.NET string split method strange behavior

I have a problem with string.split using.NET 3.5:

String to split is:

dim source as string = "ab|foo|bar|bar|bar-foo|ab|ezrezertr|ghghhjhj|ab|foo|xxx|"
dim result() as string = source.split("ab|")

When used within a Winforms applicaton, the result is "correct":

result(0) is "foo|bar|bar|bar-foo|"

result(1) is "ezrezertr|ghghhjhj|"

result(2) is "foo|xxx|"

And I'm happy!

When used within an ASP.NET code behind, the result is:

result(0) is "b|foo|bar|bar|bar-foo|"

result(1) is "b|ezrezertr|ghghhjhj|"

result(2) is "b|foo|xxx|"

In other words, the split function only get rid of the 1st character of the separator string! Does someone kno开发者_如何转开发w why?

The ASP.NET results look like you are using Regex.Split rather than String.Split. The string "ab|" will be interpreted as a regular expression for "a" followed by "b" or nothing, so just an "a" matches.

Later: Second Theory:

• There is no overload of String.Split that takes a single string argument. The only single argument overload takes an array of char.
• String.Splt(char()) will split on any of the passed characters.
• VB, without Option Strict On will implicitly convert a string to an array of chars.

Hence I think in the ASP.NET case you don't have option strict on, therefore theString.Split(anotherString) is being treated as theString.Split(anotherString.ToCharArray()).

Thus splitting on just a "b".

However this leaves the question of how the first cases acts as passing a string, but there is no overload taking a String() without extra parameters (a SplitOptions)...

Summary: Visual Basic's extra implicit conversions and behaviour set at a file/project/language level can make identical code behave differently.

If the input string is the same, then I can nearly guarantee you that the split method of a Console app and ASP.NET behave the same.

Maybe the output is not displayed correctly?