Closed candid82 closed 4 years ago
I'm having a hard time finding documentation about this IE 10 bug. Could you point me to some writeup about the existence of the quirk or documentation about what IE 10 reads -ms-flex: auto
as?
Hum... Just noticed this at the bottom of that document: "The third flex item has auto defined for the flex property. This is equivalent to flex: 1 1 auto.". This kind of conflicts with "if omitted" behavior specified for flex-grow and flex-shrink parts, but I guess this issue is fixed in IE11. Older document for the prefixed version (supported by IE10) doesn't mention anything about "flex: auto" case: http://msdn.microsoft.com/en-us/library/ie/hh673531(v=vs.85).aspx. I definitely saw this issue in IE10. Unfortunately I no longer have a test environment for IE10, so cannot double check.
ah, this is noted on http://caniuse.com/#feat=flexbox:
IE10 and IE11 default values for flex are
0 0 auto
rather than0 1 auto
, as per the draft spec, as of September 2013.
And this question on StackOverflow: http://stackoverflow.com/questions/16808815/ie10-flexbox-p-element-non-wrapping
So, I think that if the default is 0 1 auto
we should be applying that (not 1 1 auto
), right?
It's been a while since I looked at this issue, but I think default != auto. According to the MDN, auto
means 1 1 main-size
(they changed auto
to main-size
since last time I read it.... sigh). I guess the closest to that would be to translate auto
to 1 1 auto
in IE 10.
flex: auto
should be translated to-ms-flex: 1 1 auto
because-ms-flex: auto
is interpreted as-ms-flex: 1 0 auto
by IE 10.