• Remember Mailing List

  • disabled fields in personalize_form for zope root admin

    from jhb on May 22, 2010 04:14 AM 
    Hi *,

I basically have the same behaviour as in

http://www.coactivate.org/projects/remember/lists/remember/archive/2007/12/1197298124524

which is that my admin user from the zope root can't change settings in
personilze form. This happens as soon as remember is installed on a
fresh site. I have read Robs answers to the email above, but don't see
how to make use of his pointers.

I am using a buildout which extends

http://dist.plone.org/release/3.3.5/versions.cfg

Any ideas or pointers?

Cheers,

  Joerg
    Thread Outline:

  • Re: disabled fields in personalize_form for zope root admin

    from jhb on May 22, 2010 05:55 AM 
    > http://www.coactivate.org/projects/remember/lists/remember/archive/2007/12/1197298124524

I poked a bit more around the issue:
http://baach.de/Members/jhb/research-notes/plone-can-edit-admin-user-in-personalize_form-if-remember-is-installed

Turns out that PAS checks:

mdata = getToolByName(self, 'portal_memberdata', None)
    if mdata:

Due to remembers MemberDataContainer being a BTree, and having a __len__
method, but the tree being empty, this evaluates to false.

Now, I can fix it by giving the MemberDataContainer a __nonzero__ (and
file a bug+patch to remember), or change the PAS so that it does 'if
mdata != None' (and file a bug with PAS).

Which way to go?

Cheers,

  Joerg

    • Re: disabled fields in personalize_form for zope root admin

      from ra on May 22, 2010 02:14 PM 
      On 05/22/2010 02:54 AM, Joerg Baach wrote:
>> http://www.coactivate.org/projects/remember/lists/remember/archive/2007/12/1197298124524
>
> I poked a bit more around the issue:
> http://baach.de/Members/jhb/research-notes/plone-can-edit-admin-user-in-personalize_form-if-remember-is-installed
>
> Turns out that PAS checks:
>
> mdata = getToolByName(self, 'portal_memberdata', None)
>      if mdata:
>
> Due to remembers MemberDataContainer being a BTree, and having a __len__
> method, but the tree being empty, this evaluates to false.
>
> Now, I can fix it by giving the MemberDataContainer a __nonzero__ (and
> file a bug+patch to remember), or change the PAS so that it does 'if
> mdata != None' (and file a bug with PAS).
>
> Which way to go?

Personally, I consider this a PAS bug.  I think it's A Bad Idea® to do an "if 
foo:" check when what you really mean is "if foo is None:", since with duck 
typing you never know what circumstances might cause an object to evaluate to 
False, as you've discovered here.

-r
Return to conversation view