From 402577677641fd64c9b89696782ffe6ba5e9a6c3 Mon Sep 17 00:00:00 2001 From: Jay McCarthy Date: Fri, 3 Dec 2010 23:47:51 -0700 Subject: [PATCH] Elaboration --- collects/web-server/compat/0/README | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) diff --git a/collects/web-server/compat/0/README b/collects/web-server/compat/0/README index 260a9f6004..c82c2938be 100644 --- a/collects/web-server/compat/0/README +++ b/collects/web-server/compat/0/README @@ -5,10 +5,13 @@ conversion of X-expressions and lists with the format (cons/c bytes? The compatibility binding for normalize-response is a coercion from the OLD responses to the NEW response structure. -Other incompatibilities introduced: +Other incompatibilities introduced without compatibility bindings: -The contract on read-mime-types specifies what kind of hash it returns. send/formlet requires that the wrapper return an Xexpr + +Other incompatibilities introduced with compatibility bindings: + +http/response-structs: response/basic was removed. response/full was removed. response/port was removed [1] @@ -16,6 +19,8 @@ The response/incremental structure was removed. response/c was removed. make-xexpr-response was renamed response/xexpr and extended. normalize-response was removed. + +http/cookie: xexpr-response/cookies was removed and folded into response/xexpr. The following places are where old responses were accepted and no longer are: @@ -49,7 +54,7 @@ Here are some ideas that could be use to ease backwards compatibility: Footnotes: -1. response/port was present for about a week, so no compatibility is provided. +1. response/port was present for a few weeks, so no compatibility is provided. 2. If anyone can suggest a good way to provide a compatibility layer without duplicating code, I'm interested. The problem is that