Summary of error handling

[Cantor, Scott]

On 6/10/14, 3:01 AM, "Rod Widdowson" <rdw at steadingsoftware.com> wrote:
>
>Slightly orthogonal question - which I think I know the answer for:  How
>much do we care about the jsps in these new situations?  That is, they may
>work but how much effort do we put in beyond that?

For something like this, I'll at least replicate what I did in Velocity.

>However, and I think this is the answer to my question,  jsp by its nature
>is far more powerful than velocity, hence we have to actively support this
>for the 1% who need this power (out of the 10% who will make any changes
>to
>the out of the box config).

Yes, I think so. I also have been playing with Jetty 9.2, and so far I see
no way to run an application with a packed war. I'm still working on it,
but the bugs around this have very disturbing comments in them that make
it sound like Jetty has just given up on the idea.

If that's true, then you're in a situation where you might as well use JSP
because you get all the power and the tree is unpacked anyway, so you can
change them in real time.

>So to take the case of I18N - if someone needs to do this properly (and
>I'm
>guessing Switch are at the forefront of that) then they are probably going
>to need the "get out of jail card" that is jsp, and live with the pain.

Yeah, but as I said, I don't think we should actually fix Spring to do
this right. So you'd have to invent your own i18n layer I guess. My view
is it's a Spring bug and we should just wait for them to fix the API to
take Locale[] instead of Locale.

>Therefore we need to ensure that we have enough example code in our jsps
>that the 0.1% can make progress.
>
>Make sense?

To a point, but when we're talking about the .1%, I'm not a big believer
in spoon feeding examples. That's what Javadocs are for.

-- Scott