This patch *should* have the effect of a minor speed up
in the decode method (the envelope length-decoding
method). However, the real-world speedup is minimal in
my tests. What it does do is simplify the code
somewhat while generalising it ever so slightly (to
allow for length strings of any length that can be
specified by the format). There's also a few notes
about where there might be a few more operations that
could be eliminated.
Profiling is showing this method to be the current
hotspot (~5% total time (~19% cummulative)).
Log in to post a comment.