[css-sizing-4] Remove cyclic a-r section. It's old text, and somewhat confused, from before we laid out the ordering of size transfers more explicitly. (It seems to be assuming that we apply min/max to an axis's size *then* transfer it, which would indeed by cyclic, but that's not how this works.)

Author: tabatkins

css-sizing-4/Overview.bs

@@ -447,40 +447,6 @@ Min/Max Size Transfers</h3>
 		to the width of the word “content”.
 	</div>
 
-<h4 id='aspect-ratio-cyclic'>
-Cyclic Aspect-Ratio Calculations</h4>
-
-	On non-replaced elements with an ''width/auto'' <a>inline size</a>
-	(or any replaced element whose intrinsic <a>block size</a>
-	depends on its <a>inline size</a>),
-	the intrinsic <a>block sizes</a> of the box
-	are assumed to be zero
-	when applying 'min-block-size' to the aspect ratio calculations,
-	and assumed to be infinity
-	when applying 'max-block-size' to the aspect ratio calculations.
-	These limits take effect as usual for non-replaced elements
-	once the <a>inline size</a> is resolved.
-
-	<div class="example">
-		For example, given a (<a>horizontal writing mode</a>) element with
-
-		<pre highlight=css>
-		  width: auto;
-		  min-height: min-content;
-		  aspect-ratio: 1/1;
-		</pre>
-
-		using the aspect ratio to resolve the width
-		would require knowing the min-content height of the element.
-
-		Since for a non-replaced element,
-		this value is not known until the width is resolved,
-		we assume the 'min-height' value is zero
-		while resolving the width.
-
-		Then, once the width is known, we resolve the height,
-		taking into account the ''height/min-content'' minimum.
-	</div>
 
 <!--
 ████ ██    ██ ████████ ████████  ████ ██    ██  ██████  ████  ██████