Bugzilla

Re comment 11:

> I actually think it feels right for that to always be auto on the table-wrapper frame, even in cases where the author has specified some min-width/min-height value on the table, since tables don't support being smaller than their content-size anyway; and if the specified value is larger, then we'll still pick that up & honor it when we measure content to resolve the auto size on the table wrapper.

I feel this make sense to me. We probably shouldn't look into inner table's property except when absolutely necessary. I'll explore your idea.

Re commend 15:

Thank you for taking a look into these testcases, and giving the outline of the possible fix. You're right that original Part 5 (remove ComputeSizeFlag::UseAutoISize) is trying to prevent inner `<table>`'s specified size from influencing the content-based minimum size.

Re comment 16:

Yeah, its painful to fix this bug if we want to consider all the non-standard `caption-side` value, but bug 1688695 probably won't be fix soon. I think we should at least make the standard `caption-side` correct without worrying too much about breaking those non-standard ones.

Re comment 11:

> I actually think it feels right for that to always be auto on the table-wrapper frame, even in cases where the author has specified some min-width/min-height value on the table, since tables don't support being smaller than their content-size anyway; and if the specified value is larger, then we'll still pick that up & honor it when we measure content to resolve the auto size on the table wrapper.

I feel this make sense to me. We probably shouldn't look into inner table's property except when absolutely necessary. I'll explore your idea.

Re comment 15:

Thank you for taking a look into these testcases, and giving the outline of the possible fix. You're right that original Part 5 (remove ComputeSizeFlag::UseAutoISize) is trying to prevent inner `<table>`'s specified size from influencing the content-based minimum size.

Re comment 16:

Yeah, its painful to fix this bug if we want to consider all the non-standard `caption-side` value, but bug 1688695 probably won't be fix soon. I think we should at least make the standard `caption-side` correct without worrying too much about breaking those non-standard ones.