On Sun, 2 Mar 2003 10:54:29 +0000, Tim Bunce wrote:
>On Sat, Mar 01, 2003 at 07:43:14PM -0800, Dean Arnold wrote:
>>"Jonathan Leffler" <[EMAIL PROTECTED]>�wrote:
>>>
>>>You show the 'install_method' working on
>>>DBD::Driver::db->install_method(...). �That appears to make it
>>>available on a $dbh - correct? �Is there an equivalent for the
>>>statement handle? �DBD::Driver::st->install_method()?
>>>
>>>After invoking DBD::Informix::db->install_method('ix_funky',
>>>undef),
>>>I believe I'd be able to run $dbh->ix_funky(...) without further
>>>ado
>>>(assuming $dbh was a DBD::Informix database handle). �Is that
>>>correct?
>>>
>>>I rather like the possibilities that this opens up, I have to
>>>say.
>>
>>Megadittos (to coin an overused phrase). Certainly is much more
>>aesthetically acceptable and easier to explain to users when
>>driver-specific functions are supplied.
>>
>>One little question (not entirely due to being too lazy to test
>>it):
>>does this all work with subclassed DBI drivers ? E.g., will it
>>work as is in DBIx::Chart wo/ any changes to DBIx::Chart ?
>
>Yeap.
>
>>Hmmm, actually, I suppose I should make the same inquiry
>>about all the changes that have been made since 1.32...at which
>>point I'm certain the answer is
>>"give it a shot and see what shakes out"...wish I had more cycles..
>
>"give it a shot and see what shakes out" :)
>
>>...wish I had more cycles..
>
>Ditto.
>
>Tim.Does this make sense: Suggestion to all driver writers. Check our your APIs for the most likely candidates to be used in an 'install_method' call. Make the calls in a static init section for the driver. Document in the driver's POD. -- Matthew O. Persico
