You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Some MathML elements perform horizontal centering (namely math@display=block, mfrac, munder, mover, munderover).
In #75 we decided to support the width property, allowing the content box to be wider/narrower than the math content box. Current spec essentially says that the math content for the elements mentioned above is centered within the specified width. However, in the case where the width is narrower than the math content and some overflow happens, negative offsets show up and it is no longer possible to scroll to the leftmost part of the math content.
and is fixed in Chrome Canary by clamping the offset of the math content to 0. This can be easily be done for other elements, but we will likely lose the fact that the middle of the children are aligned:
A slightly more complex fix is to shift all the children by the opposite of the smallest offset (negative, so largest in absolute value) so that the offset of the widest child is set to 0 and its respective alignment with the other children is preserved.
The text was updated successfully, but these errors were encountered:
cc @bfgeek @distler
Some MathML elements perform horizontal centering (namely math@display=block, mfrac, munder, mover, munderover).
In #75 we decided to support the width property, allowing the content box to be wider/narrower than the math content box. Current spec essentially says that the math content for the elements mentioned above is centered within the specified width. However, in the case where the width is narrower than the math content and some overflow happens, negative offsets show up and it is no longer possible to scroll to the leftmost part of the math content.
This was reported in Chromium #1416539 for math@display=block:
and is fixed in Chrome Canary by clamping the offset of the math content to 0. This can be easily be done for other elements, but we will likely lose the fact that the middle of the children are aligned:
A slightly more complex fix is to shift all the children by the opposite of the smallest offset (negative, so largest in absolute value) so that the offset of the widest child is set to 0 and its respective alignment with the other children is preserved.
The text was updated successfully, but these errors were encountered: