0.19.4

CHANGELOG.md

@@ -1,4 +1,9 @@
-# 0.19.3 - 2016-11-21
+# 0.19.4 - 2016-11-24
+
+- 🐛 Fixed: click on anchors are now scrolling has expected
+([#665](https://github.com/MoOx/phenomic/issues/665) - @MoOx)
+
+# 0.19.3 - 2016-11-22
 
 🍭 This bugfixes are a good thing and have been detected since we use 2
 different Phenomic projects for our documentation website:
@@ -22,16 +27,16 @@ We might revisit this issue when refactoring how collection is handled and
 rendered.
 (@MoOx)
 
-🐛 Fixed: when a ``<Link>`` explicitly receive an absolute url from the same
+- 🐛 Fixed: when a ``<Link>`` explicitly receive an absolute url from the same
 domain, but not the same project base path, a normal ``<a>`` is rendered
 (@MoOx)
 
-# 0.19.2 - 2016-11-21
+# 0.19.2 - 2016-11-22
 
 **broken**: npm published a incomplete package which resulted in an error during
 installation: ``Cannot find module '...phenomic/npm/postinstall.js'``
 
-# 0.19.1 - 2016-11-21
+# 0.19.1 - 2016-11-22
 
 - 🐛 Fixed: static build does not write phenomic assets as if they were handled by react-helmet.
 This fix the issue where phenomic CSS are removed from the page when

package.json

@@ -1,6 +1,6 @@
 {
   "name": "phenomic",
-  "version": "0.19.3",
+  "version": "0.19.4",
   "description": "Modern website generator based on the React and Webpack ecosystem.",
   "keywords": [
     "react",

src/client/should-update-scroll.js

@@ -8,18 +8,22 @@ type RouterPropsType = {
   }
 }
 
-// we skip scroll update if only hash changes
+// only scroll when complete url changes
+// (except for hash change - this example provide a natural feeling when you
+// use simple link to internal anchor (eg: a table of content that use hashes
+// inside a page))
 const shouldUpdateScroll =
 (prevProps: RouterPropsType | null, props: RouterPropsType): boolean => {
   if (prevProps) {
     const prev = prevProps.location
     const curr = props.location
     // if "page" is the same we don't scroll
     if (prev.pathname + prev.search === curr.pathname + curr.search) {
-      // except if hash is empty, because it should scroll if
-      // - same url, without hash
-      // - hash to no hash
-      // both cases are kind of a "back to top"
+      // except if hash is empty, because it should keep browser natural
+      // behavior if:
+      // - same url, without considering the hash
+      // - previous url contains a hash to an new url does not
+      // both cases are kind of a "back to top" scenario
       if (curr.hash === "") {
         return true
       }