Hey there Bjorn -- Finally found the time to try out dump_obj, and it's showing me that everything *should* be stored correctly in Unicode. Furthermore, the properties page itself (sample attached) does correctly identify the input as UTF-8. However, the results of my simple dtml method still show question marks, even when I look at the source of the generated page (attached, along with the results of dump_obj). Given that the dtml-method result page *source* shows question marks in place of proper Japanese, it suggests to me that something isn't returning properly. Perhaps some intermediary function is not Unicode-aware and fills in the non-ascii chars with question marks? Thanks too for your Plone suggestion. I'm wading through the Plone and Archetype documentation, but it's a bit thin on the right kind of detail for a Plone/Zope newbie like me. I appreciate your time in helping me out. Cheers, Erik Anderson --- Bjorn Stabell <bjorn@stabell.name> wrote:
Hi Erik,
I'd recommend you get and install Plone+CMF+Zope from plone.org instead of trying to find all the correct products of the right versions yourself.
As for your problem displaying Japanese characters, it does sound like you're running into a unicode encoding/decoding problem.
If your properties are stored as unicode, Zope will automatically encode them in the output character encoding (in your case, this seems to be UTF-8).
Getting the input from your form decoded propertly takes a bit more work, though. Make sure all HTML text field names have :ustring:utf8 (or uline, utokens, or utext) appended to them, e.g.,
<input type="text" name="title:ustring:utf8">
Zope will look at the name of the field when guessing at which character encoding the text is in, and then decode it into a unicode string (which is stored using Python's internal encoding). This type of string will be of type "ustring".
To double-check that your strings are really stored as ustrings (they will look like u"hello", not "hello"), install and run the attached dump_obj external method on your object.
If this discussion continues, let's move this discussion to the Zope list (zope@zope.org) so that it gets archived and others can learn from it and help with it as well.
Bye, -- Bjorn
ATTACHMENT part 2 message/rfc822 From: "Admin-Unix2" <admin@exoweb.net> To: "Bjorn Stabell" <bjorn@exoweb.net> Date: Thu, 4 Dec 2003 14:12:59 +0800
# TODO: # - editing interface (open new window for each property) # - also list functions # - searching/ filtering # - collapse/ expand (dynamically load) # - include size information # - include cache information (ghosted or active) # # IDEAS: # - add debug to URL to have execution traced in a separate window, e.g., # .../debug/login_form
from DocumentTemplate.html_quote import html_quote
def dump_obj(self, depth=10): # # don't go too deep # if depth<0: return `self` out = ""
# # images; also show them # if hasattr(self, 'tag'): out += '<IMG SRC="'+self.getId()+'" style="float: right;"/>' #return self.tag(self) TODO: acquisition doesn't work? # files TODO: add link for downloading
# # list-like types # if type(self) in [type([]), type(())]: out += "\n" for item in self: out += "<DIV style='border-top: solid #cccccc 1px;'>"+dump_obj(item)+"</DIV>" return out
# # dictionary-like types (including objects) # d = {} if type(self) is type({}): # dictionaries d = self elif hasattr(self, '__dict__'): # objects d = self.__dict__ if d: out += "<TABLE>\n" for key,val in d.items(): out += "<TR>" out += "<TH align=left valign=top>" out += "<a title=\""+html_quote(str(type(val)))+"\">"+str(key)+"</a>" out += "</TH>" out += "<TD valign=top style='border: solid #669966 1px;'>" if type(d) is type({}): out += dump_obj(val, depth-1) else: out += dump_obj(getattr(d, key), depth-1) out += "</TD></TR>\n" out += "</TABLE>\n" return out
# # basic types # s = `self` if len(s) > 400: out += html_quote(s[:400])+'<b style="color: #cc3300;">...</b>' else: out += html_quote(s) return out