[css-grid] fit-content with indefinite percentage

[Loirooriol]

According to https://drafts.csswg.org/css-grid/#valdef-grid-template-columns-length-percentage,

If the size of the grid container depends on the size of its tracks, then the <percentage> must be treated as auto, for the purpose of calculating the intrinsic sizes of the grid container and then resolve against that resulting grid container size for the purpose of laying out the grid and its items.

But what if the indefinite percentage is inside fit-content()? https://drafts.csswg.org/css-grid/#valdef-grid-template-columns-fit-content doesn't explain this explicitly, and it can't just be treated as auto like in <length-percentage> because fit-content(auto) is not valid syntax.

I guess it should be treated as a normal minmax(auto, max-content), without the special fit-content behavior? That would match this fit-content description:

This is essentially calculated as the smaller of minmax(auto, max-content) and minmax(auto, limit).

[fantasai]

Yes. The invalidity of fit-content(auto) is because such a syntax would be meaningless: it's roughly equivalent to just auto. I don't think the spec is inconsistent or unclear about what happens here, but if you think something needs fixing lmk?

[Loirooriol]

Yes, I would like https://drafts.csswg.org/css-grid/#valdef-grid-template-columns-fit-content to explain what to do with indefinite percentages, just like https://drafts.csswg.org/css-grid/#valdef-grid-template-columns-length-percentage does.

If it's supposed to behave like minmax(auto, max-content), explicitly say so. Seems the most reasonable, but e.g. auto could also make sense. I believe there is no difference in practice, since when the align-content and justify-content properties are used, the percentages should be definite. But it should be properly defined anyways.

[fantasai]

If the two possible interpretations you can come up with both yield the same behavior, then does it really require clarification?

(This section of the spec is referenced by authors. If it's not necessary for correctness to add a note about obscure details, it's probably better to leave them out.)

If it's not clear clear that the argument to fit-content() is to be interpreted as a track size like any other track size, though, I can clear that up... d110d63

[Loirooriol]

If the two possible interpretations you can come up with both yield the same behavior

Well I'm not completely sure if they always yield the same behavior. Can you confirm if it's the case? I still haven't pored over the part about track expansion by align-content and justify-content.

If it's not clear clear that the argument to fit-content() is to be interpreted as a track size like any other track size

But the problem is that this is not really the case, because indefinite percentages are treated as auto, but fit-content(auto) is not allowed. I guess if it existed it would behave according to

This is essentially calculated as the smaller of minmax(auto, max-content) and minmax(auto, limit).

and AFAIK minmax(auto, max-content) <= minmax(auto, auto), so the fit-content would be treated as minmax(auto, max-content)

So if you want to leave the spec as-is it's not that bad but I don't think it's completely clear.

[fantasai]

Well I'm not completely sure if they always yield the same behavior. Can you confirm if it's the case?

Well, yeah. auto track size >= max-content, and fit-content() intrinsically imposes a max-content maximum.