Front page | perl.perl5.porters |
Postings from March 2003
On Sat, Mar 29, 2003 at 04:15:58PM +0900, Dan Kogai wrote:
> perldoc -f require says;
> >sub require {
> > my($filename) = @_;
> > return 1 if $INC{$filename};
> > my($realfilename,$result);
> > ITER: {
> > foreach $prefix (@INC) {
> > $realfilename = "$prefix/$filename";
> > if (-f $realfilename) {
> > $INC{$filename} = $realfilename;
> > $result = do $realfilename;
> > last ITER;
> > }
> > }
> > die "Can't find $filename in \@INC";
> > }
> > delete $INC{$filename} if $@ || !$result;
> > die $@ if $@;
> > die "$filename did not return true value" unless $result;
> > return $result;
> >}
>
> Therefore it is perl 5.8.0 that is behaving correctly. And "eval
> q(require Module)" still works since It does not kill perl itself.
By that logic, "require Module" should die, but its not.
The behavior of:
# loads Pod::Man fine, doesn't die.
require Pod::Man;
and
# doesn't load Pod::Man, dies.
eval q{require Pod::Man}; die $@ if $@;
and
# loads Pod::Man fine, doesnt' die.
eval q{require Pod::Man; 1}; die $@ if $@;
should be consistent. In 5.8.0 they are not. There *is* a bug in eval
STRING here.
Here's another interesting data point:
$ perl5.8.0 -wle 'print require Pod::Man;'
142
$ perl5.8.0 -wle 'print eval q[require Pod::Man]; print $@'
142
So it *is* returning a true value. Not sure where its coming from, but its
there. Its also bizarre that "print eval q[require Pod::Man]" is different
from "eval q[require Pod::Man]" in that the module is loaded and $@ is not
set.
Slapping "1;" on the end of modules might be useful, that's a different
debate, but it will not fix this bug.
--
You're smoother than a tunnel of shining sorrow.
Thread Previous
|
Thread Next