Help with a regex that matches something either before OR after something else

Question

I have a bunch of XML that has lines that look like this

<_char font_name="/ITC Stone Serif Std Bold" italic="true" />

but sometimes look like this

<_char font_size="88175" italic="true" font_name="/ITC Stone Serif Std Bold" />

Here's what I need to do

• Replace italic="true" with italic="false for every line that contains ITC Stone Serif Std Bold, regardless of whether it comes before OR after the italic part.

Can this be done with a single regex?

I'm not looking for a real-time solution. I just have a ton of XML files that have this "mistake" in them and I'm trying to do a global search-and-replace with PowerGrep which would require a single regex. If scripting's the only way to do it, then so be it.

Answer 1

I'm not 100% sure, but I feel like something like XSLT might be better suited to this task, and depending upon what Regex engine you're using, I'm not sure that you can do this with a single regular expression.

Answer 2

Well, in general, using RE for XML parsing isn't a great idea. But if you really wanted, the easiest way would be to just do it in two lines:

if (/ITC Stone Serif Std Bold/) {
   s/italic="true"/italic="false"/g;
}

Answer 3

Does the simple use of '|' operator satisfy you ?

name="/ITC Stone Sans Std Bold"[^>]italic="(true)"|italic="(true)"[^>]font_name="/ITC Stone Serif Std Bold"

That should detect any line with the attribute name before of after attribute italic with value true.

Answer 4

In Perl - untested:

while (<>)
{
    s/italic="true"/italic="false"/ if m%font_name="/ITC Stone Sans Std Bold" italic="true"|italic="true" font_name="/ITC Stone Serif Std Bold"%;
    print;
}

Very simple minded - might need a global qualifier, might need a more complex substitute if other parts of the same line could contain italic options.

Also - a thought - should you take this opportunity to make the notation uniform, so always put italic in front of (or behind) the font name?

Answer 5

Pattern: /(<_char(?=(?:\s+\w+="[^"]*")*?\s+font_name="[^"]*?ITC Stone Serif Std Bold[^"]*")(?:\s+\w+="[^"]*")*?\s+italic=")true(?=")/
Replacement: '$1false'

Answer 6

Perl 5.10

Using new features of Perl 5.10.

s(
 <_char \s* [^>]*? \K (?: (?&font) \s+ (?&italic) | (?&italic) \s+ (?&font) )
 (?(DEFINE)
  (?<font>font_name="/ITC[ ]Stone[ ]Serif[ ]Std[ ]Bold")
  (?<italic>italic="true")
 )
){
 $+{font} . 'italic="false"'
}xge

Warning: not tested.