On Thu, Feb 14, 2013 at 3:04 PM, Daniel Vandersluis <
reality.tv.add...@gmail.com> wrote:

> Is it just me, or is Haml 4 slower than Haml 3.2? I created a sample haml
> page using html2haml (about 200 lines of haml code), and used the benchmark
> script I found at
> http://www.coffeepowered.net/2011/06/17/making-haml-faster/:


TL;DR - you're right.

Haml 3.1.4 contains an incompatibility with Rails 3.2.3+ whose fix impacted
performance. You can read about this in the commit which introduced it:

https://github.com/haml/haml/commit/cda687aaf3d8f89f9fba28c4bc137640afe98562

So ideally you should compare Haml 3.1.8 against 4.0, since those are the
latest versions of each stable release that work with the current version
of Rails. In that comparison Haml 3.1 still comes out slightly ahead,
except on parsing performance (though that's not especially important in
production).

On 1.9.3 the difference is slight, but unfortunately on JRuby Haml 3
is surprisingly twice as fast as Haml 4. Here are the results:

https://gist.github.com/norman/4977796

The benchmark source is here: https://github.com/norman/haml-bench

I'm looking into what caused the performance regression today. Performance
wasn't a focus of the 4.0 release but we obviously didn't intend to
introduce any regressions either. The earlier alphas of Haml 4.0 did in
fact have slightly better performance across the board and I had expected
to maintain that into the release. Mea culpa for not keeping a closer eye
on the benchmarks.

Akira Matsuda is currently working on improving performance, and you can
see his commits in the master branch right now. We should have a 4.0.x
release out within the next few weeks that will include his improvements,
and I'll keep the list posted regarding the regressions from 3.1 to 4.0.

Regards,

Norman

-- 
You received this message because you are subscribed to the Google Groups 
"Haml" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to haml+unsubscr...@googlegroups.com.
To post to this group, send email to haml@googlegroups.com.
Visit this group at http://groups.google.com/group/haml?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.


Reply via email to