On Apr 17, 2008, at 12:48 PM, Michael G Schwern wrote:
Chris Dolan wrote:
Interesting reactions:
* People were appalled that Test::Class invokes methods in
alphabetic order instead of lexical order
It seems odd to me that people would be so concerned about the
order of execution, given the methods should be able to be executed
in any order. Any idea why that was an issue?
Compare the lexical order to the TAP output:
-------------------------------
package My::Test;
use base 'Test::Class';
sub foo :Test(1) {
pass('foo');
}
sub bar :Test(1) {
pass('bar');
}
1;
-------------------------------
1..2
ok 1 - bar
ok 2 - foo
-------------------------------
That's counter-intuitive. I appreciate that the order is highly
deterministic, but it's baffling to anyone who hasn't read the
Test::Class documentation carefully. Of course, that's just the
harness-less output. With a harness, T::C emits diag()s with the
method names, which helps dramatically.
I don't want to make too big a deal about this -- the invocation
order IS documented adequately -- but it was my first reaction and it
was a common reaction at my presentation.
Chris
