multiple results for regex

Question

question

dclaar Suspended asked Dec 10, '09 dclaar Suspended posted Dec 10, '09

multiple results for regex

I have a simple analysis rule:

analyze regex .* (line mode not checked, full match, declare issue if no matches found)

assert $value eq ""

if true, fail

if false, OK

When I run this, I get two assertions: One with all of the text from the command, and one with nothing in it. So I pass and fail the analysis.

This is just wrong. If I ask for everything, I should get everything. And isn't that what .* means?

I'm really just trying to see if I got anything back, so I'm open to some other way of doing that. But this still looks like a bug!

iTest bug

10 |950

Attachments: Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

Answer 1 · 2009-12-10T00:44:41Z

dclaar Suspended answered Dec 10, '09 dclaar Suspended posted Dec 10, '09

Setting the regex to ^.*$ causes it to only return one response. Still calling bug. :-)

Oh, 3.4.1.

10 |950

Attachments: Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

Answer 2 · 2009-12-10T01:22:18Z

YujieL answered Dec 10, '09 YujieL posted Dec 10, '09

it sounds like a bug...what is the response/text that you're doing the analysis on? I'd like to try it myself and see what the analysis returns.

3

10 |950

Attachments: Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

dclaar commented · Dec 10, 2009 at 01:25 AM

Attaching test case. Really just call foo

foo

write this

write is

write a

write test

Of course, in real life, it was the output of a command, but the above/attached will reproduce it.

0 ·

unregex.fftc (3.9 KiB)

YujieL dclaar commented · Dec 10, 2009 at 02:14 AM

I see the same behavior, but this doesn't look like a bug. Some regex testers online will show you the number of matches, and it will find two matches, one which is the entire string, one which is an empty string. I'm no expert on how regex engine works, but it's quite possible that the engine starts matching at position 0, and it finds "thisisatest", then it tries to find a second match starting at position 11, which is just an empty string. Since .* matches anything, it would match the empty string.

0 ·

dclaar YujieL commented · Dec 10, 2009 at 02:18 AM

Wow, I guess you're right. Even perl does it that way:

perl -le '$a="this is a test";$a=~s/.*/a/g;print $a';
aa

Learn something new every day! I guess in perl, I've always anchored the expression with ^ and $.

0 ·

question