Re: Catching errors

/ Alessandro Vernet <avernet@orbeon.com> was heard to say:
| On 7/10/07, Norman Walsh <ndw@nwalsh.com> wrote:
|> I can't think of one, but I'm not sure I've ever used "finally" in a
|> Java program either. Given that steps are self-contained, I really
|> don't think we need a finally. Though, I suppose someone will raise
|> the case of an extension step that opens a database and needs a
|> finally to close it.
|
| I am convinced by Rui's arguments for introducing the ability to catch
| specific errors and to provide a "finally" block. My arguments are as
| follows:

I've come around to think "not for V1".

If we have just a p:catch in V1 then we can extend it in the future if
we really see that error codes returned by steps are useful and used.

I'm afraid that we don't really know if steps are going to return
meaningful error codes often enough to make the multiple-catches
approach usable.

If we add it now, we're stuck with it forever.

In short: good idea, probably the right thing, not in V1.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | Upon the whole I dislike mankind:
http://nwalsh.com/            | whatever people on the other side of
                              | the question may advance, they cannot
                              | deny that they are always surprised at
                              | hearing of a good action and never of a
                              | bad one.-- Keats