ansaurus

Question

Answer 1

+2 A:

This is not possible using RegEx. HTML is not a regular language, so incomplete pages cannot be verified by RegEx.

See here for why parsing HTML with RegEx is a bad idea.

Use a validating HTML parser for your platform to load the HTML and check for validation errors.

You need to figure out if a file is HTML or not before you try to detect whether it is incomplete HTML. You may be able to do this using the file extension, if available, however you did not provide enough information about your environment for any recommendation to be made.

It is easy enough to find whether <html and </html> exist in a string and if not return a "special string". Normal string functions should suffice - no need for RegEx.

Oded 2010-09-24 09:46:08

I'm not asking for much. I'm not parsing HTML. All I need is a regex that would insert a bit of text if there is an <html> string, but no </html> string.

Ed 2010-09-24 09:49:49

@Ed - then please _edit_ your question and clarify it.

Oded 2010-09-24 09:50:51

If I could, I would use something else, but for this problem I can only use a search and replace regex.

Ed 2010-09-24 10:15:46

@Ed - again, please _edit_ your question and _add_ these bits of information to it. It will help anyone reading your question to answer it better.

Oded 2010-09-24 10:17:24

Answer 2

+3 A:

Oded is correct. You cannot parse HTML with regex. But of course you can see whether some (multiline) string contains <html> not followed by </html>. If you are sure that whatever your web request returns will be consistent and not contain any weird things like html tags inside comments, then

<html\b[^>]*>(?:(?!<\s*/\s*html).)*\Z

will find such a string, if you set the "dot matches newlines" option. How to do this depends on the regex implementation which you didn't provide yet.

<html\b[^>]*>          # match <html> tag
(?:                    # match the following:
  (?!<\s*/\s*html)     # If it's impossible to match </html here
  .                    # then match any character
)*                     # zero or more times.
\Z                     # Then assert that we are indeed at the end of the string

Tim Pietzcker 2010-09-24 09:54:14

See [Bart K.’s comment](http://stackoverflow.com/questions/3785861/regex-to-detect-incomplete-html#comment-4010395).

Gumbo 2010-09-24 10:05:36

@Gumbo: Of course I did. I do know the limitations (and have mentioned them). But in his case it seems workable since he can exclude cases that would trip this regex up.

Tim Pietzcker 2010-09-24 10:21:00

This hit the spot! Thank you.

Ed 2010-09-24 10:26:24

Answer 3

A:

In general, regular expressions are not suitable to describe HTML because regular expressions can only describe regular languages but HTML is not a regular language. The fact that HTML allows to nest elements in an arbitrary depth makes it irregular.

Although there are regular expression implementations that support recursive patterns (e.g. PCRE, .NET, et al.), it would require a horrible complex regular expression to describe every valid HTML code.

Gumbo 2010-09-24 10:01:11

ansaurus

tags:

views:

answers:

Regex to detect incomplete HTML

related questions