Re: svn commit: r729586 - in /httpd/httpd/trunk: CHANGES server/util_script.c

[Joe Orton <jo...@redhat.com>]

On Wed, Jan 07, 2009 at 02:34:29PM -0500, Eric Covener wrote:
> On Fri, Dec 26, 2008 at 10:53 PM,  <ni...@apache.org> wrote:
> > Author: niq
> > Date: Fri Dec 26 19:53:32 2008
> > New Revision: 729586
> >
> > URL: http://svn.apache.org/viewvc?rev=729586&view=rev
> > Log:
> > CGI: return 504 (Gateway timeout) rather than 500 when a script
> > times out before returning status line/headers.
> > PR 42190
> 
> Any concern that canned message for 504 is going to cause confusion?

I don't see why 504 is more appropriate than 500 for this case.

504 is specifically defined for cases where the server is acting as a 
gateway or proxy, which it is not here.  (by the 2616 definitions of 
gateway and proxy)

joe

Re: svn commit: r729586 - in /httpd/httpd/trunk: CHANGES server/util_script.c

[André Malo <nd...@perlig.de>]

* Nick Kew wrote: 


> On 8 Jan 2009, at 10:34, Joe Orton wrote:
> > I don't see why 504 is more appropriate than 500 for this case.
> >
> > 504 is specifically defined for cases where the server is acting as a
> > gateway or proxy, which it is not here.  (by the 2616 definitions of
> > gateway and proxy)
> >
> > joe
>
> One might consider the G of CGI a clue.
>
> The fact that the backend is (usually) an application running locally
> on the
> same machine as the webserver doesn't preclude the latter being a
> gateway.
>
> Come to think of it, CGI errors fall into more categories than we allow.
> A misconfiguration is indeed Internal Server Error.  But a script
> that generates
> garbage is an External Server Error, and a 502 response would be in
> order.
> It would be no bad thing to point the finger of blame at broken scripts
> rather than confuse the authors with "internal" errors.

Generally spoken, the message ist mostly not seen by authors, but by users. 
For *them* it's an opaque error (and should be), no matter what.

nd

Re: svn commit: r729586 - in /httpd/httpd/trunk: CHANGES server/util_script.c

[Nick Kew <ni...@webthing.com>]

On 8 Jan 2009, at 10:34, Joe Orton wrote:

>
> I don't see why 504 is more appropriate than 500 for this case.
>
> 504 is specifically defined for cases where the server is acting as a
> gateway or proxy, which it is not here.  (by the 2616 definitions of
> gateway and proxy)
>
> joe

One might consider the G of CGI a clue.

The fact that the backend is (usually) an application running locally  
on the
same machine as the webserver doesn't preclude the latter being a  
gateway.

Come to think of it, CGI errors fall into more categories than we allow.
A misconfiguration is indeed Internal Server Error.  But a script  
that generates
garbage is an External Server Error, and a 502 response would be in  
order.
It would be no bad thing to point the finger of blame at broken scripts
rather than confuse the authors with "internal" errors.

-- 
Nick Kew