Front page | perl.perl5.porters |
Postings from October 2011
On Fri Oct 28 18:02:14 2011, sprout wrote: > On Fri Oct 28 07:40:43 2011, jpl@research.att.com wrote: > > The best should not be the enemy of the good. At minimum, there should > > be some documentation changes to File::Glob, removing the ':glob' from > > the synopsis, and warning about its use. Done with commit 5144542d35. > > We can also add the :bsd_glob export that I suggested, but without > support for lists at first--that can come later. Done with commit f4cbf9907d. > > > I'm not sufficiently > > knowledgeable about keyword overrides to offer any other ideas. (But at > > least I'm reading this:-) ) -- jpl > > > > I’m still thinking about it. I think that the glob() magic should > depend solely on the scope in which the subroutine is defined, and not > on the caller, as a sub defined under ‘use v5.16’ will not be expecting > the magic second argument and should not have to check the caller’s > feature hints to determine whether the second argument was passed in by > the user or was a ‘perquisite’. > > Now, should the same thing apply to require? I.e., should subs compiled > under ‘use v5.16’ get the magical bareword treatment (automatic > s|::|/|gr . ".pm")? I think they should. > I searched CPAN and did not find any non-core modules that use the magic second argument to glob(), so it sounds as though we don’t need any fancy magic. We could just make glob() fully overridable (under ‘use v5.16’), plain and simple. However, this is still just making my brain go round in circles, so I will set it aside for now. -- Father ChrysostomosThread Previous | Thread Next