Date in HTTP headers

2001-07-26 Thread GOMEZ Henri

Hi,

TC 3.2.3 and 3.3b1 didn't send the actually the Date header
in http 1.0 connector.

RFC2616 indicate that the Date header should be present.

What's your opinion adding it to both implementation ?
FYI, Tomcat 4.0 send the Date header :)



Re: Date in HTTP headers

2001-07-26 Thread Mike Anderson

+1

Mike Anderson

>>> [EMAIL PROTECTED] 07/26/01 12:03PM >>>
Hi,

TC 3.2.3 and 3.3b1 didn't send the actually the Date header
in http 1.0 connector.

RFC2616 indicate that the Date header should be present.

What's your opinion adding it to both implementation ?
FYI, Tomcat 4.0 send the Date header :)




RE: Date in HTTP headers

2001-07-26 Thread Ignacio J. Ortega

Hola Henri:

First at all i'm +1 on adding Date header aswell ( is on my todo list
for 3 months now ;)

But just a comment: RFC2616 is the "Hypertext Transfer Protocol --
HTTP/1.1" RFC, AFAIK HTTP 1.0 doesnt have a real RFC, well it has one,
the RFC1945 but this as the RFC itself recognizes is not a real RFC (
one of obligated honoring ) it is more a set of comments and indications
and advices for compatibility than a complete RFC..  notably in RFC1945
there is no support for Modified-Since header we are sending now i.e.
and a limited Host: Header support

And remember we only claim to have HTTP 1.0 support, if you continue
looking at RFC2616, you will see a big number of other characteristics
of HTTP 1.1 protocol we are not implementing, notably the ones that
makes HTTP 1.1 such a good and low overhead protocol ( connection
persistance, chunking i.e )..

Said that, we are not obligated to implement the complete rfc2616, but
we can implement those features in RFC2616 that mades users have a
better experience with tc33 standalone use..

Too. when using TC33 behind a supported webserver, that makes TC33 use
all that features present on HTTP1.1 without any problems...

Remmy could give us ( another time more :) some light on that .. 

only a light remark..

Saludos ,
Ignacio J. Ortega


> -Mensaje original-
> De: GOMEZ Henri [mailto:[EMAIL PROTECTED]]
> Enviado el: jueves 26 de julio de 2001 20:03
> Para: [EMAIL PROTECTED]
> Asunto: Date in HTTP headers
> 
> 
> Hi,
> 
> TC 3.2.3 and 3.3b1 didn't send the actually the Date header
> in http 1.0 connector.
> 
> RFC2616 indicate that the Date header should be present.
> 
> What's your opinion adding it to both implementation ?
> FYI, Tomcat 4.0 send the Date header :)
> 



RE: Date in HTTP headers

2001-07-26 Thread GOMEZ Henri

>First at all i'm +1 on adding Date header aswell ( is on my todo list
>for 3 months now ;)

Done, I attached 2 diff one for TC 3.3 CVS and one against TC 3.2.3.
Note the little hack to speedup date generation in rfc1123 format :)

PS: TC 3.2.3 expose Servlet-Engine but not TC 3.3. What about adding
this one also ? We won't loose to many time, especially for those of
us using Ajp12/Ajp13 :)

>But just a comment: RFC2616 is the "Hypertext Transfer Protocol --
>HTTP/1.1" RFC, AFAIK HTTP 1.0 doesnt have a real RFC, well it has one,
>the RFC1945 but this as the RFC itself recognizes is not a real RFC (
>one of obligated honoring ) it is more a set of comments and 
>indications
>and advices for compatibility than a complete RFC..  notably in RFC1945
>there is no support for Modified-Since header we are sending now i.e.
>and a limited Host: Header support

HTTP 1.1 support for TC 3.2.3 / 3.3 are in my project list but since TC
3.2.3
is feature freeze, it will be only for TC 3.3. But after ajp14 completion...

PS: I implemented the same code that in TC 3.2.x, ie Date header is set 
only if the user didn't allready set the Date header. Did it meet spec
or should we replace the user Date header by the Http connector computed one
?


 date-fix-3.2.diff
 date-fix-3.3.diff


RE: Date in HTTP headers

2001-07-26 Thread Ignacio J. Ortega

+1 for this two patches..

> -Mensaje original-
> De: GOMEZ Henri [mailto:[EMAIL PROTECTED]]
> Enviado el: jueves 26 de julio de 2001 22:17
> Para: [EMAIL PROTECTED]
> Asunto: RE: Date in HTTP headers
> 
> 
> >First at all i'm +1 on adding Date header aswell ( is on my todo list
> >for 3 months now ;)
> 
> Done, I attached 2 diff one for TC 3.3 CVS and one against TC 3.2.3.
> Note the little hack to speedup date generation in rfc1123 format :)
> 
> PS: TC 3.2.3 expose Servlet-Engine but not TC 3.3. What about adding
> this one also ? We won't loose to many time, especially for those of
> us using Ajp12/Ajp13 :)
> 

I dont understand my friend, what do you want to say in that
PostScriptum ?

> >But just a comment: RFC2616 is the "Hypertext Transfer Protocol --
> >HTTP/1.1" RFC, AFAIK HTTP 1.0 doesnt have a real RFC, well 
> it has one,
> >the RFC1945 but this as the RFC itself recognizes is not a real RFC (
> >one of obligated honoring ) it is more a set of comments and 
> >indications
> >and advices for compatibility than a complete RFC..  notably 
> in RFC1945
> >there is no support for Modified-Since header we are sending now i.e.
> >and a limited Host: Header support
> 
> HTTP 1.1 support for TC 3.2.3 / 3.3 are in my project list 
> but since TC
> 3.2.3
> is feature freeze, it will be only for TC 3.3. But after 
> ajp14 completion...
> 

Mine too since 3 or4 months ago, i have started to do it a couple of
times, nothing so far , :( 

> PS: I implemented the same code that in TC 3.2.x, ie Date 
> header is set 
> only if the user didn't allready set the Date header. Did it meet spec
> or should we replace the user Date header by the Http 
> connector computed one
> ?
> 
> 

For me it's ok, although not digged in the spec to say that ;)


Saludos ,
Ignacio J. Ortega