Re: [csswg-drafts] [cssom] Resolved value for font-size should perhaps take user's minimum font-size into account (#10479)

The CSS Working Group just discussed `[cssom] Resolved value for font-size should perhaps take user's minimum font-size into account`, and agreed to the following:

* `SUMMARY: Dig into compat issues to see what's likely to break for which behaviors`

<details><summary>The full IRC log of that discussion</summary>
&lt;lwarlow> q+<br>
&lt;fantasai> emilio: I think we consciously made it not apply to 1em resolution.<br>
&lt;astearns> ack lwarlow<br>
&lt;fantasai> lwarlow: Is the user's minimum font size exposed anywhere else already? Maybe adds more fingerprinting behavior.<br>
&lt;fantasai> emilio: It is exposed because it affects geometry, so even if it isn't exposed in gCS() it's easy to get<br>
&lt;fantasai> lwarlow: ok then it's fine<br>
&lt;fantasai> fantasai: 3 things generally correspond: used font-size, getComputedStyle(font-size), and resolution of 1em.<br>
&lt;fantasai> fantasai: Breaking connection across any of these would be weird<br>
&lt;fantasai> emilio: We disconnected from 1em because some people made the used font size tiny to have a unit they wanted<br>
&lt;fantasai> fantasai: Wouldn't that be 'rem'? Not 'em'<br>
&lt;fantasai> emilio: If you have a very small font size, and then that font-size 2em ...<br>
&lt;fantasai> emilio: It could go both ways<br>
&lt;astearns> ack dbaron<br>
&lt;fantasai> dbaron: I have a vague memory that some of the motivation might have been 'em'<br>
&lt;fantasai> dbaron: Some of the prefs are for a rather large minimum font size, e.g. 36px<br>
&lt;fantasai> dbaron: One of the goals for that case was, if the page had headers that were 3x the font-size, the 36px minimum wouldn't actualy make the headers bigger<br>
&lt;fantasai> dbaron: This ridiculously large font-size wouldn't inflate the headers to e.g. 108px<br>
&lt;fantasai> dbaron: There's tradeoffs both ways, but one of the use cases for this was users who really need rather big text<br>
&lt;lwarlow> Doesn't that apply to the used value too not just computed though?<br>
&lt;fantasai> dbaron: but don't want to inflate all of the text on the page, even text that's already big<br>
&lt;flackr> +1 that makes sense to me<br>
&lt;fantasai> fantasai: I think we need to distinguish between the resolution of 'em' on 'font-size' (which looks at the parent) vs 'em' on the element itself.<br>
&lt;fantasai> fantasai: You want to resolve 'font-size' without considering the parent's minimum font size<br>
&lt;fantasai> fantasai: but apply the minimum when resolving 'em' units on the element itself<br>
&lt;fantasai> fantasai: Otherwise e.g. 'line-height: 1em' would not be big enough for the text that's been inflated<br>
&lt;fantasai> emilio: [something about handling this case already]<br>
&lt;flackr> fantasai: it's not just line-height, there's vertical-align, underlying properties, width of box fitting some amount of content<br>
&lt;flackr> fantasai: if the site is using ems for sizing, most things relative to size of font should stay relative to size of font<br>
&lt;flackr> fantasai: question is are there sites doing weird things with ems that would cause breakage<br>
&lt;fantasai> fantasai: and is that breakage worse than the breakage of *not* scaling 'em' on sites that are doing proper use of units<br>
&lt;fantasai> astearns: Seems like we need more info.<br>
&lt;fantasai> smfr: I'd like to debug the issue reported against WK first.<br>
&lt;fantasai> smfr: Do we have sites where this is a problem? I've seen some bug reports but haven't investigated yet<br>
&lt;fantasai> emilio: I can do some digging about that<br>
&lt;flackr> fantasai: we've got browsers with diff behaviors, the sites that are broken are like the opposite of that? [missed]<br>
&lt;fantasai> astearns: ok let's take back to issue<br>
&lt;fantasai> SUMMARY: Dig into compat issues to see what's likely to break for which behaviors<br>
</details>


-- 
GitHub Notification of comment by css-meeting-bot
Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/10479#issuecomment-4345379497 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Wednesday, 29 April 2026 15:58:10 UTC