Well, I ran into a snag. I tried doing this:
{
my($Skipped, $Skip_Reason);
sub test_these (&;$) {
my($code, $how_many) = @_;
goto $code;
# skip() will jump to here to bail out of the $code.
CONSIDERED_HARMFUL: if( $Skipped ) {
$how_many ||= 1;
_skipped($Skip_Reason) for 1..$how_many;
}
# Reset our state.
$Skipped = $Todo = 0;
$Skip_Reason = $Todo_Reason = undef;
}
sub skip {
$Skipped = 1;
$Skip_Reason = shift;
# Yes, an actual goto LABEL! This is what lets skip() abort the
# test_these block.
goto CONSIDERED_HARMFUL;
}
sub todo {
$Todo = 1;
$Todo_Reason = shift;
}
}
Problem is, the goto() inside skip() won't jump to the label inside
test_these(). The trouble is the goto $code, used to preserve the
caller's context (I want the test_these() block to act as much like a
normal block as possibe). goto LABEL looks up the callstack for the
label, but the C<goto $code> hijacks the callstack, so it never sees
test_these() as part of it and thus can't see CONSIDERED_HARMFUL.
Two ways around this:
1) Don't worry about preserving caller() information in the
test_these() block.
This bothers me. Ideally, a test function should interfere as little
as possible with the code.
2) Try a real block.
I can make this work easily for skips:
SKIP: {
skip($why, $how_many) if $condition;
ok( $some_test );
ok( $some_other_test );
}
using a little known hack that you can use C<last> from outside
(lexically outside) a block.
sub skip {
my($why, $how_many) = @_;
$how_many ||= 1;
_ok_skipped($why) for 1..$how_many;
no warnings;
last SKIP;
}
that takes care of skip. What about todo?
TODO: {
todo($why);
ok( $some_test );
ok( $some_other_test );
}
and write todo() the same way as before. ok() would note that the
$Todo flag was set and act accordingly. Problem is, I need to unset
that $Todo flag at the end of the block. Can't think of a way to do
that.
--
Michael G. Schwern <[EMAIL PROTECTED]> http://www.pobox.com/~schwern/
Perl6 Quality Assurance <[EMAIL PROTECTED]> Kwalitee Is Job One
A tickle of paste
Teases down my sweaty thigh
Oh, plug it again!
-- ignatz
