OpenEuphoria

1. regex observation & question regarding re:escape()

• Posted by jessedavis Sep 18, 2016
• 1621 views

Windows 7 pro sp1 euphoria: 4.0.5 (362497032f33, 2012-10-11)

I guess I am missing something here. I don't see what it is.

From the manual ==> 8.20.6.1 escape()

 Escape special regular expression characters that may be entered into a search string from user input. 
Notes: 
Special regex characters are:  
 
. \ + * ? [ ^ ] $ ( ) { } = ! < > | : -   
 
Example 1: 
 
include std/regex.e as re 
sequence search_s = re:escape("Payroll is $***15.00") 
-- search_s = "Payroll is \\$\\*\\*\\*15\\.00" 

What is actually produced by the above eucode is:

"Payroll is \$\*\*\*15\.00" which will not compile as a eu regex

another special character (from above list) is the backslash.
escape() cannot deal with it at all giving an unknown escaped character error..
So, use the alternative entry method offered in the docs: use #/.../ instead of "..." as an example: regex will not compile the following:
regex p = re:new(#/\*/)
from the docs:

The alternative writing style lets you write a regex without doubling up on the escape characters: 
 
regex sample = re:new( #/\*/ ) 

Giving the error message: hex number not formed correctly.
Using a double escape does still work.
Regular expressions are hairy enough without all the extra backslashes.
In any event, the docs should agree with the performance!

regards, jd

sequence p1 = re:Parse( "Payroll_is_ /$ /*/*/* 15.00") 

integer try = match( "Payroll is $***15.00", s) 

sequence p1 = re:Parse( "Payroll_is_ /$ [*]{0,4} /d+ . /d{2}") -- Alas, non-standard regex.. 

3. Re: regex observation & question regarding re:escape()

• Posted by jessedavis Sep 19, 2016
• 1580 views

The expression doesn't do much, true. These two regex expressions were taken from the docs directly as they were examples of how to do it. I was particularly interested in the second entry format that uses the #/.../ format because it has the potential to ease the confusion factor. I am building a small single line parser. I thought regex would be easier to implement since it recognizes patterns whereas the built in search routines recognize characters.
In any event, thanks for your help.

Regards,
jd

5. Re: regex observation & question regarding re:escape()

• Posted by _tom (admin) Sep 19, 2016
• 1543 views

My fault.

I can't find out how I got #/\*/ to work. It must have been some weird stuff I had on my computer.

I did have time to retry example 1

    include std/console.e 
 
include std/regex.e as re 
 
sequence s0 = "Payroll is $***15.00" 
display( s0 ) 
    --> Payroll is $***15.00 
    -- this is a valid Euphoria string 
    -- this IS NOT a valid regex string 
 
-- when s0 is used with the regex language 
    --> $ is a keyword 
    --> * is a keyword 
    --> . is a keyword 
 
    -- you must escape these characters before using them in a regex 
    -- then, you must escape the \ when writing a string 
 
sequence s1 = "Payroll is \\$\\*\\*\\*15\\.00" 
display( s1 ) 
    --> Payroll is \$\*\*\*15\.00 
    -- the form required when writing a regex 
 
 
-- writing a string with a ` backtik is a bit shorter 
 
sequence s2 = `Payroll is \$\*\*\*15\.00` 
display( s2) 
    --> Payroll is \$\*\*\*15\.00 
    -- the form required when writing a regex 
 
-- the Euphoria re:escape function adds the required backslashes 
-- to a sequence written with " delimiters 
sequence s3 = re:escape( s0 ) 
display( s3 ) 
    --> Payroll is \$\*\*\*15\.00 
    -- the form required when writing a regex 
 
--Note: 
-- the strings re:escape(s0), s1, and s2 are only useful as literal patterns 

thanks for pointing out my mistakes

_tom

7. Re: regex observation & question regarding re:escape()

• Posted by Spock Sep 20, 2016
• 1556 views

Just to clarify - 'anything' is referring to characters in a text string. Special characters treated as literals in the regex itself must still be escaped.

Spock

"foo\\.txt" 
 
-- could be written as: 
 
`foo\.txt` 

9. Re: regex observation & question regarding re:escape()

• Posted by jessedavis Sep 22, 2016
• 1445 views

Thanks tom for taking the time to demonstrate the proper usage. I appreciate your efforts. When I read the docs I took away an erroneous expectation about both escape() and the #/.../ form. I had completely forgotten the backtik form. Ya live & learn; and then, some of us just live!
I'm a long time user of euphoria - since 2.something. Now that I have retired from engineering I enjoy pushing ever deep into euphoria. You guys do a great job.

Thanks again,
jd

"foo\\.txt" 
 
-- could be written as: 
 
`foo\.txt`