I'd be happy to allow "local_", "_", and perhaps "x_" to be used as 'private use' prefixes that would never be registered to a driver.
Tim. On Wed, Nov 01, 2006 at 06:49:37PM -0800, Ammon Riley wrote: > Hi, > > Is there a way to register a database prefix in the prefix registry, > without modifying DBI? > > We have a set of proxy machines sitting in front of an Informix database. > The proxies work by communicating a fairly strict protocol, which prevents > us from using DBD::Proxy. Consequently, I'm writing (or, rewriting, really) > a DBD:: driver to talk to our proxy hosts. > > The previous version of the driver was very rudimentary. I've now > implemented something that actually follows DBI spec, though some of > the more esoteric features don't work (but we don't use them, anyway). > > We have a few driver-specific functions that we use frequently. > Currently, that means calling them through $dbh->func(). That's really > annoying. I'd like to have the driver-specific functions available as > first-class methods, so that I can call them via $dbh->pts_func(), just > like everything else > > However, install_method() will only work if my driver prefix (I've > chosen "pts_") is a registered prefix. Obviously, since this DBD:: > driver has no application outside this particular company, it's not > going to be included in a distributed version of DBI. > > Consequently, I have two options, if I want to achieve my goal: > > 1) modify the DBI installed on our systems, making it more difficult > for our sys-admins to routinely update the DBI module. > > 2) register my driver in DBI, somehow. > > After looking at the documentation for DBI, and DBI::DBD, and then, > as a last hope, DBI::FAQ, it appears that the "somehow" for option > two does not exist. If I'm wrong, please enlighten me. If I'm not > wrong, then would it be possible for DBI to provide a way for DBD:: > writers to update the prefix registry, allowing install_method() > to work on an unregistered prefix? > > (I don't, at the moment, understand the inner workings of DBI well > enough to provide any possible solutions as to how it could work, > let alone a patch, and since I'm going to be on vacation for the > next bit, I figured someone else may have a good response in the > meantime.) > > Of course, if there are any options I haven't considered, please feel > free to point them out, as well. > > Cheers, > Ammon > > -- > Ammon Riley || Pipeline Setup || Rhythm & Hues || LA, CA || [EMAIL > PROTECTED]
