On Tue, Nov 21, 2017 at 08:46:59PM +0000, Alex Bennée wrote:
>
> Eric Sunshine <sunshine@xxxxxxxxxxxxxx> writes:
> > Aside from those observations, it looks like the tokenizer in this
> > function is broken. For any input with the address enclosed in "<" and
> > ">", the comment is lost entirely; it doesn't even end up in the
> > @tokens array. Since you're already fixing bugs/regressions in this
> > code, perhaps that's something you'd like to tackle as well in a
> > separate patch? ("No" is an acceptable answer, of course.)
> >
> >> } elsif ($token eq "<") {
> >> push @phrase, (splice @address), (splice @buffer);
> >> } elsif ($token eq ">") {
>
> I can have a go but my perl-fu has weakened somewhat since I stopped
> having to maintain perl code for a living. It's almost as though my
> brain was glad to dump the knowledge ;-)
>
> I guess we could maintain a nesting count and a current token type and
> use that to more intelligently direct the nested portions to the
> appropriate bits. Maybe Matthieu or Remi (CC'ed) might want to chime in
> on other options?
Trying to come up with a reinvention of regexps for email addresses is asking
for trouble, not to mention a crappy rod for your own back. Don't do that.
This is why people use Mail::Address.
https://metacpan.org/pod/distribution/MailTools/lib/Mail/Address.pod
-- Thomas Adam
