conceptual questions

Hello Developers, The idea of PECL sounds really great but I'm wondering how you could handle the differences in detail. First of all my scepticism: I'm really happy about the project and look forward to get it for testing. But I got are a few questions to your concept: [1] Do you want to parse a SQL-statement coming from the userland by analysing whether it was been created for i.e. for Postgres and than translate it in the specific language for the target dbs like MySQL. Or do you wanna use an own META-Language like XML like it will propably be used by PEAR::MDB2. And if so, is that really fast than? [2] Another problem are UNION Commands I think. Do you want to split a UNION-Command coming from the application i.e. for MySQL? I believe this style of SQL you can't execute at MySQL 4.x at present. [3] What about the following construct: SELECT FROM tab1 WHERE tab1_id IN ( SELECT tab2_id FROM tab2 ); OK - the given example could likely be transformed into a JOIN construct, but it is a valid SQL and should be possible to execute. How do you want to handle this f.e. on MySQL? [4] Another case for a problem-situation could be the follwing if the name of col1 is a reserved word like 'count'. Then you have to set the col-name in quotes. PG: INSERT INTO tab ("col1","col2") VALUES ( 'val1', 'val2' ); MySQL: INSERT INTO tab (Žcol1Ž,Žcol2Ž) VALUES ( 'val1', 'val2' ); This is surely a detail and easy to handle by string-replacement but nevertheless interesting to know how you would like to handle this cases in PECL. [5] There a lot of questions more like how to react on a CREATE function statement in specific language [like PL/PGSQL] or TRIGGERS, CONSTRAINS a.s.o. f.e. on MySQL. I would be glad to here from you because I hope that cases are considered somehow in the PECL-concept. Which PHP-Version will be the base for PECL? When do you think it is ready for a seroius test [beta-state or something like that]? Regards Rico Ortmann [ro@radax.net]