Great!
They are reasonable, but this needs to go into 1.0 versus 0.9, so next release.
(There is no numerical significance to 1.0)
One quick follo wup.
Also, I may suggest waiting until we have cleaned up our template
logic. We need to do the "multiple basedir" thing first, as this is
already somewhat a source of confusion where a relatively pathed file
lives with includes at multiple directly levels.
When do you think that the template logic will be cleaned up and the multiple basedir will be done? Is this something that I could also help with?
No idea. Template code must be refactored first, and this can't
start until we release 1.0
It is unlikely that this is something I'm going to work on, help is
welcome, but this area of code is both extremely important and
complicated, so just changing it to do multiple basedirs is not the
solution we are looking for.
It's going to be closer to a mini-rewrite.
FYI should you want to try this
It at least affects the signature of every template and template_ds
call in the app, and probably indicates that various things should
keep up a self.basedirs and just pass that context into the template
class, rather than passing a basedir around.
In some cases basedirs will be the union of the play and the task
basedirs, etc. It would also need to consider handler basedirs (if
imported), the cwd, etc.
Michael,
I saw that ansible 1.0 now supports arbitrary depth playbook inclusion so having me augment the code will no longer be necessary. I did a github fork of the project for version 0.8 that added rudimentary recursion. I see about removing that fork.
Thanks.