[Zope3-dev] Re: unicode problems !?

[Bjorn Tillenius]

On Mon, Oct 11, 2004 at 11:00:09PM +0200, Philipp von Weitershausen wrote:
> Bjorn Tillenius wrote:
> >On Sat, Oct 09, 2004 at 11:42:53AM +0200, Philipp von Weitershausen wrote:
> >
> >>Files could grow an 'encoding' attribute that is updated by the editing 
> >>view according to the encoding of the request.
> >>I think text/* files should still be editable; we should fix up the 
> >>Bytes widget.
> >
> >
> >Yes, adding an 'encoding' attribute is one option. Although I don't
> >understand what you mean by that it should be updated by the editing
> >view according to to request.
> >
> >Often when users upload files, the files will be encoded in an encoding
> >different from UTF-8.
> 
> Why?

Are you asking why the uploaded file is not UTF-8? Have you ever heard
of a company called Microsoft? Can you show me a MS Windows system which
defaults to UTF-8?

At this point I'm not talking about creating an empty file, and enter
some text in a text box. I'm talking about creating a file, then upload
some local file. For example this e-mail, containing the letter 'ö'. Now
this e-mail isn't encoded using UTF-8...

> The user writes some chinese text into the text box. The user doesn't 
> care about an encoding at that point. The user just wants to enter text 
> in a textarea. The view should deduce the encoding from the request.

Why do you want to deduce the encoding? So if I edit the file using
Latin-1, it gets encoded as Latin-1. The you come and edit the file as
well, using UTF-8. How do you suggest that would work? Zope gets what's
in the text box as unicode in both cases already.

In other words, today the user doesn't have to care about the encoding
when he inputs text in a text box, that works already. (By deducing the
encoding from the request)

> If you upload a non "text/*" file, yes, then you should be required to 
> specify an encoding. But then you won't be shown a textarea anyway.

Now you lost me, why do you want an encoding for non text files?

> >The File must also use its 'encoding' attribute to set the encoding of
> >the response.
> 
> But of course :)

Yes, only wanted to point that out so that it wouldn't be forgotten if
someone chooses to implement this.

Regards,

Bjorn