http parser problem in V 0.8.14

[kaoshijuan]

i created an http client to get the URL http://www.simsimi.com/func/req?msg=test&lc=ch , and the response made the node got an error HPE_INVALID_HEADER_TOKEN, i'm wondering there is something wrong with the parser of the http protocol.

it's OK while I use the browser like IE or chrome, and it also works well while using Python 2.7.3 to access this URL.

[bnoordhuis]

The server returns an invalid response.

$ curl -v 'http://www.simsimi.com/func/req?msg=test&lc=ch'
* About to connect() to www.simsimi.com port 80 (#0)
*   Trying 50.16.220.34... connected
> GET /func/req?msg=test&lc=ch HTTP/1.1
> User-Agent: curl/7.22.0 (x86_64-pc-linux-gnu) libcurl/7.22.0 OpenSSL/1.0.1 zlib/1.2.3.4 libidn/1.23 librtmp/2.3
> Host: www.simsimi.com
> Accept: */*
> 
< HTTP/1.1 200 OK
< Cache-Control: no-cache
< Content-Language: en
< Content-Type: text/html;charset=UTF-8
< Date: Mon, 10 Dec 2012 09:36:43 GMT
< Expires: Thu, 01 Jan 1970 00:00:00 GMT
< Pragma: no-cache
< Server: Apache-Coyote/1.1
< Set-Cookie: JSESSIONID=095D7ED1038F217E6148B3321F6821B2; Path=/; HttpOnly
< text/html: utf-8
< Content-Length: 55
< Connection: keep-alive
< 
* Connection #0 to host www.simsimi.com left intact
* Closing connection #0
{"response":"hi","id":1466090,"result":100,"msg":"OK."}

Note the line that says text/html: utf-8. The HTTP parser in node.js is unforgiving of such spec violations.

[kaoshijuan]

I think the parser should be more robust and it can tolerate the violations.

[indutny]

For what reason? Also you can always use raw net interface and implement your own http client on top of it.

Parser is intolerant exactly for robustness and performance reasons.