Bug #76331 [->Opn]: Location header overrides Content-Type
| From: | requinix@php.net | Date: | Sat, 09 Jun 2018 06:54:39 +0000 |
| Subject: | Bug #76331 [->Opn]: Location header overrides Content-Type | ||
| References: | 1 | Groups: | php.bugs |
| Request: | Send a blank email to php-bugs+get-215555@lists.php.net to get a copy of this message | ||
Edit report at https://bugs.php.net/bug.php?id=76331&edit=1
ID: 76331
Updated by: requinix@php.net
Reported by: Andy_Schmidt at HM-Software dot com
Summary: Location header overrides Content-Type
-Status: Openp
+Status: Open
Type: Bug
Package: HTTP related
Operating System: Win 2012
PHP Version: 7.2.5
Block user comment: N
Private report: N
Previous Comments:
------------------------------------------------------------------------
[2018-06-08 21:28:59] Andy_Schmidt at HM-Software dot com
There are additional complications due to this behavior, in case it helps tracking it down. I just
lost a few days tracking down impossible problems in my asynchronous application (claiming that
there was additional output after the headers had been written and the buffers had been flush()ed
and ob_flushed() )
At the end those turned down to be another side effect of this behavior. Because, NOT ONLY does the
system override the "Content-Type", it actually injects a default HTML content with the
title "Document Moved", and an H1 of "Object Moved" and a body of "This
document may be found here". To make things even worse, it leaves the original
"Content-Length" of the image file in place.
If THAT default HTML redirect content is part of the PHP code, then this would suggest that this
behavior is actually native to PHP!
From here things go quickly downhill. Either due to the excessive (= wrong) content length, or due
to the previous content type, or maybe PHP still has the original jpeg data stream in a buffer, the
FCGI handler will then be handed an extraneous data stream starting with hex FF D8 FF E0 ... which
is the JPEG filetype signature. The extra data (after the request had supposedly already completed)
causes the web server/FCGI to log a system error and terminate the instance.
------------------------------------------------------------------------
[2018-05-11 21:20:47] cmb@php.net
> [â¦] one yet need to show me any client not following a http
> redirect unconditional
<https://curl.haxx.se/> does not even follow redirects by
default.
Anyway, this bug tracker is most certainly not the appropriate
place to discuss the reasonableness of Internet standards and
possibly divergent behavior of clients. Please let's stick to the
issue at hand, which is that the supplied Content-Type header is
overridden for apparently no good reason.
------------------------------------------------------------------------
[2018-05-11 20:57:46] spam2 at rhsoft dot net
WTF - especially after a POST request succeeded automatic redirects to a confirmation page are
common to avoid multiple submits and one yet need to show me any client not following a http
redirect unconditional
------------------------------------------------------------------------
[2018-05-11 20:55:28] cmb@php.net
> SAPI is standard IIS 8.5 FastCGI.
Thanks. So this is likely an (F)CGI issue (I've tested with
Apache mod_php).
------------------------------------------------------------------------
[2018-05-11 20:42:03] Andy_Schmidt at HM-Software dot com
>> spam2@rhsoft.net: ...should not contain any http body at all <<
Automatic redirection is NOT a requirement, moreover, inclusion of a payload IS explicitly mentioned
as being common - exactly the opposite of your opinion. Please note https://tools.ietf.org/html/rfc7231#section-6.4.2:
"The server SHOULD generate a Location header field in the response
containing a preferred URI reference for the new permanent URI. The
user agent MAY use the Location field value for automatic
redirection. The server's response payload usually contains a short
hypertext note with a hyperlink to the new URI(s)."
------------------------------------------------------------------------
The remainder of the comments for this report are too long. To view
the rest of the comments, please view the bug report online at
https://bugs.php.net/bug.php?id=76331
--
Edit this bug report at https://bugs.php.net/bug.php?id=76331&edit=1