PHP string pattern

I've come across the following string pattern in some PHP code and am trying to understand exactly what it means. I know it's used in an ereg() which should be replaced by preg_match(), that's another job.

PHP
1	ereg("^[A-Za-z0-9'\r\n\/,\.#()!?�$*+= -]+$",$value)

As far as I can see that's checking to see if any (1 or more) of the specified characters exist in $value. The characters being checked for being

any upper or lower case letter
any digit
' [apostrophe]
/ [forward slash]
, [comma]
. [period or full stop]
# [hash]
( [open round bracket]
) [close round bracket]
! [exclamation mark or shriek]
? [question mark]
£ [pound sign]
$ [dollar sign]
* [asterisk]
+ [plus sign]
= [equals sign]
[space]
- [minus sign or hyphen]

I have no idea what the � is there for, perhaps it's some sort of typo. (It's immediately before the apostrophe but has been removed by the forum code.)

Also I cannot understand the use of escape r and escape n.

Have I understood the pattern correctly?

Any thoughts on the � and the escape r and n ?

Edited by cheshire_man (Thu 09-Sep-10 16:03:19)

\r\n are carriage return, linefeed

Your understanding is correct.

For preg_match, you will need:

Text
1	preg_match("/^[A-Za-z0-9'\r\n\/,\.#()!?�$*+= -]+$/",$value)

(the string contains a regexp including the // delimiters)

Edited by micksharpe (Thu 09-Sep-10 16:18:57)

Do you mean greek beta, or german double s (not that I know what either of them would be in this context)

I was aware of that in the context of generating HTML from PHP. But you've made me wonder - does that mean that, for example, if the ereg() was being used to check data input in an HTML form and the user had used the return/enter key \r\n would need to be permitted as they'd be in the data that comes in from the <textarea> ?

German double s - actually it's hex DF (dec.223) which I think is the German double s

I don't think you have accounted for the initial caret, which is saying that the character being matched must be the first one in $value, or the "+$" at the end.

+ matches any single character and $ matches the end of the string. Hence (I think) the pattern is looking to match $value consisting of one of the characters that you mentioned (or CR or LF) followed by any one character. So just a two-character string.

I'm just a little bit worried about the double quotes and the effect this would have on, in particular, the two $ characters. It's more usual to see regular expressions like this in single quotes.

Take your point (I think) about the start ^ and the end $ although does the use of the [ ] brackets have an effect?

I understood the + to mean one or more of the literal characters enclosed in the [ ] brackets.

The double quotes are needed - aren't they? - as there is a single quote inside them, changing the double quotes to single quotes would cause a syntax problem, wouldn't it?

Sorry - you are correct about the plus sign. So I think that $value must be just a single character - one from the set in [].

I'm still a little confused about the fact the some characters that I would expect to be escaped (for example the first $ and the ') aren't, but that may be because I am more used to Perl than PHP. Perhaps the fact that these characters are inside the [] has some significance here. I always need several goes at constructing a regular expression before I arrive at what I intend! A little experimentation might make things clearer in this case.

In reply to a post by cheshire_man:

Take your point (I think) about the start ^ and the end $ although does the use of the [ ] brackets have an effect?
I understood the + to mean one or more of the literal characters enclosed in the [ ] brackets.

The ^ and $ are from the beginning and to the end, so you are matching one or more (the +) of the characters in square brackets

The double quotes are needed - aren't they? - as there is a single quote inside them, changing the double quotes to single quotes would cause a syntax problem, wouldn't it?

yes, they are needed