#130 Needs Unicode Support

[Jayesh Sheth]

The following was posted by a visitor to the Rilke CMS
demo site:

"
Lacking unicode support
THis test shows that Rilke CMS can't work with
multi-script texts. Here is a russian letter "я"
and here is a swedish "ה". The swedish is perhaps
readable while the russiand זרו
абвг appears as a numerical
character reference. However, if you try to edit it,
inside HTMLarea again, the russian letter pops up
again, as it is supposed to do.

It is problably something with the publishing process
... because I think HTMLarea are good enough?

Btw, Rilke CMS lacks a preview of postings, doesn't it?
"

We need to test Rilke CMS with non-English languages. I
think this is a server side processing issue. It
probably has something to do with the html tag
indicating the encoding, which in turn is interpreted
by the server.

This issue needs more investigation ...

[Jayesh Sheth]

Logged In: YES
user_id=809642

I got the following note from a user about this problem:

Submitted By: Leif Halvard Silli (hyperlekken [at] lenk.no)
As proved by the use of Midas inside Radio Userland, one can
publish multilingual text, e.g. russian and norwegian, with
Midas in Radio Blog. My trial account of Radio demonstrates
that.

http://radio.weblogs.com/0136031/

But when I tested Rilke CMS, which uses Midas in the form of
HTMLarea (right, HTMLarea is a Midas thing?), then I landed
in problems. Just look for yourself at this link:

http://www.rilkecms.com/demo/?id=63

The test on that link shows that Rilke CMS can't work with
multi-script texts. Here is a russian letter "я" and here is
a swedish "". The swedish is perhaps readable while the
russian appears as a numerical character reference. However,
if you try to edit it, inside HTMLarea again, the russian
letter pops up again, as it is supposed to do.

The same problem does not appear to be built into HTMLarea
itself, because if you try out HTML area at Mishoos website

http://dynarch.com/htmlarea/examples/core.html

then there are no problems, except that you must remember to
change the encoding manually to UTF-8.

I have some ideas about where to look for the reason of the
problem. You see Radio uses ISO-8859-1 pages by default.
Therefore all non-ASCII text is inserted as html/numerical
character entities. THis works 100% ok. But Mishoo's web
site use UTF-8. Therefore the text one types in are also
published using UTF-8 character representation instead of
entities. But in Rilke CMS you uses ISO-8859-15 on the
editing page while you uses ISO-8859-1 on the published
pages. I suppose that if you just makes the encoding of both
editing and publishing the same, then this will work fine.
At least for Mozilla browser. So you might i.e. set it to
ISO-8859-1 and we will all be happy.

But you might also try UTF-8. I suppose that will make Rilke
CMS more multilingualy "safe" for other browsers than the
Mozilla ones.

[Jayesh Sheth]

Logged In: YES
user_id=809642

Most probably solution (as per previous comment):

Ensure that the encoding for the Administration Screen and
the main page is the same using a tag such as the following:

<meta http-equiv="content-type" content="text/html;
charset=iso-8859-1">