<script>
alert(encodeURIComponent('a å /'));
</script>
[download]

The thing you are referring to as "two sequences" is actually the two-byte sequence for the utf-8 encoded character U00E5. (updated the grammar slightly to make more sense)

Naturally, we'd love to have an elegant and concise way to interpret this correctly as utf8 text, but I don't know enough about the URI modules to provide much guidance in that direction.

So instead, I'll offer an ad-hoc (but still somewhat concise) work-around -- it's a kluge, but it should work until you or some other monk can find the needed gems in the appropriate module(s):

use Encode;

# ... get the uri string into $_ by whatever means ...

$_ = "a%20%C3%A5%20%2F";

# first, let's turn the uri encoded string (with "%HH" for some bytes)
+ into binary:

s/\%([0-9a-f]{2})/chr(hex($1))/egi;

# then, since this produces a utf-8 byte sequence, let's "decode" that
+ into utf-8

$_ = decode( 'utf8', $_ );

# $_ now has utf8 flag set, and contains the string with expected unic
+ode characters

binmode STDOUT, ":utf8";

print;
[download]

The "binmode STDOUT" thing could be taken out if you add a "-CO" flag on the shebang line, I believe -- that "perlrun" option does the same thing as 'binmode STDOUT, ":utf8";'.

I think that CGI's 'unescape' method does the same thing.

use CGI qw(unescape);
$_ = "a%20%C3%A5%20%2F";
$value = CGI::unescape($_);
[download]

(it's in CGI/Util.pm)

Thanks for your suggestion.
This does not solve my problem, unfortunately. I have found that the decode function in the Encode module should be able to do the job.


use Encode;
print decode('utf8', "\xC3\xA5");
[download]

This prints the character I need.
The problem now is:
How to go from %C3%A5 etc. to \xC3\xA5?
I tried

use Encode;
$_='%C3%A5';
 s/%/\\x/g;
eval { $_=$_};
print decode('utf8', $_);
[download]

but it does not work as expected. :-)

This was actually solved in the previous post by graff.