Some Spring Boot thoughts

[Cantor, Scott]

On 1/9/14, 4:47 PM, "Tom Zeller" <tzeller at dragonacea.biz> wrote:
>
>(a) Spring Boot supports externalized configuration with
>self-referential properties, e.g. :
> idp.properties:
>  idp.home=/opt/idp
>  idp.keystore.path=${idp.home}/creds/idp.jks

That might be a strong argument, since it solves the problem of embedding
assumptions about resource types in the Spring config.

>For the v3 AACLI, I thought we might skip a "traditional" application
>layout :
>
> bin/aacli.sh
> bin/aacli.bat
> lib/idp-attribute-cli.jar
> lib/...dependencies..
>
>and replace it with
>
> java -jar <something>.jar

Doesn't create any problems to me, though usually it seems like even that
kind of model requires batch scripts to get JAVA_HOME or other variables
set.

>I guess I like the
>idea of not maintaing shell or batch files in lieu of Java.

I'm not optimistic we'd have none, but they would be boilerplate.

>Just some thoughts. In the last few days I looked for a way to support
>self-referential properties and loading dependencies from a WAR, and
>it seems like Spring Boot might help, hence this message.

I think we're at a point we have to skip experimentation and just evaluate
complete solutions. As a comparative example, I definitely could have
worked with different serialization libraries, but I knew that I could
build something a certain way that would work, so I just did it. So I
think we have to finish formulating requirements, where any are missing,
and then assign the job to somebody to just "do it" in whatever way they
think is best.

-- Scott