On Wed, 2009-05-27 at 16:41 -0600, Alex Rousskov wrote: > On 05/10/2009 11:08 PM, Robert Collins wrote: > > On Mon, 2009-05-11 at 16:12 +1200, Amos Jeffries wrote: > >> > >> PS. I took another look at the breakage that made me RFC, and I got > >> the auth and ACL code mixed up. It's auth that is currently fubar again. > >> But the principle holds: model is broken and we do need to fix it. > > > > I'd be strongly inclined to use a linker script to add references rather > > than the reg.cc files; this would mean we don't need to change code, and > > should be able to move to dynamic loading if desired (a bunch of .so > > files on disk, dlopen to use). > > What is a linker script? Do we have the expertise to maintain that?
A linker script is simply a set of instructions to the linker telling it some specific symbols to always include. Yes, I think we do have such expertise. > FWIW, I do not think the Reg.cc approach precludes the use of > dynamically loadable libraries, especially in the C++ world. For > example, eCAP modules auto-register themselves when dynamically loaded. > > Calling Init() in Reg.cc simply makes sure stuff is initialized before > use. If we move to dynamic libraries, the library can call the same > Init() method automagically. We may run into initialization order > problems, but those are unrelated to Reg.cc approach. Sure, I'll grant that. -Rob
signature.asc
Description: This is a digitally signed message part
