Re: clarification on the performance of require_once andAPC
| From: | Alexey Borzov | Date: | Mon, 27 Aug 2007 11:52:03 +0000 |
| Subject: | Re: clarification on the performance of require_once andAPC | ||
| References: | 1 2 3 4 5 6 7 8 9 | Groups: | php.pear.dev |
| Request: | Send a blank email to pear-dev+get-47861@lists.php.net to get a copy of this message | ||
Hi,
Lukas Kahwe Smith wrote:
And here we will know that there's our code running processed by our build script.Lukas Kahwe Smith wrote:Yes. At least then we know that its our code running plus some class loading mechanism. Modifying code on install not only poses risk in the actual modification process,...but we definitely do want to support both "unzip-and-go" and "pyrus-installed" versions of packages, as well as third-party __autoload() implementations. ::snicker::For the performance freak, have a phing tool which strips it out and creates "allfiles.php" (I don't like that name at all).Stripping out code on the fly in the installer is crazy. We do not want to have magically adapted code versions flying around.
but also in other areas like the well known issue that many users report stuff like "on line xyz there is foobar".I'd take a bug report with proper "actual output" vs. "expected output" over that with explicitly stated line numbers any day. The latter come out bogus too often. Not mentioning the fact that using the same build script on the same files will produce the same output.
Your attempted comparison does not hold any water.Try again.