On Fri, May 4, 2012 at 12:57 PM, Shlomi Fish <[email protected]> wrote:
> Hi all,
>
> in reflection about Gabor’s talk yesterday, I’ve been spending some time
> refactoring the code of the Config::IniFiles CPAN module, of which I am a
> co-maintainer, and which I adopted because I’ve made use of it for one
> project.
>
> While doing that I’ve ran into this gem:
>
> sub GetFileName
> {
> my $self = shift;
> my $filename;
> if (exists $self->{cf}) {
> $filename = $self->{cf};
> } else {
> undef $filename;
> }
> return $filename;
> }
>
> It's indicative of a lot of ignorance of how Perl 5 works.
>
> This ended up being shortened into:
>
> sub GetFileName
> {
> my $self = shift;
>
> return $self->{cf};
> }
>
This is wonderful. When you see more such examples, please share with us.
I think we can all learn from these!
> Anyway, later I've ran into this loop:
>
> # process continuation lines, if any
> while($self->{allowcontinue} && $val =~ s/\\$//) {
> $line = $self->_read_next_line($fh);
> $val .= $line;
> }
>
> Now, $self->{allowcontinue} is not modified within _read_next_line() (or
> anywhere else except for initialisation) so in case it is true, it is
> evaluated every time the condition is processed (not to mention, it gave me a
> "WTF?" moment).
This is a tricky one. There might have been some earlier version where the
$self->{allowcontinue} was changed during the call and this might be a
left over or
maybe this is was a way to avoid another level of indentation.
I know I wrote such code and I am sure I still do even though this is
probably some
misguided optimization of the code.
Gabor
_______________________________________________
Perl mailing list
[email protected]
http://mail.perl.org.il/mailman/listinfo/perl
