2008/12/7 Harold Fuchs <[EMAIL PROTECTED]>
> On 07/12/2008 11:16, Johnny Rosenberg wrote: > >> I have tried almost everything, but I just can't get this particular >> search >> to find anything (I know it's there to find, because I can find it >> manually…). >> >> The search string I want to find looks like this, kind of: >> >> The text "</configItem>", sometimes followed by a space but not necessary. >> Then a paragraph break, then zero or more spaces and then the text >> "</layout>". >> >> Here's an example, I don't know if all spaces etc makes it through all the >> email systems on the way, though: >> >> <languageList><iso639Id>swe</iso639Id></languageList> >> *</configItem> >> </layout> >> * <layout> >> >> Those bold characters is supposed to be found in this case. >> >> I checked the Regular expression check box, of course. One of the things >> that I tried was the following: >> >> </configItem>[:space:]*$[:space:]*</layout> >> >> It seems like whenever I put something after the $, it won't find >> anything. >> So obviously that's not hte way to go, but what is? >> >> J.R. >> >> > I am fairly certain that Regular Expressions (REs) can't be used to search > across paragraph boundaries. I believe that this is true of pretty much any > editor that supports REs. (For plain text editors substitute "line ends" for > "paragraph boundaries".) > > So you need a kludge. > > 1. Use REs to change *all* the paragraph marks in your text to some > (preferably short) string that you are sure doesn't occur anywhere in that > text: > a) in the Search box enter "$" without the quotes > b) in the Replace box enter the "magic" string, for example "zxcvb", > without the quotes. > > 2. Now use REs to find your pattern but with the "$" replaced by our magic > string, so "</configItem>[:space:]*zxcvb[:space:]*</layout>", again without > the quotes. > > 3. Having found the strings you want and done to them whatever you want, > use REs to change *all* the magic strings back to paragraph marks: > a) In the Search box enter "zxcvb" - no quotes > b) In the Replace box enter "\n" > > Of course you could do step 2 & 3 for each separate instance by only > finding the "next" instance of your pattern, working on it and changing its > magic string back to a paragraph mark: > a) change all paragraph marks to magic string > b) do > find next pattern including magic string > make any changes in this neighbourhood > change this magic string to a paragraph mark > until done > > The problem becomes *much* more complicated if the two parts of your > pattern can be separated by multiple lines each of which might contain text; > I haven't tried to solve that problem because you didn't pose it ;-) I'm not > at all sure that I can :-( > > -- > Harold Fuchs > London, England > Please reply *only* to users@openoffice.org > > Thanks, that worked, but it was a bit hard to read when there was text all over the place. Actually I solved it myself before reading your answer. Since the text I was searching was imported from a text file (/usr/share/X11/xkb/rules/evdev.xml) I found that I could select CR as paragraph break rather than LF, which made it possible to use \n in the search string. If there was a wey to search and replace paragraph breaks by line breaks (or whatever they are called - Shift+Enter) it would be an acceptable solution. Maybe I should write a bug report about it, asking for this as an enhancement, being able to seach for paragraph breaks. The following search string, for instance, was allowed, there would not be a problem (at least that I can see): Text On One Line$Text On The Next Line I don't really know why this isn't allowed. I would actually consider it a bug. J.R.
