You can just do Test::Builder->new to get the Test::Builder object. It will be the same one used by Test::More because it's a singleton. That way you should need no patches,
In the end I came up with this code. It's pretty simple and straightforward and maybe would be nice to include with Test::More.
====================================================== =head2 todo_some
Execute given test subroutine while setting $TODO flag for given test numbers.
IN: 1 hash reference, with testnumber / reason strings pairs
2 reference to subroutine to execute test
3..N parameters to pass to test subroutine
OUT: 1 whatever test sub returnsSometimes when writing tests, some tests may fail. You can mark these tests as TODO but this is only convenient for a single range of tests.
Sometimes a few tests may fail inside a set of nested loops. But marking them in the flow of the program is very complicated.
The C<todo_some> method allows you to specify which tests are supposed to fail, specified by test number and reason for the failure.
Example:
my %todo = ( 1023 => "d f g still fails, but it's not that important", 1567 => 'yikes, we need to fix this one!', );
foreach my $foo (@foo) {
foreach my $bar (@bar) {
foreach my $car (@car) {
todo_some( \%todo,\&ok,func($foo,$bar,$car),"check result of func()" );
}
}
}
=cut
sub todo_some {
local $TODO = shift->{$Test->current_test + 1} || '';
no strict 'subs'; # allow test func specification by name
my $sub = shift;
$sub->( @_ );
} #todo_some
======================================================Liz
