Closes #2870

Addresses the doc part of #2872.

Solaris, (remember Solaris?) will return connection related errnos
on fd metadata operations such as ioctl(2), [gs]etsockopt(2) etc.

We propagated these errors up from VTC_(non)blocking() which could,
theoretically, bail out earlier than it would otherwise have done.
This complicated code far more than it speeded anything up.

Instead just ensure that any errors returned are indeed connection
related, and then ignore them, and let Solaris deal with the broken
connection same way as other operating systems.

This is more complex than it sounds because the active VCL at the
time the topreq got scheduled, may no longer be by the time
the esi:include req gets processed.

Fixes: #2849

again and all the overhead it brought.

Fixes #2846

Fixes #2880

Fixes #2879

but only "errno.h"

When a VMOD returns a STRANDS, it is a requirement that the storage
be stable, but that is not really different from any other type of
return value.

Completes #2873

This has bugged me (slightly) for ages.

Too trivial to warrant its own line(s) in test-case/u00008.vtc.

Strenghen the case against STRING_LIST in the docs.

See also: #2873

Should address "HTTP cache must not reuse a response with an invalid
Cache-Control: max-age (trailing alpha)" from https://cache-tests.fyi.

Should address "HTTP cache must not use a cached response with
Cache-Control: No-CaChE, even with max-age and Expires" from
https://cache-tests.fyi.