Ruby Regex optimization

Question

Regex newbie here. I have a regular expression that matches Windows pathnames and UNC paths, terminated by '\'.

Working examples:

c:\windows\
c:\
\\server\share\
\\server\sh are\

Invalid:

c:\windows
\\server
\\server\share
\\server\ share \

However, it works as expected (at least i hope so), but it's pretty unreadable and not very performant, so any tips for optimization are greatly appreciated...

/\A(
  ([a-z]:\\(([a-zA-Z0-9äöüÄÖÜß_.$]+|[a-zA-Z0-9äöüÄÖÜß_.$]+[a-zA-Z0-9äöüÄÖÜß_.$\ ]*[a-zA-Z0-9äöüÄÖÜß_.$]+)\\)*)|
  (\\\\(([a-zA-Z0-9äöüÄÖÜß_.$]+|[a-zA-Z0-9äöüÄÖÜß_.$]+[a-zA-Z0-9äöüÄÖÜß_.$\ ]*[a-zA-Z0-9äöüÄÖÜß_.$]+)\\)+(([a-zA-Z0-9äöüÄÖÜß_.$]+|
  [a-zA-Z0-9äöüÄÖÜß_.$]+[a-zA-Z0-9äöüÄÖÜß_.$\ ]*[a-zA-Z0-9äöüÄÖÜß_.$]+)\\)+)
)\z/

You're missing out on an enormous number of valid filenames (there are lots of other characters allowed) - isn't that a problem? And you're allowing many invalid filenames (for example con.txt). What exactly are you planning to do? Checking for validity? — Tim Pietzcker
– Tim Pietzcker, Commented Sep 21, 2011 at 7:58
Sorry for the lack of information on this, i'm using Ruby 1.9 any you are right, i'm trying to check for validity... — tobmatth
– tobmatth, Commented Sep 21, 2011 at 9:03

Tim Pietzcker · Accepted Answer · 2011-09-21 10:28:40Z

4

In Ruby 1.9, the following should work:

if subject =~ 
    /\A(?:(?!.*\\(?:con|prn|aux|nul|com\d|lpt\d)\\)  # exclude invalid names
    (?:                                              # Either match        
     [a-z]:\\                                        # drive letter 
    |                                                # or
     \\\\(?:[^\\\/:*?"<>|\s]+\\){2}                  # UNC share name
    )                                                # End of alternation
    (?:                                              # Try to match:
     (?!\s)                                          # (Assert no starting space)
     [^\\\/:*?"<>|\r\n]+                             # a valid directory name
     (?<!\s)                                         # (Assert no ending space)
     \\                                              # backslash
    )*                                               # repeat as needed
    )\Z/mix
    # Successful match
else
    # Match attempt failed
end

edited Sep 21, 2011 at 10:28

answered Sep 21, 2011 at 8:26

Tim Pietzcker

337k59 gold badges520 silver badges572 bronze badges

Sign up to request clarification or add additional context in comments.

5 Comments

tobmatth Over a year ago

This allows UNC paths without share names (\\server\), but i'll get this to work and it looks way cleaner than my regzilla, thanks...

Tim Pietzcker Over a year ago

Ah, ok, shouldn't they be allowed? No problem.

tobmatth Over a year ago

Hmm, this still allows \\server\. Isn't something like (pseudocode): if unc share at least one valid directory name required needed?

Tim Pietzcker Over a year ago

You're right, sorry. Now it shouldn't do this any more. Spaces are not permitted in server and share names, is that correct?

tobmatth Over a year ago

I'm not sure about that, afair they are permitted for share names, but spaces in servernames are filled with '-', e.g. '\\ser-ver\my share\', i'll have to look this up. However, thanks again for your answer...

Collectives™ on Stack Overflow

Ruby Regex optimization

1 Answer 1

5 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

5 Comments

Your Answer

Sign up or log in

Post as a guest

Related