- From: Daniel Holbert <dholbert@mozilla.com>
- Date: Tue, 24 Mar 2015 19:32:19 -0700
- To: www-style list <www-style@w3.org>
- Cc: Christian Biesinger <cbiesinger@google.com>
On 03/24/2015 06:44 PM, Daniel Holbert wrote: > On 03/24/2015 06:26 PM, Christian Biesinger wrote: >>> If the item’s computed flex-basis is definite, or the item’s computed flex-basis is content and its computed main size property is definite [...] > Stepping back: so, for that testcase to work, the spec needs to > effectively say: "If you have a flex item with an explicit height & > default flex properties, don't let its min-height:auto value resolve to > anything larger than its explicit height." > > So in the testcase, biesi linked to, where the div has explicit "height: > 38px", the content doesn't matter -- its min-height:auto resolves to (at > most) 38px. > > I think the first half of this spec-sentence (RE the item's computed > flex-basis) is trying to cover that case, but it doesn't quite cover it, > because the item's computed flex-basis is actually "auto". (which means > we "retrieve" the height property and use that for the flex-basis. But > the computed flex-basis is still auto, IIUC.) Tab/fantasai, I think this spec-text might need a tweak. As noted above, I think this spec-text might be assuming that "flex-basis:auto" actually *computes* to the main-size property-value (so that "computed flex-basis is definite" includes e.g. flex-basis:auto;width:50px). But I don't think that assumption holds up. It's not clearly specced that way, at least -- and browsers seem to agree that "flex-basis:auto" just computes to "auto", using this testcase which probes the computed value via inheritance: http://jsfiddle.net/uxrtm0au/ So, I think this sentence might need s/computed/used/. Or alternately, it might just need s/'content'/'content' or 'auto'/. ~Daniel
Received on Wednesday, 25 March 2015 02:32:51 UTC