Re: Internal Filename Representations (was Re: Summary of I/O related RFCs)
[email protected] (Nathan Wiger) Sat, 12 Aug 2000 12:38:07 -0700
| Newsgroups | perl.perl6.language.io,perl.perl6.internals |
|---|---|
| Message-ID | <[email protected]> |
> I think it's far from trivial, especially if you want to take into > account network names, file versions, protection attributes and ACLs, ... Sure, this all depends on what we're going to do with this data and at what level. The biggest problem that modules like CGI.pm have to deal with is just putting together simple filenames. My example just seeks to address that problem. However, as you point out if we want to really make distinctions between directories, filenames, extensions, drives, servers, acls, etc, and moreover keep this stuff in a form that makes it directly usable (like "C" as the drive instead of "C:"), this all gets really complex really quickly. I was only addressing trying to split up "C:\Windows\System" and "/etc/motd" into a single format with one set of functions. I would argue that on this particular issue "less is more". If we could just deal with filenames universally this would be a big win (leave acls, permissions, versions, etc to something else). -Nate