Odd S&R with non breaking space

[BobC]

Calibre 1.46 running on WindowsXP SR3.

I've got an epub where someone has hard-coded an indent at the start of each paragraph using four or five non breaking space characters.

This is how they appear in the editor :

  <p>****** At last the office was clear again.</p>

The * is a nbsp character - MR renders is thus.
I want to change these to use a css class to provide the indent :

<p class="indent">At last the office was clear again.</p>

So I constructed a S&R :

Search :<p>*{1,} 
Replace: <p class="indent">

Where the character after the <p> is the non-breaking space inserted using the editor's "Special Character" tool , with options Regex, Wrap, Dotall and applied to "All text files"

If I then click on "Replace All" the replace is only applied to the current file so I have to work my way through seventy odd files to do the job.

The same applies if I use

Search : <p>\s{3,}

Anybody any ideas why and how I can apply the S&R to "All Text Files"

BobC

EDIT : I think I have solved this - the non-breaking spaces are coded as html entities. Until I open the individual file they are not converted to a form S&R can find. By running "fix HTML" on the book it converts them all and then the S&R appears to work.

Found it by using Sigil to open the file and realised it was a html entity coding issue.

[eschwartz]

You can also use

( )+

to search for multiple entity non-breaking spaces.

Or use

(*| )+

(where asterisk is unicode nbsp) to find either one.

(The plus and {1,} do exactly the same thing. opinion: plus is easier to read.)

[BobC]

Constructing a search string isn't a problem - the problem I hit was recognising that what I could see when viewing the file (unicode non-breaking spaces) wasn't what was in unviewed files so wouldn't be found by an "all files" search. I must include "beautify all files" in my workflow before doing any global S&R in the future, this will force the conversion. Having the editor automatically beautify the files on opening masked the issue - viewed files had been converted, unviewed ones hadn't.

Being pedantic wouldn't your

( )+

search find single as well as multiple non-breaking spaces ? I tend to use the
{2,} format for multiple spaces which is why I also use that format ({1,}) for "one or more", just habit I suppose.

BobC