uhoh-itsmaciek

I've also been blocked from your blog. Firefox on Android.

So what are you proposing? Force people to marry and have children? Ignore the problem?

Yeah, that was kind of ignored in the article and it's definitely an issue for some platform features. Date pickers are another one where it's easy to outgrow the native implementation.

So allowing projects like Nitter and XCancel must also be understandable, in order for recipients to be able to actually read these messages.

In Firefox, the page tells me neither my browser nor my display support HDR. In Chrome, it works fine. Latest browsers on a Pixel 8.

This doesn't really scale, and it doesn't really solve the problem for your (potential) users to wash your hands of software delivery like this.

Fair enough, but that move should make you think twice before tying your fate to theirs.

I ran into this problem recently on one of our blog posts: we used some Claude output which included tables drawn with Unicode line drawing characters. However, our monospace font did not include these characters, and so rendering fell back to another font in our font stack with different width metrics. I fixed it by using a font that had similar metrics and did include those characters with `unicode-range` (to only select characters we needed) and `size-adjust` (to match font width more exactly), and adding it to the stack. It's a little hacky but works pretty well in practice.

Could you elaborate?

Great point, but if the 2% are served graceful degradation rather than a broken site, that's probably okay.

Right. Backups don't matter. Restores matter.

HN submission titles don't need emoji.

>Without those periodic full page images in the log, the storage layer would have to replay an infinitely long chain of small deltas to reconstruct a page for a read request. What was once a bounded O(checkpoint frequency) replay becomes an unbounded chain, leading to a spike in read latency and resource consumption.

I don't follow: read requests are not served from the WAL. They read the current state of the page from the buffer cache, where the page is updated after the change (FPI or not) is written to the WAL.