1) Could you provide an executive summary of what DBI::Pool is intended to do ? (to refresh my addled memory, and so we don't all have to crawl the mail archives; I didn't find anything meaningful in the POD or readme)
The idea was to create a module which does connection pooling under perl threads, just like Apache::DBI does connection pooling for mod_perl processes (and eventually may provide a superclass for Apache::DBI which will be able to handle Apache2 threaded mpm implementations).
The implementation is very simple (you can check the code it's just a few dozens lines of code):
DBI::Pool hijacks the connect method, on the first login the operation will be normal, but instead of destroying the connection object, it'll rip the low-level connection data structure which is perl interpreter-independant and store it in the storage. The next time a thread wants a connection it will be served from the local storage if there are available or a new connection will be created). To serve a cached connection, a perl dbh will need to transplant the low-level connection data.
2) What impact, if any, does this have on pure perl drivers ?
I'm not familiar with the implementation to be able to tell. The main problem is that in order to share the connections between threads, you need to make that data perl-interpreter independant. If you store any perl scalar you can't use it in a perl thread which didn't create it. If the pure perl driver can somehow freeze a connection data which can then be thawed by a different perl, then it will certainly work. so you tell me :)
p.s. notice that the name of this module may change, this is just something I've started to prototype with. Suggestions are welcome.
-- __________________________________________________________________ Stas Bekman JAm_pH ------> Just Another mod_perl Hacker http://stason.org/ mod_perl Guide ---> http://perl.apache.org mailto:[EMAIL PROTECTED] http://use.perl.org http://apacheweek.com http://modperlbook.org http://apache.org http://ticketmaster.com
