Currently, one of the PerlIO layers internal to core is :mmap. It is documented (per perliol) as «it is mainly there as a proof of concept. It is likely to be unbundled from the core at some point». There are three things we can do with it, each depending on the ones above it: 1. We can move it from core-internal to a core module. The only visible consequence of that is that (like :encoding) it's no longer possible to run «PERLIO=mmap ./foo.pl». I don't think that's a particularly sane thing to do anyway (it surprises me only one test segfaults when trying to run core tests with PERLIO=mmap). Obvious consequence is making the core slightly smaller and simpler. 2. We can then dual-life it. I can't see any reason not to do this. 3. We could deprecate it and later remove it from core. This obviously is user visible, though only one module on CPAN seems to be are currently using it. Quite frankly I see no convincing reason not to do 1 & 2. I don't care much about 3, but if we ever are going to do a major sweep of core modules it should be an obvious candidate. LeonThread Next