RE: [Zope-dev] ANN: Proposal ContentNegotiation

2001-05-04 Thread Toby Dickenson



 -Original Message-
 From: Andreas Jung [mailto:[EMAIL PROTECTED]]
 Sent: 03 May 2001 20:23
 To: Toby Dickenson
 Cc: zope-dev
 Subject: Re: [Zope-dev] ANN: Proposal ContentNegotiation
 
 
 
 From: Toby Dickenson [EMAIL PROTECTED]
 To: 'Andreas Jung' [EMAIL PROTECTED]
 Sent: Thursday, May 03, 2001 1:04 PM
 Subject: RE: [Zope-dev] ANN: Proposal ContentNegotiation
 
 
  Your proposal suggests that published objects should 
 generally return
  pre-encoded objects in 8-bit strings. This is the one 
 detail which in my
  experience looks *wrong* and is very much dependant on my proposal.
 
 Not really - I only say that the internal character set is currently
 ascii/iso-8859-1.
 I did not mention that this will be the default character set 
 in the future.

Its not a question of the 'default' encoding if (as in your proposal) the
published object explicity specifies the character encoding of a pre-encoded
object.

Im referring to the section where you say:

: If the object creates a RESPONSE in another character set
: than the internal character set it must indicate this by
: setting an attribute of the RESPONSE object to the used
: character set e.g.:

:   RESPONSE.charset = 'utf-8'(either as attribute or by a method
call)

I think we are mixing up 'character set' and 'character encoding' here.
The character set of html (and xml) is the Unicode set.
utf-8 is a character encoding, not a character set.
Given the semantic of this proposed attribute, the name 'character_encoding'
would be more appropriate.

In my experience, forcing published methods to deal with character encoding
is a bad idea (I know because I tried that in my first iteration). Your
proposal suggests that methods return 8-bit strings, specifying the
character encoding of that string in the RESPONSE. This makes it hard to
combine methods from two different components that happen use different
character encodings.

Note that since many components already use latin-1, the usefulness of this
aspect proposal is low.

If they need a wider character set then using Unicode is the only practical
approach.

Ill write up a full proposal for how this would fit together on the wiki.

(On a different subject, I think the default encoding will always have to
stay as latin1 in order to support pre-existing object databases )







___
Zope-Dev maillist  -  [EMAIL PROTECTED]
http://lists.zope.org/mailman/listinfo/zope-dev
**  No cross posts or HTML encoding!  **
(Related lists - 
 http://lists.zope.org/mailman/listinfo/zope-announce
 http://lists.zope.org/mailman/listinfo/zope )



[Zope-dev] ANN: Proposal ContentNegotiation

2001-05-03 Thread Andreas Jung

Hi folks,

feel free to review and comment the  proposal for extending the content 
negotiation capabilities of Zope:

 http://dev.zope.org/Wikis/DevSite/Proposals/HTTPContentNegotiation

Cheers,
Andreas
Digital Creations


___
Zope-Dev maillist  -  [EMAIL PROTECTED]
http://lists.zope.org/mailman/listinfo/zope-dev
**  No cross posts or HTML encoding!  **
(Related lists - 
 http://lists.zope.org/mailman/listinfo/zope-announce
 http://lists.zope.org/mailman/listinfo/zope )



Re: [Zope-dev] ANN: Proposal ContentNegotiation

2001-05-03 Thread Toby Dickenson

On Thu, 3 May 2001 11:45:58 -0400, Andreas Jung
[EMAIL PROTECTED] wrote:

Hi folks,

feel free to review and comment the  proposal for extending the content 
negotiation capabilities of Zope:

 http://dev.zope.org/Wikis/DevSite/Proposals/HTTPContentNegotiation

Cheers,
Andreas
Digital Creations

Andreas, Are you aware that I have a mature patch to give ZPublisher
full Unicode support? It has been stable for six months, and is being
used in at least two commercial products.

http://www.zope.org/Members/htrd/wstring

Ive expressed before that I would be more than happy for these to be
rolled into the Zope core, and they do cover a large part of your CSN
problem (plus several related problems that you havent realised
yet. ;-)



Toby Dickenson
[EMAIL PROTECTED]

___
Zope-Dev maillist  -  [EMAIL PROTECTED]
http://lists.zope.org/mailman/listinfo/zope-dev
**  No cross posts or HTML encoding!  **
(Related lists -
 http://lists.zope.org/mailman/listinfo/zope-announce
 http://lists.zope.org/mailman/listinfo/zope )



Re: [Zope-dev] ANN: Proposal ContentNegotiation

2001-05-03 Thread Toby Dickenson

On Thu, 03 May 2001 17:14:28 +0100, Toby Dickenson
[EMAIL PROTECTED] wrote:

On Thu, 3 May 2001 11:45:58 -0400, Andreas Jung
[EMAIL PROTECTED] wrote:

 http://dev.zope.org/Wikis/DevSite/Proposals/HTTPContentNegotiation

http://www.zope.org/Members/htrd/wstring

Ive expressed before that I would be more than happy for these to be
rolled into the Zope core, and they do cover a large part of your CSN
problem (plus several related problems that you havent realised
yet. ;-)


I should also mention.

http://dev.zope.org/Wikis/DevSite/Proposals/UnicodeSupport

which is my original proposal to include optional Unicode Support
in the 2.3.0 release. Eventually only the first half of that proposal
was implemented, that is support for python 2.0 but not unicode. This
proposal is still the best overview of what my patch achieves.


Toby Dickenson
[EMAIL PROTECTED]

___
Zope-Dev maillist  -  [EMAIL PROTECTED]
http://lists.zope.org/mailman/listinfo/zope-dev
**  No cross posts or HTML encoding!  **
(Related lists -
 http://lists.zope.org/mailman/listinfo/zope-announce
 http://lists.zope.org/mailman/listinfo/zope )