add breaking changes 6.2 recipe

docs/recipes/breaking-changes-6-2.md

@@ -0,0 +1,34 @@
+<!--
+{
+  "title": "Breaking Changes Between Lucee 6.1 and 6.2",
+  "id": "breaking-changes-6-1-to-6-2",
+  "categories": ["breaking changes", "migration","compat"],
+  "description": "A guide to breaking changes introduced in Lucee between version 6.1 and 6.2",
+  "keywords": ["breaking changes", "Lucee 6.1", "Lucee 6.2", "migration", "upgrade"],
+  "related": [
+    "mathematical-precision"
+  ]
+}
+-->
+
+# Breaking Changes between Lucee 6.1 and 6.2
+
+This document outlines the breaking changes introduced when upgrading from Lucee 6.1 to Lucee 6.2. Be aware of these changes when migrating your applications to ensure smooth compatibility.
+
+## Changing PreciseMath to be off by default
+
+With Lucee 6, we introduced support for higher precision maths, by switching the underlying Java class from Double to BigInteger.
+
+During the development of 6.2, which was heavily focussed on increasing Lucee's overall performance, it became apparent that the overhead of BigDecimal, both in terms of performance and memory usage was too simply high.
+
+So, we decided to switch the default back to preciseMath=false, which really improved performance as this affects any use of numbers. For most Lucee applications, this change will have no functional affect.
+
+Dynamically switching preciseMath on and off as required is recommended, rather than using it all the time.
+
+As part of this change, we updated all our test cases to test switching dynamically during a request and identified a few problems which we addressed.
+
+## Lucee is up to 50% for some operations than Lucee 5.4
+
+While not exactly what you might expect as a breaking change, we did find that all the improvements made with 6.2 managed to surface some other underlying bugs, simply because Lucee got faster.
+
+You also may find some race conditions, etc within your own code / applications.

docs/recipes/mathematical-precision.md

@@ -24,7 +24,7 @@ This makes math operations much more precise (but slower) and there is no need a
 
 Since version 6.0, all numbers Lucee uses in the runtime are by default BigDecimal based and no longer double as before. 
 
-However, for performance reasons, with Lucee 6.2, we reverted the default to the old behaviour as it's much faster.
+However, for performance reasons, with Lucee 6.2, we reverted the default to the old behavior as it's much faster.
 
 You can toggle precise math in the `Application.cfc` as follows: