[Zope-dev] [Zope-Checkins] SVN: Zope/trunk/ merged haufe-legacy-integration branch

[Chris Withers]

Hi All,

I have some real issues with this work being done in the fashion it has:

- It's being done into the beta phase of Zope 2.12

- There's been zero discussion with the community before this was done.

Have I missed something here?

In particular:
> Modified: Zope/trunk/src/ZPublisher/Publish.py
> +from zope.event import notify
>  
> +from pubevents import PubStart, PubSuccess, PubFailure, \
> +     PubBeforeCommit, PubAfterTraversal
> +

Has anyone looked at the performance implications of all these events 
(which most of us will never use) being fired off?

>      except:
> +        # save in order to give 'PubFailure' the original exception info
> +        exc_info = sys.exc_info()

As this is never deleted, does it not leak stack frames?

> Modified: Zope/trunk/src/Zope2/Startup/handlers.py
> ===================================================================
> --- Zope/trunk/src/Zope2/Startup/handlers.py	2009-05-12 08:32:23 UTC (rev 99865)
> +++ Zope/trunk/src/Zope2/Startup/handlers.py	2009-05-12 08:49:29 UTC (rev 99866)
> @@ -184,8 +184,10 @@
>      """
>  
>      # Set environment variables
> -    for k,v in config.environment.items():
> -        os.environ[k] = v
> +    d = {}
> +    for s in config.environment: 
> +        d.update(s)
> +    os.environ.update(d)

The change in the config that makes this necessary is pretty 
unacceptable. I have lots of code that looks in config.environment, now 
I have to find all that code and change it in a similar fashion. WTF?

> -  <section type="environment" attribute="environment" name="*">
> +  <multisection type="environment" attribute="environment" name="*">
>      <description>
>       A section which allows a user to define arbitrary key-value pairs for
>       use as environment variables during Zope's run cycle.  It
>       is not recommended to set system-related environment variables such as
>       PYTHONPATH within this section.
>      </description>
> -  </section>
> +  </multisection>

...this is the offending change.

Andreas, please revert this change until people have had a chance to 
look at it properly.

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk