Re: pdo: binding variables supplied to execute() is NotVeryUseful(tm)...

Hello Thies, Friday, April 1, 2005, 3:13:10 AM, you wrote: > Am 31.03.2005 um 17:28 schrieb Marcus Boerger: >> Hello Thies, >> >> Friday, March 25, 2005, 1:55:30 PM, you wrote: >> >>> and this is 'cause pdo _expects_ the user to prefix the bound >>> variables >>> with a colon. grr... >> >>> ppls, lets change it before it's too late. this "tiny bit" makes >>> binding >>> harder than it should be, and we all know and understand that all user >>> of php should use bound variables all the time! >> >>> suggestion: auto-trim a leading colon from the internal binding >>> tables. >>> that way "oldish" code would not break... >> >> >>> re, thies >> >>> PS: looking into that right now - hopefully "patch follows";-) >> >> >> We could easily add this behavior and i think it makes somewhat sense. >> It just looks a bit like __wakepup/__sleep. The only 'but' is that i >> suggest the behavior has its own method like 'bindParamArray'. If you >> cannot work out a patch feel free to contact me. Maybe i'll find some >> time during the conf. > hey marcus! > i don't quite understand the relation to sleep/wakeup here (i'm not > reading all the php lists, maybe you could point me to the right > thread?) > adding a new 'bindParamArray' with different semantics doesn't really > make things easier. and i think we should have once consistent > parameter-bind interface: > - a method bindParam > - an optinal associative array passed to execute. > all those functions expect to bind a <parameter name> to a statement. > this <parameter name> should work both with or without a leading colon. > by the way (just checked) - this is how the "native" oracle-oci binding > functions work. so oracle decided to make the colon in front of the > placeholder for bind optional. at the end of the day you can only bind > to a placeholder defined in your sql-statement. in the sql you have to > "mark" the placeholder by prepending a colon - but for the bind you > really don't need the colon. really! > so - again - lets make the colon *optional* - everywhere we allow > binding in PDO! > re, thies hmmm, this is just a copy'n'paste mistake somehow.... -- Best regards, Marcus mailto:mail@marcus-boerger.de