[MRG+1] dissolve contrib_exp

[nyov]

Here's a start for #1063, getting rid of contrib_exp.
I haven't deleted the old xml_iter yet, I think it should be done in the next commit so we have at least a point in history where bot are available.
The decompression middleware I moved to contrib, django item was only a deprecation warning anyway.

edit: I don't think we need backwards-compat. for this one, it said "experimental" for a reason.

[kmike]

1. What about keeping both xmliter_lxml and xmliter for now? Before removing xmliter we should check that xmliter_lxml works better.
2. Even though contrib_exp was experimental, people used it, so +1 from me to add a backwards compatibility shim.

[nyov]

Not my decision to remove xmliter, I wouldn't mind to keep both, @curita mentioned that in #1063 (comment).

On the compat shims, would you agree that we don't need one for djangoitem? It was already a shim (since 70f8e51).

[kmike]

On the compat shims, would you agree that we don't need one for djangoitem? It was already a shim (since 70f8e51).

Yes, +1 to dropping the existing djangoitem shim.

[curita]

I don't recall why we decided on removing xmliter really (/cc @nramirezuy, @pablohoffman). I'm OK with keeping it too.

[kmike]

Thanks @nyov, the PR looks good to me.
+1 to merge it after #963.

[pablohoffman]

needs rebase after merging #963

[nyov]

done