diff options
| -rw-r--r-- | lib/charnames.t | 19 |
1 files changed, 13 insertions, 6 deletions
diff --git a/lib/charnames.t b/lib/charnames.t index 75a3aaccb1..74b0999242 100644 --- a/lib/charnames.t +++ b/lib/charnames.t @@ -780,20 +780,27 @@ is("\N{U+1D0C5}", "\N{BYZANTINE MUSICAL SYMBOL FTHORA SKLIRON CHROMA VASIS}"); defined $ENV{PERL_TEST_CHARNAMES_SEED}; $seed = srand($seed); + # We will look at the data grouped in "blocks" of the following + # size. + my $block_size_bits = 7; # above 16 is not sensible + my $block_size = 2**$block_size_bits; + # There are the regular names, like "SPACE", plus the ones # that are algorithmically determinable, such as "CKJ UNIFIED # IDEOGRAPH-hhhh" where the hhhh is the actual hex code point number # of the character. The percentage of each type to test is # independently settable. my $percentage_of_regular_names = 25; - my $percentage_of_algorithmic_names = 1; + my $percentage_of_algorithmic_names = 100 / $block_size; # 1 test/block - my @names; # The names of every code point. + # Changing the block size doesn't change anything with regards to + # testing the regular names, but will affect the algorithmic names. + # If you make the size too big so that blocks include both regular + # names and algorithmic, the whole block will be sampled at the sum + # of the two rates. If you make it too small, then more algorithmic + # names will be tested than you probably intended. - # We will look at the data grouped in "blocks" of the following - # size. - my $block_size_bits = 7; # above 16 is not sensible - my $block_size = 2**$block_size_bits; + my @names; # The names of every code point. # We look at one block past the Unicode maximum, to verify there are # no names in it. |