Bugzilla

Re comment 17:
> So: the next-best option I can think of to address the fundamental issue here is to use an int64_t (with its larger coordinate space) for tracking the flex container's space-to-be-distributed (which is hugely negative value in this case).  This might actually make for more robust results in general, so I'm tempted to do this.

Do you feel it's worth adding an int64_t strong type for the above usage? It can also serve the purpose to explore bug 1602669 as its APIs should be similar to strong type `nscoord`.

Re comment 17:
> So: the next-best option I can think of to address the fundamental issue here is to use an int64_t (with its larger coordinate space) for tracking the flex container's space-to-be-distributed (which is hugely negative value in this case).  This might actually make for more robust results in general, so I'm tempted to do this.

Do you feel it's worth adding an int64_t strong coordinate type for the above usage? It can also serve the purpose to explore bug 1602669 as its APIs should be similar to strong type `nscoord`.