Encode module and inverted commas - perl

Question

I am scraping a web page, and extracting a specific section from it. That section includes inverted commas (’, character 146). I'm trying to print my extracted data to a text file, but it's giving me ’ instead of the inverted comma. I have tried the following:
$content =~ s/’/'/g;
my $invComma = chr 146;
$content =~ s/$invComma/'/g;
$content =~ s/\x{0092}/'/g;
None of it has worked. I can't decode('UTF-8', $content) because it has wide characters. When I try to encode('UTF-8', $content) the ’ changes to ’ instead. I have already tried use utf8 as well, to no effect.
I know that my text file viewer can display inverted commas, because I printed one to a test file and opened it. The problem is therefore in my script.
What am I doing wrong, and how do I fix it?
UPDATE: I am able to do $content =~ s/’/'/g to replace it with a simple apostrophe, but I still don't know why nothing else works. I'd also like a fix that actually solves the problem, instead of just solving one of the symptoms.
UPDATE 2: I have been informed by hobbs that the character is actually U+2019 RIGHT SINGLE QUOTATION MARK and changed my regex to use chr 0x2019 which now works.

Answer 1

The character you're trying to replace is only 0x92 / 146 in the Windows-1252 encoding. Perl uses Unicode, where that character is U+2019 RIGHT SINGLE QUOTATION MARK, aka "\x{2019}", chr(0x2019), or chr(8217).

Answer 2

Start by finding out what $content contains. You can use the following:
use Data::Dumper;
local $Data::Dumper::Useqq = 1;
warn(Dumper($content));
If you get the following, $content is decoded
$VAR1 = "...\x{2019}...";
Any of the following will work.
use utf8; # Source code is encoded using UTF-8.
$content =~ s/’/'/g;
$content =~ s/\x{2019}/'/g;
$content =~ s/\N{U+2019}/'/g;
$content =~ s/\N{RIGHT SINGLE QUOTATION MARK}/'/g;
If you get the following, $content is encoded using UTF-8.
$VAR1 = "...\342\200\231...";
Start by decoding the value of $content using either of the following:
utf8::decode($content) or die;
use Encode qw( decode_utf8 );
$content = decode_utf8($content);
Then use any of the solutions for decoded content (above).
If you get the following, $content is encoded using cp1252.
$VAR1 = "...\222...";
Start by decoding the value of $content.
use Encode qw( decode );
$content = decode("cp1252", $content);
Then use any of the solutions for decoded content (above).
By the way, ’ is what the UTF-8 encoding of ’ (E2 80 99) would look like if decoded as cp1252.

Answer 3

The problem was not in my script, it was in my editor. The script works properly, and the question is based on false pretenses. I was using gVim on Windows, which did not play nicely with Unicode. My script was properly decoding the content, but when I opened the output file in gVim, it mangled the text and displayed it incorrectly. My attempts to use regular expressions to change the characters failed because I was using the wrong codepoint - it wasn't 0x92, it was 0x2019. This was another failing of gVim. Thanks to hobbs and ikegami for helping me figure this out.