On Saturday 12 November 2011, 14:23:30, Shakthi Kannan wrote:

Hi,

--- On Sat, Nov 12, 2011 at 5:44 PM, Daniel Fischer

wrote: | Maybe the backreferences numbering starts at 0?

Not backreferences, but who cares?

| Worth a try.

\--

\0 represents the entire string match:

The entire *match*, that is, the part of the input matched by the regexp. The other entries correspond to parts matched by certain subregexen in the match.

http://cvs.haskell.org/Hugs/pages/libraries/base/Text-Regex.html

May I suggest using the docs at hackage, hugs hasn't had a release since 2006, I don't think the docs are up to date. Unless you're actually using hugs, in which case I suggest switching to ghc. http://hackage.haskell.org/package/regex-compat

Prelude Text.Regex> subRegex (mkRegex "e") "hello" "\\0" "hello"

Heh, I didn't see it immediately either ;) Prelude Text.Regex> subRegex (mkRegex "e") "hello" ">\\0<" "h>e subRegex (mkRegex "https?[^\\s\n\r]+") "The best is http://haskell.org" "there</a>" "The best is http://ha\">there</a>skell.org" Not what you want, character classes don't work that way, Prelude Text.Regex> subRegex (mkRegex "https?[^[:space:]]+") "The best is http://haskell.org\n" "there</a>" "The best is http://haskell.org\">there</a>\n" but that. However, Prelude Text.Regex> subRegex (mkRegex "https?[^[:space:]]+") "The best is http://haskell.org." "there</a>" "The best is http://haskell.org.\">there</a>" please make an effort to not include final punctuation in the href, it's rather annoying how many 404s I get from that.

On Saturday 12 November 2011, 16:10:58, Shakthi Kannan wrote:

Hi,

--- On Sat, Nov 12, 2011 at 8:20 PM, Daniel Fischer

wrote: | Prelude Text.Regex> subRegex (mkRegex "https?[^[:space:]]+") "The best | is http://haskell.org\n" "there</a>" | "The best is http://haskell.org\">there</a>\n"

\--

Is there any way we can escape the " within <a href></a>, and not get the \" in the output?

The \" are just because that's Haskell's way of showing Strings. Strings are shown enclosed in quotes, and quotes (and other characters) appearing in the String are escaped. You can see what would get written to the file by outputting the String with putStrLn, Prelude Text.Regex> putStrLn $ subRegex (mkRegex "https?://([:alpha:]| [0-9]|.[[:alpha:]0-9])+") "The best is http://haskell.org." "there</a>" The best is <a href="http://haskell.org">there</a>. See, just like it should be.

I am producing HTML in the output, and the \" doesn't work with the anchor tag. Expected output is:

"The best is <a href="http://haskell.org">there</a>\n"

Thanks for your prompt help!

SK