[ZCM] [ZC] 695/ 7 Resolve "object id limitations are inconsistent at best"

[Collector: Zope Bugs, Features, and Patches ...]

Issue #695 Update (Resolve) "object id limitations are inconsistent at best"
 Status Resolved, Zope/bug low
To followup, visit:
  http://collector.zope.org/Zope/695

==============================================================
= Resolve - Entry #7 by ajung on Feb 2, 2003 7:19 am

 Status: Pending => Resolved

Patch applied to HEAD (will appear in Zope 2.7).

-aj
________________________________________
= Comment - Entry #6 by ajung on Feb 2, 2003 7:19 am

Patch applied to HEAD (will appear in Zope 2.7).

-aj
________________________________________
= Comment - Entry #5 by leper on Nov 24, 2002 6:10 pm


Uploaded:  "om.diff"
 - http://collector.zope.org/Zope/695/om.diff/view
Here's a patch to reject the . and .. cases which aren't traversable anyway, and follow PEP 8 more closely.
As for the other limitations, zope-dev it is, though as it stands I see no further reason to keep this bug open.
________________________________________
= Comment - Entry #4 by ajung on Nov 24, 2002 3:11 am

>> you can not check for every single forbidden id...

>Really? huh. Why doesn't the conventional wisdom of "define a >grammar describing the valid namespace and use it to reject >anything invalid" apply?

That's because no one has defined a grammer at the point when
Zope has been written. This is just a cruft. So there are
two solutions for you:

a) live with the current implementation

b) provide a suitable implementation

I suggest to continue the discussion on the zope-dev.
The collector is not a discussion forum.

-aj
________________________________________
= Comment - Entry #3 by leper on Nov 23, 2002 4:39 pm

> + and % are forbidden because they might be used in URLs
> for a replacement notation

Sorry, I don't mean to be thick but what do you mean by 'replacement notation'?  Is this a Zope thing, or are you just refering to the generic URI escaping encoding?

> you can not check for every single forbidden id...

Really? huh. Why doesn't the conventional wisdom of "define a grammar describing the valid namespace and use it to reject anything invalid" apply?
________________________________________
= Comment - Entry #2 by ajung on Nov 23, 2002 12:32 pm

+ and % are forbidden because they might be used in URLs
for a replacement notation and are reserved for this case
to avoid clashes. An id '.' or '..' does not make sense but
you can not check for every single forbidden id...I assume
you can construct much more valid Zope IDs that make less
sense.

-aj
________________________________________
= Request - Entry #1 by leper on Nov 23, 2002 12:19 pm

Creating an object with an id of . or .. is allowed while
creating objects with ids containing + or % is not.  Something
somewhere isn't quite right.  I'll probably open this up for
discussion on the zope-dev list shortly.

==============================================================