has a large repo: `npx oxlint@latest`.
English
9
8
117
15.2K
Stas Lashmanov
Stas Lashmanov@CyberAP·
@xanf_ua Красивого и точного нет, а так "Реактивность без холостого хода". Ещё можно "Адресная реактивность", но уже граничит с запутыванием.
Русский
2
0
1
208
Evan You
Evan You@evanyou·
Something is cooking
English
24
8
668
61.8K
Stas Lashmanov
Stas Lashmanov@CyberAP·
@Rich_Harris @timfishy @jaffathecake To calc height and set content visibility for example. I get the idea to delegate user interactions and that works in 90% of the cases. But in these 10% of cases when you need stuff on mount it becomes increasingly more complex to deal with MutationObserver.
English
0
0
0
45
Rich Harris
Rich Harris@Rich_Harris·
this seems extremely cursed
Rich Harris tweet media
English
32
16
619
105.6K