Regex.IsMatch(expected, Regex.Escape(expected).Replace(@"5", @"\d+"))

The query contains regex characters which need to be escaped (i.e. '.').

It's a bit messy ... if this is going to be used a lot, would probably be good to replace the whitespace with '\s+', etc.

Michael: '.' will match any single character, so that will still work; it will just match more than expected.

I might guess that the newlines are confusing the regex engine.

But regex "/* " won't match string "/* ".

it's the /* */ comments...

Maybe Regex.Escape before you do your Replace might help...

It's a problem with the comment tags in the SQL statement. Remove the comments and the regular expression passes.

So, is Replace doing something funny to the asterisk or what?...

Aren't tests supposed to be readable? I have no idea what this is supposed to do from staring at the code. Might be just me though.

Regex.IsMatch(expected,Regex.Escape(expected).Replace("5", @"\d+"))

passes

There isn't anything wrong with the string, but that regex will never work, as it's going to change all the digits to 5. Should be this: @"\d+_(?=0)" That does a lookahead and only matches digits followed by an _ then a 0. The 0 is then not part of the grouping and so won't be changed.

It doesn't make sense, why would you want to do that anyway?

        const string EXPECTED = @"Cached query: 

SELECT this.Id as Id50_,

   this_.Title          as Title5_0_,

   this_.Subtitle       as Subtitle5_0_,

   this_.AllowsComments as AllowsCo4_5_0_,

   this_.CreatedAt      as CreatedAt5_0_

FROM Blogs this_

WHERE this_.Title = 'The lazy blog' /* @p0 */

   and this_.Id = 1 /* @p1 */";


        string pattern = Regex.Escape(EXPECTED).Replace("5_", @"\d+_");


        Assert.True(Regex.IsMatch(EXPECTED, pattern));

@Chris I was only providing one example of what would be escaped. As Sergey and bobo mentioned, the comment tags also cause an issue because '*' is a regex quantifier, and as such, is escaped by Regex.Escape as well.

Bertrand,

Well, when you only see a very small part, it is not surprising

Cool. That part does look convoluted :)

Why would we care about the regex when the entire principle this is based on is a big hairy WTF?

FYI - I know the multiline syntax is handy from a readability perspective, but tests like that may not pass on Mono/Linux or any other OS that doesn't use \r\n as the line terminator:

www.cornetdesign.com/.../...nvironmentnewline.html

To get it to pass, you'd have to recompile on the platform. Well, unless you are developing this on Mono/Linux, but I don't think that's the case. :)

Cory,

Yes, I know about that, but since this is a WPF app, and Mono doesn't have that, I don't worry too much about it

Don't know if you're still having issues here, but it looks like you need to escape the '' and '.' characters in the pattern. Also, you probably need to compact the whitespace in the pattern to as few as possible, then replace them with \s

Ooops, didn't see James' comment above... and it looks like it works.

The Regex.Escape() method does pretty much what I described above. From the docs: "Escapes a minimal set of characters (\, *, +, ?, |, {, [, (,), ^, $,., #, and white space) by replacing them with their escape codes."

That's handy.