On 13/11/02 16:08 +0100, [EMAIL PROTECTED] wrote:
>
> A problem we had with dynamic loading of Inline modules
> has been tracked down to the use of $_ in Inline::denter.
> With the attached patch it works OK
Wow. That's crazy. How could $_ affect anything? Would you mind sending a
brief expanation? I'll apply the patch.
Cheers, Brian
>
> /g�ran
>
>
>
>
> *** Inline-0.44/lib/Inline/denter.pm-orig 2001-06-13 06:30:50.000000000 +0200
> --- Inline-0.44/lib/Inline/denter.pm 2002-11-13 15:43:41.000000000 +0100
> ***************
> *** 177,182 ****
> --- 177,183 ----
> }
>
> sub _next_line {
> + local ($_);
> my $o = shift;
> $o->{done}++, $o->{level} = -1, return unless @{$o->{lines}};
> $_ = shift @{$o->{lines}};
> ***************
> *** 188,193 ****
> --- 189,195 ----
> $o->{done}++, $o->{level} = -1, return unless @{$o->{lines}};
> my ($width, $tabwidth) = @{$o}{qw(width tabwidth)};
> while (1) {
> + local ($_);
> $_ = $o->{lines}[0];
> # expand tabs in leading whitespace;
> $o->next_line, next if /^(\s*$|\#)/; # skip comments and blank lines
> ***************
> *** 221,226 ****
> --- 223,229 ----
>
> sub indent_data {
> my $o = shift;
> + local ($_);
> $_ = shift;
> return $o->indent_undef($_)
> if not defined;
