Re: Re: cvs: php-src(PHP_5_3) / Makefile.frag NEWS acinclude.m4 configure.in /ext/spl php_spl.c /ext/standard/tests/strings 006.phpt 007.phpt
highlight_file.phpt strip_tags_basic1.phpt strip_tags_basic2.phpt strip_tags_error.phpt strip_tags_variation1.phpt s

From: Date: Sun, 16 Mar 2008 21:54:02 +0000
Subject: Re: Re: cvs: php-src(PHP_5_3) / Makefile.frag NEWS acinclude.m4 configure.in /ext/spl php_spl.c /ext/standard/tests/strings 006.phpt 007.phpt
highlight_file.phpt strip_tags_basic1.phpt strip_tags_basic2.phpt strip_tags_error.phpt strip_tags_variation1.phpt s
References: 1 2 3 4  Groups: php.internals 
Request: Send a blank email to internals+get-36154@lists.php.net to get a copy of this message
Hello Edward, Sunday, March 16, 2008, 10:42:43 PM, you wrote: > Marcus Boerger wrote: >> - the tokenizer works a bit different as tokenizing works better now, that >> is uses less tokens. >> - the scanner is about 20% faster, overall execution is about 2% faster >> when not using compiler caches like APC >> - in some edgecases error messages are no longer corrupted > I'm not so much interested in performance changes as in end-user > behavior changes; will this affect *any* PHP code, or is it completely > backwards-compatible? (not counting multibyte support, of course) Some > possible areas I'm thinking based on browsing the patch are the > tokenizer functions, the ini parser interface, filename limits, cli > interface, etc. > It looks like some of the compilation docs/toolsets will need to be > updated. I'll try compiling 5.3 on Windows later today. Of course it affects that kind of tools. But the scanner was changed a lot of times for 5.3 already, so you have to do that anyway. That is the cost of making progress - changes. Best regards, Marcus

« previous php.internals (#36154) next »