Need to build regex to find specific hex characters

mphare · July 27th, 2009, 02:52 PM

Hi,

I have a statement: <xsl:analyze-string select="$cmdSyntax" regex="^([A-Z0-9]+)-.*$">

that I need to adjust.

Rather than just looking for the dash after the alpha numerics, I need to search for a specific string of hex characters (xE2, x80 then x91). They will all three be there for the targets I'm looking for.

I'm not seeing how to build a regex that locates specific hex values.

Can you help?

Thanks,

- mike

samjudson · July 27th, 2009, 03:19 PM

I think hex characters are referred to as "&#xE2" etc. Try that see if it works.

mphare · July 27th, 2009, 03:43 PM

Sounds logical, but doesn't seem to work.

My expression became: <xsl:analyze-string select="." regex="^([A-Z0-9]+)â.*$">, but I'm still not getting a hit.

I did notice that the expression \xnn where nn is a hex value is strictly disallowed by the definition and the interpreter should reject the use of such a construct.

Too bad, \xE2 is just what I was looking for.

Appreciate the effort though.

mhkay · July 27th, 2009, 04:23 PM

Could you show a sample of your XML?

I'm confused as to whether it contains something like

Code:

<element>E2</element>

or something like

Code:

<element>&#xE2;</element>

In the first case the regex to match it is 'E2', in the second case it is 'â'

mphare · July 27th, 2009, 04:37 PM

Here's the XML:

Code:

<Syntax>SET‑ATTR‑EQPT:[&lt;TID>]:&lt;AID>:&lt;CTAG>::[&lt;NTFCNCDE>],&lt;CONDTYPE>,[&lt;LOCN>],[&lt;DIRN>],,[&lt;SRVEFF>];</Syntax>

In this example, the dashes between SET and ATTR and between ATTR and EQPT are not real dashes (0x2D), but are a secret code embedded when Adobe Framemaker exported an Escaped-Hyphen-H character sequence into XML. ESC-Hyphen-H is something Framemaker uses to create hyphens that don't cause line wraps. In a long line, Framemaker will attempt to line wrap at a dash if one exists in the line. So people use ESC-Hyphen-H to inform Framemaker to wrap at the 80th character, not arbitrily at a dash. The secret code is 0xE2 0x80 0x91, at least that what my hex editor shows the file to contain.

Thanks,

- mike

mhkay · July 27th, 2009, 04:53 PM

I suspect your hex editor is showing you the UTF-8 encoding of the Unicode character. I suspect the real Unicode character is a non-breaking-hyphen, code x2011 - I haven't actually checked that E28091 is the UTF-8 encoding of x2011, but I suspect it is. If that's the case then the regex you need is '‑'

mphare · July 27th, 2009, 05:03 PM

That was it. ‑ works as expected.

Thank you!