frank5us has asked for the wisdom of the Perl Monks concerning the following question:

First, some background info. I had working perl modules for Windows 10 running DWIM perl for many years. Got a new PC last week, windows 11. Dwimperl was having problems installing with cpan for spreadsheet modules, and so I decided to remove DWIMPerl and install the latest Strawberry perl instead. cpan installed my needed spreadsheet modules correctly, so some of my code was now working. But in this module, I get this error: Wide character at C:\perl-lib-ctp/Fmsc.pm line 27, <INFILE> line 1. 
sub readfile_to_arr { # take in unicode utf8 text, return as array of 
+lines
    my $infilename = shift;
    my @lines =  ();
    use Encode;

    open INFILE,  "<:utf8" , $infilename || die ("can't open $infilena
+me") ;
 
(line 27)    while 
(line 28)  (my $line = decode_utf8(<INFILE>))
    {
        push(@lines, $line);
    }

[download]
What's the error message complaining about? A wide character in my source code (which used to work fine on Windows 10 with DWIMperl), or is it complaining about contents of the actual <INFILE> being read in?
  • Comment on Wide character at C:\perl-lib-ctp/Fmsc.pm line 27, <INFILE> line 1.
  • Download Code

Replies are listed 'Best First'.
Re: Wide character at C:\perl-lib-ctp/Fmsc.pm line 27, <INFILE> line 1.
by Corion (Patriarch) on Dec 26, 2021 at 08:27 UTC

    These two lines in combination make little sense:

        open INFILE,  "<:utf8" , $infilename || die ("can't open $infilena
+me") ;
 
    while  (my $line = decode_utf8(<INFILE>)) {
    }

    [download]

    The first line already tells Perl to decode the lines in INFILE as UTF-8 bytes.

    The second line tries to decode again.

    I would lose the second line and rewrite it as:

        open INFILE,  "<:utf8" , $infilename || die ("can't open $infilena
+me") ;
    @lines = <INFILE>;

    [download]
[reply]
[d/l]
[select]

      Precedence problem there too; you need or instead of ||, or you need to parenthesize open's arguments. As is that's checking if $infilename is true (calling die if not), and then calling open and ignoring the return from it. 

      open( FH, MODE, FILE ) || die ERROR;
open FH, MODE, FILE or die ERROR;

      [download]

      The cake is a lie.
      The cake is a lie.
      The cake is a lie.

[reply]
[d/l]
[select]

      Slight correction: the "<:utf8" asserts that the input file is UTF-8, and it is decoded as such. The preferred mode for this case these days is "<:encoding(utf-8)", which actually checks the encoding, failing if an invalid encoding is encountered. Yes, "<:utf8" appeared in older Perl documentation, which shows that better people than me failed to appreciate the distinction in the early going. It is mostly gone now.

[reply]
[d/l]
[select]
        Thanks, I've incorporated that. When I write the utf-8 as a file, do I also need to wrap it with that check? 
        ######################################################################
+##################
sub write_file_utf8 {
    my $outfilename= shift; # appears in browser tabs
    my $outstr = shift; # file as a string
    ##################################################################
+##############
    open OUTFILE,  ">:utf8" , $outfilename 
        or die "cannot open >:utf8 $outfilename: $!";
    #binmode STDOUT, ":utf8";
    print OUTFILE $outstr, "\n";
    close(OUTFILE);
} #sub

        [download]
[reply]
[d/l]
      thanks! that made the 'wide char' error go away and it works now. I guess DWIMPerl didn't care about that error.
[reply]

Back to Seekers of Perl Wisdom