[hunchentoot-devel] Fwd: Intermittent problem with codequarterly website (chunked coding glitch?)

Fri Jun 10 06:50:11 UTC 2011

> Any ideas of the top of anyone's head?

I think the problem is in any case would be solved if always specify
the "Content-Length" and not use the "Transfer-Encoding: chunked"

Andrey

2011/6/10 Peter Seibel <peter at gigamonkeys.com>:
> I have no particular reason to believe Hunchentoot is at fault here
> but I figured I'd ask in case anyone had seen anything like this. The
> codequarterly.com website is being served by Apache proxying to
> Hunchentoot (via ProxyPass and ProxyPassReverse directives in the
> Apache config) and this user reports occasionally getting back some
> hinky chunked-coded responses.
>
> Any ideas of the top of anyone's head?
>
> -Peter
>
>
>
> ---------- Forwarded message ----------
> From: Thomas (AWS), Chris <ccthomas at amazon.com>
> Date: Thu, Jun 9, 2011 at 6:14 PM
> Subject: Intermittent problem with codequarterly website (chunked
> coding glitch?)
> To: "editor at codequarterly.com" <editor at codequarterly.com>
>
>
> Greetings,
>
> This may not be the appropriate place to report this issue, but I
> couldn't find any technical contact information on your website.  Feel
> free to forward as necessary.
>
> Intermittently (maybe 10-20% of the time), when trying to view pages
> on the codequarterly website, my browser is displaying what appears to
> be the text of the raw HTML of the page, instead of the rendered page.
>  When I look closer, I see that the text begins with a hexadecimal
> number (e.g, '1FED'), followed by the usual HTML <DOCTYPE....  Header.
>
> I reproduced the problem with curl  & took a packet trace (output at
> end of email), to better see what is going on:
>
> % curl -v http://www.codequarterly.com/2011/rich-hickey/
>
> It shows that the response is coming back with chunked coding, but
> there are two hexadecimal numbers preceding  the actual HTML content,
> instead of just the single number that one would expect from a
> chunked-coded response.  The remainder of the body shows additional
> "extra" chunk length prefixes.  It is almost as if the HTTP server was
> sending the response body through two passes of chunked encoding.
>
> Anyway, you might want to look into how your website software is set
> up that might be causing this, or perhaps whether there's a proxy that
> is inappropriately adding its own chunked coding to the response as it
> passes through.
>
> Best Regards,
> Chris Thomas
>
> -- request --
> GET /2011/rich-hickey/ HTTP/1.1
> User-Agent: curl/7.18.1 (i686-pc-linux-gnu) libcurl/7.18.1
> OpenSSL/0.9.8b zlib/1.2.3
> Host: www.codequarterly.com
> Accept: */*
>
> -- response --
> HTTP/1.1 200 OK
> Date: Fri, 10 Jun 2011 00:51:12 GMT
> Transfer-Encoding: chunked
> Content-Type: text/plain
>
> 1f3e
> 1FE4<!DOCTYPE html>
> <HTML lang='en'>
>    <HEAD>
>        <META http-equiv='content-type' content='text/html; charset=UTF-8'>
>        <TITLE>Code Quarterly</TITLE>
> ....
>
>
>
>
> --
> Peter Seibel
> http://www.codequarterly.com/
>
> _______________________________________________
> tbnl-devel site list
> tbnl-devel at common-lisp.net
> http://common-lisp.net/mailman/listinfo/tbnl-devel