Question

I’ve been struggling with getting a working regular expression for a little project of mine. Can anyone help me with a regex that matches anything inside <> symbols, but only when they are not preceded by a \ symbol?

For example:

<Escaped characters \<\> are right in the middle of this sentence.>, <Here is another sentence.>

Must match

1: Square brackets \<\> are right in the middle of this sentence.
2: here is another sentence.

So far I’ve managed

/<([^\\][^>]*?)>/ig

but that gives

1: Escaped characters \<\
2: Here is another sentence.

What am I doing wrong? 🙁

Answer 1

Crimson’s answer is not working for me in testing it in the Regex Powertoy using <Escaped characters \<\> are right in the middle of this sentence.>, <Here is another sentence.> as the test but this (seems) to work:

/<(?<!\\<).*?>(?<!\\>)/gi

Gives me two matches:
<Escaped characters \<\> are right in the middle of this sentence.> and <Here is another sentence.>

Edit: I took a look at the string Gumbo said did not match. I don’t have any problems matching it in regex.powertoy.org:

alt text http://img362.imageshack.us/img362/3227/regexpowertoyorg.png

In testing I did change the original posted regex to: /(?<!\\)<(.*?)(?<!\\)>/gi which is more efficient (less probes).

Also I notice in the output of regex.powertoy.org that the forth string (\<hello <match\<this\>> but not this\> looks odd... the printed replacement is justmatchbut the match detail clearly shows that the match is correct;match\. But I also notices that the first and third test string replacements don't print the "`” escaping the angle brackets. After a bit (but not exhaustive) playing around I think that this is an issue with the display of the text via javascript, the escaped angle brackets don’t print the escape char, and non-empty angle brackets don’t get printed at all. I think this is due to the javascript seeing it as HTML. So; I think this regex is working correctly. But you should test it offline.