Index: mojo/public/dart/third_party/http_multi_server/CHANGELOG.md |
diff --git a/mojo/public/dart/third_party/http_multi_server/CHANGELOG.md b/mojo/public/dart/third_party/http_multi_server/CHANGELOG.md |
new file mode 100644 |
index 0000000000000000000000000000000000000000..df1b863bbaa090cef2e936dbc2a582a1a602777a |
--- /dev/null |
+++ b/mojo/public/dart/third_party/http_multi_server/CHANGELOG.md |
@@ -0,0 +1,36 @@ |
+## 1.3.2 |
+ |
+* Eventually stop retrying port allocation if it fails repeatedly. |
+ |
+* Properly detect socket errors caused by already-in-use addresses. |
+ |
+## 1.3.1 |
+ |
+* `loopback()` and `loopbackSecure()` recover gracefully if an ephemeral port is |
+ requested and the located port isn't available on both IPv4 and IPv6. |
+ |
+## 1.3.0 |
+ |
+* Add support for `HttpServer.autoCompress`. |
+ |
+## 1.2.0 |
+ |
+* Add support for `HttpServer.defaultResponseHeaders.clear`. |
+ |
+* Fix `HttpServer.defaultResponseHeaders.remove` and `.removeAll`. |
+ |
+## 1.1.0 |
+ |
+* Add support for `HttpServer.defaultResponseHeaders`. |
+ |
+## 1.0.2 |
+ |
+* Remove the workaround for [issue 19815][]. |
+ |
+## 1.0.1 |
+ |
+* Ignore errors from one of the servers if others are still bound. In |
+ particular, this works around [issue 19815][] on some Windows machines where |
+ IPv6 failure isn't discovered until we try to connect to the socket. |
+ |
+[issue 19815]: http://code.google.com/p/dart/issues/detail?id=19815 |