[asterisk-dev] [design] Matching algorithm

[Leif Madsen]

Tony Mountifield wrote:
>> My opinion is that treating _x and _X is probably an accidental 
>> oversight, and the code should be updated to treat them the same.
> 
> I originally got bitten by not realizing that lower-case n, x and z were
> special (all the examples in extensions.conf use upper-case). I would use
> lower-case alpha prefixes to extension numbers for certain kinds of dialplan
> processing, and the found this didn't do at all what I expected it to:
> 
> exten => _main-X.,1,Set(CONF=${EXTEN:5})
> 
> So my preference would be for only upper case N, X and Z to be special,
> and lower case to be literal characters. I don't know how many Asterisk
> users do _nxxnxxxxxx instead of _NXXNXXXXXX

I originally agreed that 'n', 'x', and 'z' should be treated exactly the 
same as 'N', 'X', and 'Z', but I can certainly see the argument about 
using the lowercase as literal characters. I've been bitten on this 
before until Jared told me that I was using a 'reserved' character in my 
string. Using your example, you'd have to do:

exten => _mai[n]-X.,1,Set(CONF=${EXTEN:5})

in order to get what you expected.

I always find it weird to read through dialplans that use lowercase 
characters for the pattern match, but have only seen it VERY rarely (and 
I don't think they were ever a production system, but rather were a test 
system someone was having me help them with -- which of course doesn't 
mean they don't exist).

--
Leif Madsen
http://www.leifmadsen.com
http://www.oreilly.com/catalog/asterisk