Comment 4 for bug 322486

Paul Carduner, Stephan Richter and I talked about this. Paul and Stephan prefer the approach Paul describes in his comment. I (and the co-worker who encountered this issue) prefer the solution I initially proposed.

I think both sides have a reasonable argument, and we didn't come to a resolution. Community (esp. papal) input would be appreciated. While each side prefers different approaches, both sides agree that either approach is preferable to what we have now.

I'll try to summarize what I believe to be the most compelling points of the two arguments.

I feel that the common case is using standard status codes. Usually, specifying a non-standard status code *is* an error, and should be treated as such. We then want an extension mechanism for the exceptional cases.

Paul is concerned that my "addStatus" module function will be used either in module/package initialization, or require a new zcml (/grok) tag. That's perhaps a bit heavyweight.

Alternative solutions might be to have a new ``forceStatus`` method on the request that would work as Paul and Stephan prefer, and changing ``setStatus`` to act as I propose.

Thoughts?

Thanks

Gary