[css-inline-3] Behavior of initial letter property with 1st argument value set to [0,1)

[chenpighead]

Per https://drafts.csswg.org/css-inline/#sizing-drop-initials, it says the first argument (size of a initial letter) accepts non-negative . If I read correctly, it means setting initial-letter with size [0, 1) is allowed. However, I wonder: (1) what is the use case for setting initial-letter's size to less than 1? (2) what is the expected behavior for setting initial-letter's size to 0?

Would it be better to specify that the first argument should be "larger or equal to 1"?

[upsuper]

In the section "3.5. Sizing Initial Letters", the formula for calculating the font size uses (N-1) * line-height, which means when the number is in [0, 1), this part of the formula would be negative, and if line-height is large enough, the result font size of that formula could be negative as well, which seems unfortunate.

Given this, I believe range of that number should be restricted to [1, +∞).

[frivoal]

I agree, values less than 1 don't seem useful. @dauwhe, are we missing something?

[dauwhe]

It appears Webkit ignores values less than one for that first argument. I'm fine with changing the spec.

[upsuper]

What's the exact meaning of "ignores values" here? Does it reject the value when parsing, or accept the value but use or compute it to a different value?

[chenpighead]

Looks like it rejects the value when parsing. It accepts positive integer only, since the non-integer part of the argument specified in the spec. is not implemented yet.

No idea where to paste the source link to webkit. Here are some references: [1] is a github unofficial mirror pointing to the parser for webkit-initial-letter. [2] is the original implementation patch.

[1] https://github.com/WebKit/webkit/blob/master/Source/WebCore/css/CSSParser.cpp#L2545
[2] https://bugs.webkit.org/attachment.cgi?id=237561&action=prettypatch