https://www.w3.org/TR/css-flexbox-1/#item-margins
"Note: This variance sucks, but it accurately captures the current state"
@Nadya678, this whole issue has been resolved in #2085. The whole note about the "variance" has been removed from both [css-flex-1] and [css-grid-1] specs, since the single option for calculating margins/paddings of flex/grid items has been chosen. All browsers decided to switch to Chrome's behavior.
P.S. Please cite the latest Editor's drafts (e.g. https://drafts.csswg.org/css-flexbox/#item-margins), which usually incorporate the most recent edits, instead of /TR/ publications, which can be rather outdated.
Not to mention, "sucks" accurately captured the suckitude of the variance.
All browsers decided to switch to Chrome's behavior.
Well, that's a I nice euphemism... The harsh truth is that Chrome's developers refused to implement what the specification said (which was a consensus decision in the CSSWG, btw) thereby _forcing_ other UAs to change for web-compatibility reasons. Shame on them.
It definitely was not a WG consensus, pretty much by definition, as Chrome/WebKit disagreed with Firefox/Edge. Please don't rewrite history.
They disagreed _after_ the decision was made. Meanwhile they shipped an intentionally non-conformant implementation thereby establishing fait accompli due to their dominant market position. _Then_ the spec was updated to allow either behavior. Please don't rewrite history.
Find some other venue if you want to continue this.
Most helpful comment
Well, that's a I nice euphemism... The harsh truth is that Chrome's developers refused to implement what the specification said (which was a consensus decision in the CSSWG, btw) thereby _forcing_ other UAs to change for web-compatibility reasons. Shame on them.