On 1/31/06, Reini Urban <[EMAIL PROTECTED]> wrote: > 2006/1/31, demerphq <[EMAIL PROTECTED]>: > > On 1/31/06, Reini Urban <[EMAIL PROTECTED]> wrote: > > > 2006/1/31, demerphq <[EMAIL PROTECTED]>: > > > > IMO we should be bundling some of the base modules from the libwin > > > > distro in core. Win32API::File being one of them, along with > > > > Win32API::Registry and Win32::TieRegistry, Win32::IPC, Win32::EventLog > > > > (the latter becuae we bundle *NIX specific counterparts so why should > > > > Win32 be left out?) > > > > > > Ok for all (esp. Win32API:: and IPC). > > > But not Win32::TieRegistry please! > > > This is very high-level and pollutes the namespace with a magic global. > > > > Only if you dont tell it to do otherwise. See the TiedRef parameter. > > > > And personally I think high-level in this area is a good idea. > > > > Are you really going to say that youd prefer to use Win32API::Registry > > instead of Win32::TieRegistry (which is as you said a wrapper around > > Win32API::Registry) > > High-level is of course ok. > I love Win32::TieRegistry and never used Win32API::Registry directly. > But putting it into core would be very dirty, with the current TiedRef > default.
Personally I dont see the issue. The module explains in detail about this subject so i cant see people using it having a problem. Its not like its going to be randomly loaded into someones namespace without them knowing. yves -- perl -Mre=debug -e "/just|another|perl|hacker/"