RE: [PEAR-DEV] yet another template

Hello Jean-Michel, >1) Yes,something like smarty's get_template_vars would be useful. It >returns the list of variables and enables automatic processing, for example >via a database repository. You told me that you were going to submit me the code that enables to do that. I will include the code after testing. >2) Basically I would like to be able to work with UTF-8 template and >generate SJIS files or whatever. >Have you thought about using mb_http_output('SJIS') to convert file output? >Or shall we use recode_file (string request, resource input, resource >output)? What is your opinion? I use mb_convert_encoding() to convert encoding of string. You can convert processed contents right before the output by this method, but this would make your code troublesome to maintain. I've never used mb_http_output(), so I am not sure if it is stable. My class uses mb* functions to parse the template file and these functions are called without encoding option. This result in these mb_* functions to use encoding set by mbstring.internal_encoding in php.ini file. If you set this option to UTF-8, used UTF-8 template and UTF-8 database, there would be no problem. You should ask this kind of question in PHP-I18N mailing-list. I think they'd better know these things. >3) Why not pearize smarty with multi-byte support and other enhancements. I >can help you if my help is welcome. I never used smarty, so I read through FAQ and QUICKSTART at their web site. It is very feature rich, and template file compiling is something I have been trying to implement in AT.php. However, I think template engine should leave anything that it doesn't have to do. I am not trying to offend anybody. It's just that my goal is slightly different from theirs. Like I said, I have no experience with smarty, so I talked mostly with assumptions. Addtionally, smarty's template syntax requires use of regex and my way of implementing multi-byte safe template file parsing cannot handle their somewhat loose syntax. I could utilize multi-byte safe regex "mbregex", but users have to fetch the module and recompile PHP with it. Maybe, we wait for another PHP release with its regex function handle multi-byte string safely. *Note: If you are developing Japanese web site, you should convert encoding of email headers and body to "JIS". (It's 7bit safe) Anyway, give me your opinion. Naoki Shima