removing features from the language

You're mistaken. Temporarily retracted while the discussion about Perl 8+ is ongoing; Perl 7 will however not remove them.

Therefore, "globals" and file-scope lexicals actually have the same scope ...

That's another whole paragraph explaining things that I know (I too have a background in C and plenty more) to support a point nobody is disputing.

... and the same risks.

No, this does not follow from the explanation you gave, since among other things, it ignores that bareword filehandles are not protected against typos like lexicals are.

File-scope lexicals look like lexicals, but have effectively the same scope as globals.

More repetition of already discussed and undisputed points.

In terms of surprise action-at-a-distance, they carry the same risks.

No, because this ignores the issues of bareword filehandles clashing with package and sub names, among others.

You continuing to "stick your head in the sand" and ignore the disadvantages of bareword filehandles and advantages of lexical filehandles will not make them go away. This is not the basis for a rational discussion, so I'm out.

Nope, again, there are plenty of ways to avoid a lexical to be closed over, like writing the main code after the subs are declared or putting a block around it.

And people who ignore lexical filehandles, don't know about packages either.

«...writing the main code after the subs are declared or putting a block around it.»

You mean something like this I guess:

#!/usr/bin/env perl

use strict;
use warnings;

KARL: {
    nose();
}

sub nose {...;}

__END__
[download]

I wrote my code many years like this and I can’t tell how often I’ve been blamed for it.

«The Crux of the Biscuit is the Apostrophe»

perl -MCrypt::CBC -E 'say Crypt::CBC->new(-key=>'kgb',-cipher=>"Blowfish")->decrypt_hex($ENV{KARL});'Help

> You mean something like this I guess:

Almost

    
use strict;
use warnings;

MAIN: {
    open my $fh,">","filename";
    ... 
    nose();
    ...
}

sub nose {...;}
[download]

nose() won't have access to any lexical vars defined inside the main block.

> I can’t tell how often I’ve been blamed for it.

It's a variation of putting the main code into a main() sub and calling it from the start.

I think it's a pattern taken from C and I've seen some monks here advocating it.

Cheers Rolf
_{(addicted to the Perl Programming Language :)
Wikisyntax for the Monastery}

If you put a block around a file-scope lexical, it is not a file-scope lexical anymore. Writing the main code after subs are defined is another "just be careful" convention — it works (until the convention is broken), but if you do not accept conventions as legitimate solutions to problems with bareword filehandles, then why do you accept conventions as solutions to problems with lexicals?

The last issue of programmer ignorance is not a problem that can be solved by changing the language, only by educating the programmers about the language in which they write.

> If you put a block around a file-scope lexical, it is not a file-scope lexical anymore.

My point was that the scope of lexical variables can be easily limited.

You can't our FHs to limit the scope so you'd need different packages.

> only by educating the programmers about the language in which they write.

Well we wouldn't need any other languages if programmers were educated enough to use assembler.

FHs need a lot of extra technique, educating a new programmer about all the tricks needed takes a while, without extra benefit.

Discouraging their use doesn't mean changing the language.

Cheers Rolf
_{(addicted to the Perl Programming Language :)
Wikisyntax for the Monastery}