[webkit-dev] MathML layout refactor proposal

Alejandro Garcia Castro alex at igalia.com
Mon Dec 28 03:06:36 PST 2015


Hi Dave,

thanks very much for your comments and feedback, this information is
great to document the whole design process of the code.

I agree with the goals you are pointing out, actually I think the
change to use flexbox was a good option to try at that point. But
currently considering the result of the design in the code we can see
that the devil is in the detail.

Trying to use box spacing to create complex anonymous render trees
that produce the complex layout result of some mathml formula, brings
a complex code difficult to improve to get the best math render
possible. That is the main reason for this change, you can see bugs
such as this were developers are not even sure how the whole layout is
happening:

https://bugs.webkit.org/show_bug.cgi?id=137330

The problem is worse if we try to add more complex layout calculations
to match best formula rendering.

Also we can find flexbox changes just avoiding mathml tests because
it is not clear how the anonymous nodes structure is handled for some
cases.

So this takes us to the current proposal, if we do not get more
feedback probably we will start asking for reviews in the next weeks,
and we can discuss in more detail in the bugs the pros and cons of the
change.

Thanks again,

Alex


More information about the webkit-dev mailing list