- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Wed, 16 Dec 2009 21:14:02 -0600
- To: "L. David Baron" <dbaron@dbaron.org>
- Cc: www-style <www-style@w3.org>
On Wed, Dec 16, 2009 at 6:17 PM, L. David Baron <dbaron@dbaron.org> wrote: > 1. While parsing the contents of the calc(), values are accepted > even if, were they the lone value of the property, they would be > rejected due to their sign (or any other range restriction on the > value of the property). > > For example: > width: calc(50% + -2px); > is a valid declaration even though -2px is not a valid value for > the 'width' property. This seems like it should be possible in the first place. Does calc() current disallow this? If so, then it should definitely be changed to allow this; this construct is pretty necessary for a lot of the uses calc() will be put to. > 2. For properties whose values have restricted ranges, the computed > value of the calc() expression is clamped to the allowed range. > > For example: > width: calc(50% - 200px); > computes to zero when the containing block's width is less than > or equal to 400px (whereas without the clamping it would become > negative when the containing block's width is less than 400px). This seems unobjectionable. I highly doubt that an author will *ever* actually want their calc() expression to become invalid some of the time. It seems much more likely that I'll write an expression like this and implicitly expect it to become 0 when the box is too small. Requiring an explicit max() call in there seems unnecessary. ~TJ
Received on Thursday, 17 December 2009 03:15:30 UTC