Re: Configuring number of parsers and buffer sizes

Nathan Bubna <[email protected]>
Newsgroups gmane.comp.jakarta.velocity.user
Message-ID <CAFyaDjF4ZeNRM=a69U5+QmHVNJKC31PkMq0THydTGA1pUaNJ+A@mail.gmail.com>
parser.pool.size

is this in 1.7?  i thought we trimmed a lot of the excess between 1.5
and 1.6, but if we can trim more, that'd be swell.

On Thu, Apr 26, 2012 at 5:59 PM, Christopher Schultz
<[email protected]> wrote:
> All,
>
> I've been looking at the docs for both Velocity Tools and Velocity
> itself and I can't seem to find out what the properties are for
> controlling the number of template parsers to create and how big their
> buffer sizes are.
>
> I've been doing some memory profiling of my webapp and it looks like
> Velocity is keeping some rather large (and empty -- probably because
> they haven't yet been used) buffers for parsing.
>
> I can see a bunch (76) of int[4096] arrays within VelocityCharStream
> (the class member is 'bufline'), and I can also see buried in
> VelocityCharStream which uses an InputStreamReader ('inputStream') a
> byte[8192] which appears to exist 48 separate times.
>
> So, the total "wasted" space due to these two arrays (duplicated a bunch
> of times) is, on the face of it, small: it's like 1-2MiB. On the other
> hand, if I don't need all that buffer space, I'd like to remove it.
>
> So, first: how can I configure the number of parsers and possibly the
> buffer sizes used by them?
>
> Second: is VelocityCharStream essentially double-buffering since the
> InputStreamReader seems to have a large buffer inside it and also
> VelocityCharStream has its own buffer?
>
> Thanks,
> -chris
>
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.