in reply to Re^5: true from (-e "") on Windoze (" is an illegal filename character
in thread true from (-e "") on Windoze

I was talking about -e '""' (a system call), not dir "" (a command line tool). What's appropriate for one is not necessarily appropriate for the other.

That said, I'm not convinced the behaviour of dir "" is the best. Note that ls "" returns ls: cannot access : No such file or directory.

Replies are listed 'Best First'.
Re^7: true from (-e "") on Windoze (" is an illegal filename character
by BrowserUk (Patriarch) on Sep 17, 2012 at 20:06 UTC
    I was talking about -e '""' (a system call), not dir ""... What's appropriate for one is not necessarily appropriate for the other.

    The system call that underlies dir is the same system call that underlies -e, thus similar behaviour is entirely appropriate.

    Making them different would be the wrong thing to do.

    I'm not convinced the behaviour of dir "" is the best.

    Convinced or not; best or not; it is the behaviour, for better or worse.

    Making Perl work differently would be as asinine as truncating 32-bit return codes to 8-bits. (Whatever happened to the the principle of maximum information preservation)

    The principle of least surprise dictates that you don't screw with the heads of developers on a given platform, in order to accommodate the expectations of programmers cross-targeting it from other platforms.

    Note that ls "" returns ls: cannot access : No such file or directory.

    The fact that *nix has an anything goes policy to filename characters, doesn't make it a good idea.

    Think of all the extra code written/maintained to accommodate for/protect against all the accidental/malicious possibilities that policy creates

    But, I doubt we'll ever agree, so let's agree to differ.

    With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
    Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
    "Science is about questioning the status quo. Questioning authority".
    In the absence of evidence, opinion is indistinguishable from prejudice.

    RIP Neil Armstrong
[reply]

      The system call that underlies dir is the same system call that underlies -e, thus similar behaviour is entirely appropriate.

      The behaviour of Perl functions is not dictated by what some system call does.

      Making Perl work differently would be as asinine as truncating 32-bit return codes to 8-bits.

      I would like Perl to stop working differently than documented. Neither that desire nor the documented behaviour is asinine.

[reply]
        I would like Perl to stop working differently than documented.

        Don't tell me, let me guess.

        On the basis of that piece of self-asserting, broken and twisted logic, you are going to change perfectly valid and reasonable platform specific behaviour on windows, to fit the undocumented and platform specific behaviour exhibited on *nix.

        Nowhere is the behaviour of -e '""' documented. Two different behaviours are observable, that fit with the local patterns and conventions of those platforms. And you are going to break a useful and logical feature of one platform to make it fit the non-useful and illogical behavour on the other. Rather than just document the platform difference.

        Now, THAT is the very definition of asininity!

        With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
        Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
        "Science is about questioning the status quo. Questioning authority".
        In the absence of evidence, opinion is indistinguishable from prejudice.

        RIP Neil Armstrong
[reply]
[d/l]

In Section Seekers of Perl Wisdom