Dieter Maurer wrote:
The message seemed to be sent only to me.
whoops... Reply/Reply All user error bug ;-) cheers, Chris
Chris Withers writes:
Dieter Maurer wrote:
(Info: ([None, (['966520543', []],)], ['966520928', '966520893',
<-- state ->
'966520543'])) Somehow, your state became a tuple. It should be a list.
Is there anything in postingValues that could cause this? I'm trying to decide whether it's something my code is doing wrong or whether it's a bug in dtml-tree.. I think, I found a bug in "TreeTag". If your "branches" method produces an empty tree, then the state is generated as a tuple and not as a list. However, it does not look as in your example.
The code is near line 249: state= state or ([id],)
If you change this to a list, this may help (though chances are not too great).
After looking at the code, I fear, your many trees on one page will not work, unless they have the "single" argument.
What's the 'single' argument? I tells the tree to have at most one branch unfolded. If you klick at one branch, any other open branch is coalesced. This tree mode does not require a cookie.
... Tree Tag rewrite .... What do people think?
I think:
* rewriting the tree tag would be a good idea.
* I would complain, if the DTML features would be restricted.
I would like to have structured variable names, e.g. "x/y/z". It will, however, lead to more problems with URL references in text included in this way. These problems exist today, but people that understand the nested with are more likely able to understand the URL problems then those that use "x/y/z".
I like your proposal to have a calling feature direct in DTML without the need to go via "expr" (and therefore Python).
Dieter