[SOLVED] PCRE Validation

Linux_Kidd · 06-24-2012, 09:00 AM

just looking for validation that my conversion to PCRE is correct.

i converted spaces to \s+
and i converted #'s to \d and where there were two digits or more i used \d{} syntax

Code:

Mozilla/5.0 (Macintosh; Intel Mac OS X 10.7; rv:10.0.1) Gecko/20100101 Firefox/10.0.1

PCRE
Mozilla/\d\.\d\s+\(Macintosh;\s+Intel\s+Mac\s+OS\s+X\s+\d{1,2}\.\d;\s+rv:\d{1,2}\.\d\.\d\)\s+Gecko/\d{1,8}\s+Firefox/\d{1,2}\.\d\.\d

does this look correct?

acid_kewpie · 06-25-2012, 02:35 AM

http://www.regextester.com/

Linux_Kidd · 06-25-2012, 05:19 AM

Quote:

Originally Posted by acid_kewpie

http://www.regextester.com/

nice site, how does it work? i put in my regex and raw text, then what?

Diantre · 06-26-2012, 12:37 AM

Quote:

Originally Posted by Linux_Kidd

just looking for validation that my conversion to PCRE is correct.

If you have the PCRE package installed, you can use the pcretest program to test your regular expressions. Have a look at its manual page for more information.

Basically, this program asks for your regex with a re> prompt, then asks for the text with a data> prompt and reports if there's a match or not.

Your regular expression is fine, just had to escape the forward slashes to use them as delimiters (or you can leave the slashes as they are and use another character for delimeter, such as #):

Code:

PCRE version 8.12 2011-01-15

  re> /Mozilla\/\d\.\d\s+\(Macintosh;\s+Intel\s+Mac\s+OS\s+X\s+\d{1,2}\.\d;\s+rv:\d{1,2}\.\d\.\d\)\s+Gecko\/\d{1,8}\s+Firefox\/\d{1,2}\.\d\.\d/
data> Mozilla/5.0 (Macintosh; Intel Mac OS X 10.7; rv:10.0.1) Gecko/20100101 Firefox/10.0.1
 0: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.7; rv:10.0.1) Gecko/20100101 Firefox/10.0.1
data> something else
No match
data>

acid_kewpie · 06-26-2012, 06:25 AM

Quote:

Originally Posted by Linux_Kidd

nice site, how does it work? i put in my regex and raw text, then what?

then you read the "result" line. Not exactly complex.

Linux_Kidd · 06-26-2012, 09:43 PM

ok, the results just say whatever i type in for text, no mater what my RE looks like, hence i dont get the tool.

i am not asking to validate if my RE will match my sample text. i am asking for RE validation in syntax because i expanded the RE to catch more stuff, etc.

Diantre · 06-28-2012, 01:11 AM

Quote:

Originally Posted by Linux_Kidd

i am asking for RE validation in syntax because i expanded the RE to catch more stuff, etc.

The pcretest tool validates regex syntax as well. If I type an incorrect regex it will complain, for instance:

Code:

PCRE version 8.12 2011-01-15

  re> /foo[bar/
Failed: missing terminating ] for character class at offset 7
  re>

The regex you provided is a valid PCRE regular expression, it works fine with pcretest. The only thing to be aware of is that forward slashes have to be escaped if this character is used as the regex delimiter.

Linux_Kidd · 06-28-2012, 05:29 AM

Quote:

Originally Posted by Diantre

The regex you provided is a valid PCRE regular expression, it works fine with pcretest. The only thing to be aware of is that forward slashes have to be escaped if this character is used as the regex delimiter.

so if want forward slash to be a pattern char i need \/
is that what you mean?

i see some PHP documentation that says forward slashes need escaping. so PCRE in general needs / escaped with \ if the / is to be a literal char

Diantre · 06-28-2012, 08:34 PM

Quote:

Originally Posted by Linux_Kidd

so if want forward slash to be a pattern char i need \/
is that what you mean?

i see some PHP documentation that says forward slashes need escaping. so PCRE in general needs / escaped with \ if the / is to be a literal char

Yes, that's the idea. However, PCRE allows the use of other characters as delimiters, with the exception of alphanumerics and the backslash. For example, if we have a regular expression to match directory names, such as /dir1/.*/dir2, the forward slashes need to be escaped if the regex delimiter is also a forward slash:

Code:

/\/dir1\/.*\/dir2/

In these cases it's convenient to use another character as the regex delimiter to have a more legible regular expression, provided the delimeter doesn't appear in the regex, otherwise it must be escaped as well. The regex above can be written with # as the regex delimeter:

Code:

#/dir1/.*/dir2#

It's not necessary to escape the slashes and the regex is much easier to read...