[asterisk-dev] zero-or-more match

[Jason Parker]

Nick Lewis wrote:
>> Well, the "easy" way would be if there was a special character which
>> repeated the previous pattern zero or more times. Like * in common
>> regular expressions.
> 
> I do not see this as a substitute for real dialplan security but it
> would be a useful new feature. 
> 
> To repeat a "previous pattern" may be too complicated and confusing but
> to repeat a "previous character" would be useful. The % character is
> little used in extensions so could potentially be reassigned as dialplan
> punctuation for the purpose of repeating the previous character. e.g.
> 
> "3%" matches "", "3", "33" etc
> "[a-z]%" matches "", "b", "yx" etc
> "X%" matches "", "0", "98" etc
> "3\%" matches "3%"
> 
> Another useful feature would be a non-repeating "." that wildcard
> matches just one character. The ? character may be suitable for this
> where "?%" is equivalent to "!" and "??%" is equivalent to "." . I
> currently use "[!-~]" extensively for this purpose but it does not deal
> with whitespace and unprintable ascii.
> 
> -- N_L

exten => /realregexhere/,1,Dial(SIP/${EXTEN})

Sadly, that doesn't actually work...


If pattern matching is going to be extended for something like you're 
suggesting, I see no reason to not just go all the way with it.  This isn't the 
first time we've seen requests to add "just a little new feature" to mimic regex.

Proper regex support (ie; man 3 regex) would satisfy many of these types of 
complaints.  It may be difficult to fit in with the current pattern matcher (I 
honestly don't know - I've not looked at it in much depth), but I have no doubt 
that a number of the people on this list would be able to accomplish this.

Obviously there are some questions that would need to be answered (for instance, 
I realize that my pseudo-syntax above could conflict with callerid matching) 
before much time is spent on it, but is this something that would make sense? 
Is it actually worth spending time on?