Re: Vote: PEAR::Strip_Comments

Thorsten Suckow-Homberg wrote:

My suspicion is that the difference in access and/or overhead performance of code with comments and code without is negligible, particularly if you're using a PHP accelerator like Turck MMCache.

Right, if you choose an accelerator the cached files do not contain any comments anymore, that's what I have heard.

If you decide to run benchmarks, please post the tests you ran and their results.

Using the PEAR::Benchmark and the following code <?php require 'Benchmark/Timer.php'; $timer = new Benchmark_Timer(); $timer->start(); //require_once("HUGE_FILE.php"); require_once("SMALL_FILE.php"); $timer->stop(); $timer->display(); ?> where HUGE_FILE.php (48288) is my Image_Imagick.php proposal (http://www.siteartwork.de/image_imagick/Imagick.tgz) with comments and the SMALL_FILE.php (18475) (manually) stripped down (leaving white-spaces and new lines), I got the following results: HUGE_FILE.php ============ Lowest: 0.013940

Peak:    0.024210

SMALL_FILE.php ============ Lowest: 0.010831

Peak:    0.015010

(WindowsXP, Apache 1.3, PHP4.3, 512 MB RAM, 1.7 Celeron)

Can you re-run these benchmarks with Benchmark_Iterate? You will need to change your test code so its not a class obviously... - Davey