[asterisk-users] escaping regular expression

[Klaus Darilion]

D Tucny schrieb:
> 2009/2/4 Klaus Darilion <klaus.mailinglists at pernau.at 
> <mailto:klaus.mailinglists at pernau.at>>
> 
> 
> 
>     D Tucny schrieb:
>      > 2009/2/4 Klaus Darilion <klaus.mailinglists at pernau.at
>     <mailto:klaus.mailinglists at pernau.at>
>      > <mailto:klaus.mailinglists at pernau.at
>     <mailto:klaus.mailinglists at pernau.at>>>
>      >
>      >     Hi!
>      >
>      >     I am going nuts using REGEXP. I just want to verify if a variable
>      >     contains a valid +E164 phone number.
>      >
>      >     These, the the pattern is ^\+[0-9]+
>      >
>      >     First I tried:
>      >
>      >       Set(pattern=^\+[0-9]+);
>      >       if (${REGEX("${pattern}" ${${var}})})
>      >
>      >     but that does not work, the backslash is removed, as seen in
>     the log
>      >     file:
>      >
>      >       func_strings.c: FUNCTION REGEX (^+[0-9]+)(+4312345
>      >     <http://www.adaanumber.com/>)
>      >
>      >     So, meanwhile I tried to escape the backslash. I tried:
>      >       Set(pattern=^\\+[0-9]+);
>      >       Set(pattern=^\\\+[0-9]+);
>      >       Set(pattern=^\\\\+[0-9]+);
>      >
>      >     But always the same result:
>      >
>      >       func_strings.c: FUNCTION REGEX (^+[0-9]+)(+4312345)
>      >
>      >     How can I solve this problem?
>      >
>      >
>      > Try something like... pattern=^[+]\{1\}[0-9]+
> 
>     Are you sure? The \ should be in front of the +
> 
> 
> Pretty sure...
> 
> exten => *56,1,NoOp("Starting regexp test")
> exten => *56,n,Set(pattern=^[+]\{1\}[0-9]+)
> exten => *56,n,Set(var=123456789)
> exten => *56,n,NoOp(${IF(${REGEX("${pattern}" ${var})}?"Match":"No 
> Match")}))
> exten => *56,n,Set(var=+123456789)
> exten => *56,n,NoOp(${IF(${REGEX("${pattern}" ${var})}?"Match":"No 
> Match")}))
> 
> 
> [Feb  4 23:49:21] DEBUG[20518] pbx.c: Launching 'NoOp'
> [Feb  4 23:49:21] VERBOSE[20518] logger.c:     -- Executing 
> [*56 at phonedefault:1] NoOp("SIP/*01-09bd8ff8", ""Starting regexp test"") 
> in new stack
> [Feb  4 23:49:21] DEBUG[20518] pbx.c: Launching 'Set'
> [Feb  4 23:49:21] VERBOSE[20518] logger.c:     -- Executing 
> [*56 at phonedefault:2] Set("SIP/*01-09bd8ff8", "pattern=^[+]\{1\}[0-9]+") 
> in new stack
> [Feb  4 23:49:21] DEBUG[20518] pbx.c: Launching 'Set'
> [Feb  4 23:49:21] VERBOSE[20518] logger.c:     -- Executing 
> [*56 at phonedefault:3] Set("SIP/*01-09bd8ff8", "var=123456789") in new stack
> [Feb  4 23:49:21] DEBUG[20518] func_strings.c: FUNCTION REGEX 
> (^[+]{1}[0-9]+)(123456789)
> [Feb  4 23:49:21] DEBUG[20518] pbx.c: Function result is '0'
> [Feb  4 23:49:21] DEBUG[20518] pbx.c: Function result is '"No Match"'
> [Feb  4 23:49:21] DEBUG[20518] pbx.c: Launching 'NoOp'
> [Feb  4 23:49:21] VERBOSE[20518] logger.c:     -- Executing 
> [*56 at phonedefault:4] NoOp("SIP/*01-09bd8ff8", ""No Match")") in new stack
> [Feb  4 23:49:21] DEBUG[20518] pbx.c: Launching 'Set'
> [Feb  4 23:49:21] VERBOSE[20518] logger.c:     -- Executing 
> [*56 at phonedefault:5] Set("SIP/*01-09bd8ff8", "var=+123456789") in new stack
> [Feb  4 23:49:21] DEBUG[20518] func_strings.c: FUNCTION REGEX 
> (^[+]{1}[0-9]+)(+123456789)
> [Feb  4 23:49:21] DEBUG[20518] pbx.c: Function result is '1'
> [Feb  4 23:49:21] DEBUG[20518] pbx.c: Function result is 'Match'
> [Feb  4 23:49:21] DEBUG[20518] pbx.c: Launching 'NoOp'
> [Feb  4 23:49:21] VERBOSE[20518] logger.c:     -- Executing 
> [*56 at phonedefault:6] NoOp("SIP/*01-09bd8ff8", "Match)") in new stack
> 
> So, the \ is still stripped (ast_app_separate_args removes \), but, it 
> doesn't matter as the + is bracketed so it's not the first character 
> after the ^ and so regcomp doesn't fail...


Ah, the trick is to put the + into [], so it need not be escaped.

Thanks
Klaus