[Zope] 'Cannot Locate' errors from Bobo using pcgi...

[Michel Pelletier]

> -----Original Message-----
> From: Tony McDonald [mailto:tony.mcdonald@ncl.ac.uk]
> Sent: Monday, June 14, 1999 5:40 AM
> To: Zope List
> Cc: Alex Rice
> Subject: Re: [Zope] 'Cannot Locate' errors from Bobo using pcgi...
> 
> 
> At 1:14 pm -0600 13/6/99, Alex Rice wrote:
> >On Sun, 13 Jun 1999 15:44:23 +0100,
> >"Tony McDonald" <tony.mcdonald@ncl.ac.uk> said:
> >
> >Tony> Bobo has encountered a problem publishing your object.
> >Tony> Cannot locate object at: http://server.net/wip/squishdo
> >
> >
> >Unfortunately, this is normal. If you search the mailing 
> list archives,
> >you'll see the folks from DC agreed that there really should 
> be hooks so
> >Publisher can use standard_error_message or something nice instead of
> >the Bobo traceback. Maybe that will show up in Zope2?
> >
> 
> cheers Alex, but Ouch!! ... that's nasty.
> DC guys: Is there a fix for Zope1.10 that I can apply whilst waiting 
> for Zope2? (it *is* fixed in Zope2 isn't it?)

Uhhmm... no.  This problem is a bit more complex than it sounds.  First
off, a 'Not Found' error (404) comes about because what you're looking
for is not in the object database, thus, the error is not raised *in*
the object database, but in ZPublisher, the ORB (unlike, for example, a
DTML error).

Due to the (very intentional) modularity of ZPublisher, ZPublisher can't
just assume that there is a standard_error_message object in the object
database (you can use ZPublisher by itself without Zope) so imagine it
raising a 404, looking for standard_error_message, which raises a 404
which...

Furthermore, ZPublisher is protocol independent.  If you wanted to
write, say, an NNTP protocol extension to ZPublisher, you could publish
your objects as newsgroups and items (sick!).  Or IMAP, or XML-RPC
(which Eric Kidd is working on as we speak) or CORBA, or whatever, pick
your protocol.  The problem is, standard_error_message doesn't make any
sense in those contexts, or maybe it does?

Jim has presented a proposal for fixing this, and a few other things
about our error model, for example, if you impliment error handling in
the ORB like his proposal, it will no longer be necesary for the object
database to grab its own errors, it can just let them float up to the
ORB and then use the new method to find and call standard_error_message.

This could be a 2.0 feature, if we can find the resources (I beg of you
to volunteer ;).

-Michel




> tone.
> ------
> Dr Tony McDonald,  FMCC, Networked Learning Environments Project
> The Medical School, Newcastle University Tel: +44 191 222 5888
> Fingerprint: 3450 876D FA41 B926 D3DD  F8C3 F2D0 C3B9 8B38 18A2
> 
> _______________________________________________
> Zope maillist  -  Zope@zope.org
> http://www.zope.org/mailman/listinfo/zope
> 
> (For developer-specific issues, use the companion list,
> zope-dev@zope.org - http://www.zope.org/mailman/listinfo/zope-dev )
>