part of the transport protocol (don't even get me started!) we need to
lift it out of HTTP/1 space and into the RFC processing.

Submitted by:       nublaii via github

Submitted by:	Clement Gautier via github

Adhere to Fedora packaging standard.

bo's workspace and eliminate the hardcoded limit of 5 VFPs.

API to a separate #include file.

busyobj if we want to examine it in any detail.

Hold that ref over all of V1D_Deliver() (different from Nils
patch) to make sure we have a consistent view on streaming
vs. no streaming throughout.

Also use the bo to clean up a hack related to ESI includes.

Diagnosed by:	Nils Goroll

Fixes #1532

Reinstall the readline handler when we are called in order to clear the
input line. Tested in a variety of versions by scn and myself.

Fixes #1531

Rename suitably and always put the objcore argument first.

beresp.cacheable is gone.

Essentially identical to patch by Martin

move n_object accounting responsibility to stevedores.

Fix n_object accounting detail in persistent.

in preparation for moving exp from obj to objcore

clearly isn't going to be a much used public interface.

we know what we're doing.

Reserving and returning space from wrk->aws is OK.

advantage from having two mechanisms.

or an string-insertion-sort feels right for me.

Using reserved workspace to tracking the params and qsort(3) to sort them
solves both problems.