In perl.git, the branch blead has been updated <http://perl5.git.perl.org/perl.git/commitdiff/b3766b12c64c46e0bcc2c1dc58cc7b96d8bef10c?hp=02512a66ad35b8ddc69cacdee9807210c5e13167>
- Log ----------------------------------------------------------------- commit b3766b12c64c46e0bcc2c1dc58cc7b96d8bef10c Author: Nicholas Clark <[email protected]> Date: Thu Oct 22 16:39:38 2009 +0100 Test requiring files with non-BMP characters (encoded as surrogate pairs). ----------------------------------------------------------------------- Summary of changes: t/comp/utf.t | 23 ++++++++++++++++++++--- 1 files changed, 20 insertions(+), 3 deletions(-) diff --git a/t/comp/utf.t b/t/comp/utf.t index 69ede95..00523f9 100644 --- a/t/comp/utf.t +++ b/t/comp/utf.t @@ -1,6 +1,6 @@ #!./perl -w -print "1..76\n"; +print "1..100\n"; my $test = 0; my %templates = ( @@ -13,7 +13,23 @@ sub bytes_to_utf { my ($enc, $content, $do_bom) = @_; my $template = $templates{$enc}; die "Unsupported encoding $enc" unless $template; - return pack "$template*", ($do_bom ? 0xFEFF : ()), unpack "U*", $content; + my @chars = unpack "U*", $content; + if ($enc ne 'utf8') { + # Make surrogate pairs + my @remember_that_utf_16_is_variable_length; + foreach my $ord (@chars) { + if ($ord < 0x10000) { + push @remember_that_utf_16_is_variable_length, + $ord; + } else { + $ord -= 0x10000; + push @remember_that_utf_16_is_variable_length, + (0xD800 | ($ord >> 10)), (0xDC00 | ($ord & 0x3FF)); + } + } + @chars = @remember_that_utf_16_is_variable_length; + } + return pack "$template*", ($do_bom ? 0xFEFF : ()), @chars; } sub test { @@ -45,8 +61,9 @@ for my $bom (0, 1) { # right now, as here we're testing the input filter itself. for my $expect ("N", "\xFF", "\x{100}", "\x{010a}", "\x{0a23}", + "\x{10000}", "\x{64321}", "\x{10FFFD}", ) { - # A space so that the UTF-16 heuristc triggers - " '" gives two + # A space so that the UTF-16 heuristic triggers - " '" gives two # characters of ASCII. my $write = " '$expect'"; my $name = 'chrs ' . join ', ', map {ord $_} split '', $expect; -- Perl5 Master Repository
