httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dean Gaudet <dgau...@arctic.org>
Subject Re: [PATCH] for 1.3: Last-Modified from scripts
Date Thu, 14 Aug 1997 21:46:42 GMT
Hmm.  I disagree with this.  If it was the CGI's responsibility to
properly implement HTTP/1.0 then the same logic could argue that it's a
module's responsibility to implement HTTP/1.0.  i.e. set_last_modified() 
shouldn't exist because modules should be smart enough to do it themself.

Also look at PR#895 where I complain that etag support should not be part
of set_last_modified.  I haven't checked if Ken's patch would cause an
ETag to be generated, if it does then it's not correct ... and both PR#894
and PR#895 need to be solved together. 

CGIs certainly don't implement HTTP/1.1.  So it's really our
responsibility (should we want it) to deal with their Last-Modified and
ETag headers.  But this is admittedly a very low priority problem.  I just
noted it as a deficiency in our API. 

Dean

On Thu, 14 Aug 1997, Alexei Kosut wrote:

> On Thu, 14 Aug 1997, Rodent of Unusual Size wrote:
> 
> > From the fingers of Alexei Kosut flowed the following:
> > >
> > >I have a vauge memory of, a year or two ago, Rob Hartill subitting a
> > >patch to do this against 0.8 (or maybe 0.7). It was decided to be a bad
> > >idea. I forget why, though.
> 
> I've found the relevant messages, from June 7, 1995 (with trimmed
> headers). I've included two messages here. After them, I don't see any
> more on this thread, although the discussions on that same day did
> involve porting Apache to NT (only took us two years, not bad):
> 
> From: Rob Hartill <hartill@ooo.lanl.gov>
> Subject: Re: 0.7.2b
> To: new-httpd@hyperreal.org
> Date: Wed, 7 Jun 95 16:29:36 MDT
> 
> > Given that that works as advertised, I still think the error code
> > response
> > should be entirely up to the CGI script w/r/t
> > last-modified/if-modified-since.
> 
> If the script tells apache, "this hasn't been modified since X", 
> apache can easily compare X with an if-modified-since and
> stop the CGI process immediately. Under what conditions would
> this be a bad idea ?
> 
> This is similar to HEAD requests, Apache will often kill the script
> before it can send content.
> 
> 
> --
> Rob Hartill                           
> http://nqcd.lanl.gov/~hartill/
> 
> Date: Wed, 7 Jun 1995 16:59:13 -0700 (PDT)
> From: Brian Behlendorf <brian@organic.com>
> Subject: Re: 0.7.2b
> To: new-httpd@hyperreal.org
> 
> On Wed, 7 Jun 1995, Rob Hartill wrote:
> > > Given that that works as advertised, I still think the error code
> > > response
> > > should be entirely up to the CGI script w/r/t
> > > last-modified/if-modified-since.
> > 
> > If the script tells apache, "this hasn't been modified since X", 
> > apache can easily compare X with an if-modified-since and
> > stop the CGI process immediately. Under what conditions would
> > this be a bad idea ?
> 
> Killing scripts before they're done is in general bad, but I suppose that
> happens anyways.  
> 
> Beyond that, a script's ability to find out if any of the modules it uses
> to render a document are older than date X is usually much quicker 
> than the rendering itself.  I suppose your proposal doesn't give these 
> kinds of scripts any grief, so whatever... it just seems odd that a CGI 
> script smart enough to return Last-Modified couldn't be smart enough to 
> check for HTTP_IF_MODIFIED_SINCE.  
> 
> Whatever happened to the CGI spec?  This could/should be a part of that.
> 
>         Brian
> 
> --=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
> brian@organic.com  brian@hyperreal.org
> http://www.[hyperreal,organic].com/
> 
> 
> -- Alexei Kosut <akosut@organic.com>
> 
> 


Mime
View raw message