eak has asked for the wisdom of the Perl Monks concerning the following question:

I ran into a very strange html/perl condition today. The following executes as valid perl: 
#!/usr/bin/perl -w
 
use strict;
use diagnostics;
 
print "hello\n";
 
</html>

[download]
If you examine $! after '</html>' you will see 'Illegal seek'. If you change </html> for <HTML>, an exception is raised. Can someone tell me what the forward slash does to the bareword? thanks, eric

Replies are listed 'Best First'.
Re: /HTML is valid perl?
by japhy (Canon) on Oct 23, 2001 at 07:52 UTC
    To Perl, <HTML> looks like a line-read from the filehandle HTML. However, if you have something that isn't a filehandle (or a variable holding a filehandle) in there, Perl sees it as a file glob. </HTML> is the glob /HTML. Even something as simple as whitespace (as shown below) triggers this distinction.

    Here's some fiendish Perl: 

    <perl >;
<head type="pragma">;
use strict;
</head >;
<body >;

my $name = "japhy";
my $code = "evil";
my $reasons = "plenty";

</perl >;

    [download]

    _____________________________________________________
    Jeff[japhy]Pinyan: Perl, regex, and perl hacker.
    s++=END;++y(;-P)}y js++=;shajsj<++y(p-q)}?print:??;

[reply]
[d/l]
Re: /HTML is valid perl?
by DamnDirtyApe (Curate) on Oct 23, 2001 at 07:56 UTC

    The forward slash mutates the angle operator into a filename globbing operator. Thus, </html> is actually returning a list of all the filenames which it matches. Similarily, </*.html> should return a list of HTML files in your root directory.

    If you can, see Programming Perl, 3rd ed., page 83.

    Good luck. ;-)

[reply]
[d/l]
[select]
(tye)Re: /HTML is valid perl?
by tye (Sage) on Oct 23, 2001 at 10:40 UTC

    But I was surprised that -w didn't elicit a "useless use of file glob in void context" warning.

            - tye (but my friends call me "Tye")
[reply]
B::Deparse knows all
by Fletch (Bishop) on Oct 23, 2001 at 21:00 UTC

    If you're ever in doubt how perl is interpreting something, just ask it to dump how it's being parsed. 

    $ perl -MO=Deparse,-p -e '</html>'
CORE::GLOBAL::glob('/html', 0);
-e syntax OK

    [download]

    See perldoc B::Deparse for the options you can give it (should it obsessively parenthesize, etc.).

[reply]
[d/l]
      While its true that B::Deparse is a cool tool it should be understood that there can be subtle differences between what it outputs and what is actually happening. I gave an example of this in Regarding B::Deparse a few days ago. Do not assume that the ouput is necessarily correct until it has been checked.

      Yves
      --
      You are not ready to use symrefs unless you already know why they are bad. -- tadmc (CLPM)

[reply]

Back to Seekers of Perl Wisdom