I am not entirely happy about this and we should try to find a better solution
to this chicken-egg-ish situation:

- part of our documentation is hand-written (in git/srcdir), part is auto-
  generated

- because we add generated rsts to the dist, they are not in srcdir when
  building from git, but they are when building from a dist (tarball)

- sphinx wants all files in one doctree - there seems to be no include path

I tend to lean towards removing generated rsts from the dist to unify the
src tree for git vs. dist builds, but I'd like to avoid too much change at
this point due to other priorities.

... which revealed that they actually fail atm, so another fix will be coming up

This allows unattended updates of vtest.sh itself by looping externally

Fixes #2021

We already had a check that took care of this in an architecture
independent way.

them a clear type.

Fixes #2016

Closes #1750

Closes #2155

Fixes #2134.

HTC_RxInit could write a single '\0' NUL character outside of the
workspace when its called and there is zero bytes left in the
workspace. This would trigger the workspace canary causing subsequent
assertion.

Fix by releaving HTC_RxInit of adding the '\0' character.

HTC_RxStuff now returns HTC_S_Overflow early if the available buffer
space is zero. The '\0' character is inserted just before calling the
completion check function.

Also fix an off-by-one error on the http_{req|resp}_size calculations,
where the maximum number of bytes accepted was one less than the
paramter indicated. c00039.vtc and c00040.vtc has been edited to
reflect that and to be more expressive about the sizes they generate.

Fixes: #1834

Make sure that the workspace can accomodate the pipelined data before
memmoving it into place.

Add a comment on an open issue in the H2 code path that could trigger
this assert.

When doing range delivery, make the delivery processing finish early
when all of the bytes of the requested range has been delivered. This
to avoid waiting around for a slow fetch to finish before handling the
next request on the connection.

Fixes: #2035

A delivery processor may want to finish the delivery early. Change the
return value from VDP_bytes and the delivery processors into 3
categories. 0 means continue, negative is error and positive is
finished.

The latching of errors (to help ESI delivery continue while still
recording the error) will latch the lowest non-zero value observed.

VDP_gunzip is fixed to not unconditionally return errors.

Also add some text to explain the behaviour of VDP_bytes and its
delivery processors.

Merges #2152

In particular VSL_Error() and VSM_Error() return a string without the
newline now.

builtin echo doesn't understand -n ?

Really ?

This is a hard dep iff the svg files need to be updated.

The `held == 0` branch in the Lck__AssertHeld function is *never*
taken and `cache.h` instructs to only use the macro. Being generated
from calling code, the panic message becomes more informative.

Panic messages were up until now not very helpful:

    Assert error in Lck__Assert(), cache/cache_lck.c line 175:
    Condition(ilck->held) not true.

E.g.

and

0 tests failed, 7 tests skipped, 558 tests passed

This could use some tlc but will do it for now.

VCC was silently accepting enum values valid for other arguments
following in the argument list as well as other identifiers in the
argument spec (for example "ENUM"). Consequently, wrong enum values in
VCL were not detected at VCC time and passed to vmod functions, which,
in the best case, would detect the error (and, if following the
varnish good practice, panic on a failed assertion).

This is another forgotten case since the enum list was changed to be
terminated by \1 in a78efad8: vcc_expr
would just loop over the \1 terminator up to the final \0 terminator
at the end of the argument spec.