I don't think defined() is gonna work. That will be true if $foo has any value and will short-circuit the regex....
What's this about a "crooked mitre"? I'm good at woodwork!
| [reply] [d/l] |
I think defined() is a good idea if there's a possibility of $foo being undefined, so as to avoid a warning about an undef being used in a string comparison/regex. Assuming an undefined value counts as a blank, I would write it like this:
if (!defined($foo) or $foo =~ m/^\??$/) {
# Foo is valid
}
Thus: "If foo is undefined, or matches an empty string or a single question mark."
But wasn't there something about \n's counting as "^" and/or "$"? Hmmm...
The Secret to Fortune Cookies in One Line
print join("... in bed", `fortune fortunes` =~ m/^(.*)(\.|\?|\!)$/), "\n"; | [reply] [d/l] [select] |
I wasn't saying that defined() doesn't have important uses, only that it use as given in the referenced post was not useful.
That said, I wouldn't use a regex as you have either. You point out one reason yourself, though m//s would probably address that concern, but, as others pointed out earlier, using a regex to test a string against a single, predefined char (or word)--especially when that char is a non-word char, meaning that the m//i option would be of no benefit--is simply overkill.
Personally, I would probably code the test as:
if ( !defined($foo) or $foo eq '' or $foo eq '?' ) {
print "\$foo OK!\n";
}
as using a regex against a fixed string seems pointless and I prefer positive conditions to negative ones.
Update:Modified condition, struck irrelevant comments. Seems I lost track of the original questioners reqs. Personally, I probably still wouldn't use a regex for this.
What's this about a "crooked mitre"? I'm good at woodwork! | [reply] [d/l] [select] |