|
Šime Vidas
@
simevidas
Zagreb, Croatia
|
|
🤓 I create news content for web developers on webplatform.news (@WebPlatformNews).
|
|
|
16.357
Tweetovi
|
339
Pratim
|
4.225
Osobe koje vas prate
|
| Tweetovi |
|
Šime Vidas
@simevidas
|
8 h |
|
Oh wow, they *did* write Ul not UI. I blame the San Francisco font. 😂
|
||
|
|
||
|
Šime Vidas
@simevidas
|
8 h |
|
Ah, Mikael meant UL, not UI. That confused me.
|
||
|
|
||
|
Šime Vidas
@simevidas
|
8 h |
|
What do you mean? There is on instance of <li class="special"> in my HTML code.
|
||
|
|
||
|
Šime Vidas
@simevidas
|
8 h |
|
<ul>
<li>
<a>Link</a>
</li>
<li class="special">
<a>Link</a>
</li>
<li>
<a>Link</a>
</li>
</ul>
:not(.special) a {
color: red;
}
How many links are red?
|
||
|
|
||
|
Šime Vidas
@simevidas
|
11 h |
|
I think the only quick way to learn webpack is have someone who has already done it explain it to you in person. Everything else takes literal days 😂
|
||
|
|
||
|
Šime Vidas
@simevidas
|
11 h |
|
For me the big thing is that Duck doesn’t remember my searches. I can search freely without thinking of consequences. It’s quite liberating.
|
||
|
|
||
|
Šime Vidas
@simevidas
|
2. velj |
|
That section explains that ::before/::after content is effectively the same as inserting a <span> into the element.
Since <img> has a content model of “nothing” (according to the HTML spec), inserting a <span> into it isn’t valid, which means ::before/::after isn’t valid either.
|
||
|
|
||
|
Šime Vidas
@simevidas
|
2. velj |
|
Good point. I didn’t consider that scrolling doesn’t count as input/interaction. I keep forgetting this.
Janky scrolling during page load is a common occurrence for me but this issue isn’t related to FID.
|
||
|
|
||
|
Šime Vidas
@simevidas
|
2. velj |
|
It’s all simulated throttling, so it depends on the hardware that runs the test. I prefer PageSpeed Insights over the others because it includes field data.
|
||
|
|
||
|
Šime Vidas
@simevidas
|
2. velj |
|
*buys a Switch Lite at the airport*
|
||
|
|
||
|
Šime Vidas
@simevidas
|
2. velj |
|
The difference is just too much. It’s contradictory.
I suspect that First CPU Idle (which is deprecated!) is just broken, i.e., its algorithm doesn’t work anymore.
Google should probably remove it from performance reports.
|
||
|
|
||
|
Šime Vidas
@simevidas
|
2. velj |
|
According to the docs (web.dev/first-cpu-idle/), First CPU Idle (a deprecated metric!) is directly tied to input delay.
In this case, it takes 20 seconds for the page to start responding to user input “in a reasonable amount of time.”
So it’s surprising that FID is go good.
|
||
|
|
||
|
Šime Vidas
@simevidas
|
2. velj |
|
This reminded me of Time to Interactive, which has this problem: webplatform.news/issues/2019-03….
|
||
|
|
||
|
Šime Vidas
@simevidas
|
2. velj |
|
|
||
|
Šime Vidas
@simevidas
|
2. velj |
|
Lab data: CPU is not idle for over 20 seconds.
Field data: The vast majority of visitors experience almost no input delay.
I’m not sure how to interpret this… pic.twitter.com/9Z79KWKccs
|
||
|
|
||
| Šime Vidas proslijedio/la je tweet | ||
|
Tyler Sticka, ☁️4
@tylersticka
|
28. sij |
|
Last month, one of my apps temporarily vanished from the App Store for some pretty silly reasons. 🤔
Silver lining: The experience reminded me why I ❤️ the web! 🙌 cloudfour.com/thinks/spoiled…
|
||
|
|
||
|
Šime Vidas
@simevidas
|
2. velj |
|
“Any impacted students will be forced to switch to a different browser.”
That doesn’t sound that bad. 😂
|
||
|
|
||
| Šime Vidas proslijedio/la je tweet | ||
|
Florens Verschelde
@fvsch
|
1. velj |
|
You don’t have to open the main @FirefoxDevTools pane to use the Eyedropper. If you put the "Developer" button in your toolbar, it‘s two clicks away. pic.twitter.com/OfKvuZf0g7
|
||
|
|
||
|
Šime Vidas
@simevidas
|
1. velj |
|
Tell Kotaku to put a The in your title on the frontpage pic.twitter.com/2dL1oyiHIC
|
||
|
|
||
|
Šime Vidas
@simevidas
|
1. velj |
|
How about the term “site domain” for domains that define websites, like w3.org and w3c.github.io?
|
||
|
|
||