need the busyobj at all now.

Dispose when refcount goes to zero.
Private objcores for pipe neither want nor need a boc, so make it optional.

I don't think this will be necessary at the end state, but it makes
the transformation easier.

now that we have a struct boc.

(struct boc) and "the backend fetch context" (struct bereq).

WS_Assert allows the workspace to be full in its check that `ws->f <= ws->e`.
WS_Reset does not honor this; if a snapshot was taken of a full workspace, and
that snapshot is reverted with WS_Reset, we panic.

This issue was identified by Jozef Hatala <jozef@fastly.com>

This (useless) range request is send by Chrome with HTML5 Video.
This was not working correctly in 4.0 and was fixed somewhere
in master.

Discussed with phk@.  Fix confirmed by geoff@.

Fixes #1844.

Fixes previous commit. Sigh.

Broken in 9bb8f962.
Committed solution proposed by phk@, discussed with phk@ and martin@.

Fixes #1845.

Discussed with mithrandir@.

Prompted by #1845.  I'm not entirely sure whether some literals like
"1.00 " should be accepted but that's the current behaviour.

into the macro-table.

Instead of counting the references, the VCL keeps track of them with up
to 31 characters in the description. It is the VMOD's responsibility to
keep track of the opaque struct vclref * and provide a meaningful user-
friendly description.

By the time we decide to switch to a VCL, it must be warm and usable.
The deprecated VCL_EVENT_USE should not get in the way.