Re: [XMLHttpRequest-tests] Following RFC2616, assert that there is no message-body, rather than no entity header (#981) from Caitlin Potter on 2014-05-10 (public-web-platform-tests-notifications@w3.org from May 2014)

From: Caitlin Potter <web-platform-tests-notifications@w3.org>
Date: Sat, 10 May 2014 08:48:58 GMT
To: public-web-platform-tests-notifications@w3.org
Message-Id: <<web-platform-tests_issue_981_1399711738945@w3.org>>

>From the Chrome source code (http://mxr.mozilla.org/chromium/source/src/net/http/http_network_transaction.cc#866), there may be a good reason for browsers to not omit a Content-Length header:

```c++
// IE and Safari also add a content length header.  Presumably it is to
// support sending a HEAD request to an URL that only expects to be sent a
// POST or some other method that normally would have a message body.
request_headers_.SetHeader(HttpRequestHeaders::kContentLength, "0");
```

How reasonable that is, well, I dunno. Presumably if Gecko doesn't do it, it's okay, but I dunno.

View on GitHub: https://github.com/w3c/web-platform-tests/pull/981#issuecomment-42736424

Received on Saturday, 10 May 2014 08:49:05 UTC