Performance improvements

[eugenpodaru]

For large strings/byte arrays the library becomes unusable. I've made some small changes that improve the performance by 100% to 200%:

• use pop/push over shift/unshift for the most common Stream methods (read, prepend)
• remove the call to endOfStream function from encode/decode since read already performs an endOfStream test

The biggest performance improvement comes from changing the Stream implementation.

[inexorabletash]

I'd prefer to keep inRange out of line. Otherwise the changes look good.

[eugenpodaru]

@inexorabletash Ok, fair enough.

Should I put inRange back out of line?

[inexorabletash]

That would make the merge easier!

On Dec 28, 2015, at 7:48 AM, Eugen Podaru notifications@github.com wrote:

@inexorabletash Ok, fair enough.

Should I put inRange back out of line?

— Reply to this email directly or view it on GitHub.

[eugenpodaru]

Ok, done. I've reverted to the commit with out of line inRange.

[inexorabletash]

Thanks!

[eugenpodaru]

Great! When is the next release?

[inexorabletash]

I pushed some refactoring changes up; can you sanity check that I didn't regress your performance? If that looks good I'll tag a release.

[eugenpodaru]

It seems to be alright