Re: ctype second opinions

On Fri, 19 Jul 2002, Brad Bulger wrote: > > > PS - fwiw, i think you're right. ereg('[[:alpha]]','') > > > returns false. > > > ctype_alpha('') should do the same thing. > > > > Thanks, but I can't continue now that I proved otherwise (unless someone > > can dispute my proof) :P > > you did? did i miss that? i saw that it got made moot by finding out that > you're not supposed to walk through strings like an array... > I had sent to Pierre how I came to the conclusion that the current behavior is correct, but forgot to CC the list. He probably thinks I'm an alien now. Given: An alpha string is defined a string in which every character is a letter. Assertion: An empty string is an alpha string. Proof: Take two strings, $a, which is a non-empty alpha string and $b, which is an empty string: a = 'Hello' // non-empty alpha string b = '' // empty string Assume that b is not an alpha string. Consider that when strings c and d are appended together, the resulting string, cd, contains the characters of c and d. If c is an alpha string and d is not, the cd is cannot be an alpha string because cd contains characters from d that are not letters. Now, consider e e = ab e contains every character from a and b. By the previous example, if b is not an alpha string, then e is not an alpha string, since e would then contain characters that are not letters. But e is an alpha string because every character in e is a letter. Therefore, b must be an alpha string. QED Really, I made too much out of this; I found that ctype_alpha(ord($c)) does what I wanted in the first place :P Brent