our import of JEmalloc, has been turned of by default and does
not improve performance when turned on.

seconds was a bit too aggresive for a default value, and after some
talk forth and back 0.7 seconds was deemed more like it.

things you should not muck about with unless I tell you to.

so I added one NUL too many to enum spec strings, preventing any
subsequent args from being handled correctly.

Spotted by:	Nils Goroll

the function-emitting function will do that.

Spotted by:	Nils Goroll

Make the default VCL explictly set the TTL to two minutes, decoupling
it from the default_ttl parameter.

This makes it clear that hit-for-pass happens, and makes it possible
to avoid the hit-for-pass object, (ie: get a plain pass) by setting
its ttl to zero in vcl_fetch.

position of the "last" bit in the included ESI object.

If an esi-processed job starts with <esi:include without any preceeding verbatim bytes, we would not emit the gzip header until after the included file.

with/without leading and training verbatim data.

preceeding verbatim bytes, we would not emit the gzip header
until after the included file.

Instead force the gzip header out as SKIP and inject a synthetic
one where needed during delivery.

Optional libedit support to varnishadm.  No saving of history or
completion yet.

workspace either.

runs into trouble when we reset the workspace as part of ESI:include
processing.

Fixes: #861

maintain proper ordering.

minimum demanded by the system.

this works better with Solaris

a smarter compiler than GCC could have seen through.

Fortunately there is a backstop, so worst case a request would
just fail.

Spotted by: Erik
Missed by: phk, twice