ngx_sprintf("%.2f", 0.999) incorrectly resulted in "0.100" instead of "1.00".

The proxy module context may be NULL in case of filter finalization
(e.g. by image_filter) followed by an internal redirect.  This needs
some better handling, but for now just check if ctx is still here.

Previously used fixed-point calculation caused wrong code path selection
in some cases, resulting in incorrect image size.

See here for report:
http://mailman.nginx.org/pipermail/nginx-devel/2012-April/002123.html

The problem occured if first uri in try_files was shorter than request uri,
resulting in reserve being 0 and hence allocation skipped.  The bug was
introduced in r4584 (1.1.19).

Instead of checking if there is events{} section present in configuration
in init_module handler we now do the same in init_conf handler.  This
allows master process to detect incorrect configuration early and
reject it.

consists solely of one unterminated token (inspired by #150).

Not using full chain passed is ok as consumers are expected to check
event's ready flag to determine if another call is needed, not the
returned size.

We now stop on IOV_MAX iovec entries only if we are going to add new one,
i.e. next buffer can't be coalesced into last iovec.

This also fixes incorrect checks for trailer creation on FreeBSD and
Mac OS X, header.nelts was checked instead of trailer.nelts.

The "complete" flag wasn't cleared on loop iteration start, resulting in
broken behaviour if there were more than IOV_MAX buffers and first
iteration was fully completed (and hence the "complete" flag was set
to 1).