<some code>Yes, it's better but it still breaks. Consider e.g. giving the modifiers in a different order (/mi instead of your /im), giving multiple modifiers (/ii is perfectly legal) using different delimiters (m!pattern!) and so on.
Better?
I'm not saying that it is impossible to parse this regex construct by hand - but why should want to do that? The eval solution always works - if $RegEx is a legal perl pattern matching operator and you can pass out anything with the construct I explained elsewhere in this thread.
Your solution restricts the number of possible patterns and the ways of providing them. The qr solution suffers from the same problems, only it is less restrictive. And the eval is slow but gives you the whole range of possible pattern matching constructs (even s/// would work) and more (which can become a problem/security concern).
To sum up, without knowing the problem better (where does the $RegEx come from, what should be achieved with it, ...) we cannot tell which solution to favour. And as always, TMTOWTDI :)
-- Hofmator
In reply to Re4: Applying a RegEx in a variable
by Hofmator
in thread Applying a RegEx in a variable
by ChOas
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |