Cfrtjryk

I have bunch of source files with very usual structure: some comments in header, some (optional) imports, and then source code, e.g.:

//

//  AppDelegate.swift

//  settings

//

//  Created by Mikhail Igonin on 14/06/2018.

//  Copyright © 2018 Mikhail Igonin. All rights reserved.

// 





import UIKit

import Fabric

import Crashlytics





@UIApplicationMain

class AppDelegate: UIResponder, UIApplicationDelegate {

    //Other comment

}

I need to add another import after comments and import block. So regex to match beginning of this file should look like this:

(([ns]*)((//.*n)|(import.*n)))+

And looks like this regex is ok: https://www.regextester.com/index.php?fam=106706

Now I'm trying to inset new import with awk and gensub:

gawk -v RS='^$' '{$0=gensub(/(([ns]*)((//.*n)|(import.*n)))+/,"\1\2nimport NEW_IMPORTn\2",1)}1' test.swift

However it doesn't work and my regex match all file:

//

//  AppDelegate.swift

//  settings

//

//  Created by Mikhail Igonin on 14/06/2018.

//  Copyright © 2018 Mikhail Igonin. All rights reserved.

//





import UIKit

import Fabric

import Crashlytics





@UIApplicationMain

class AppDelegate: UIResponder, UIApplicationDelegate {



}



import NEW_IMPORT

What's my mistake? Looks like .* works incorrect and match all file. I've tried to mark it as lazy (.*?) but without success also.

PS Solutions without awk or gensub would be also useful.

May not be the most efficient for large files, but inserts the new import line after the LAST import ... statement found:

tac file | awk '/^import/ && !I {I = 1; print "import New_IMPORT"} 1' | tac

Your mistake is to assume that . (as in .*) doesn't match a newline in awk: it does. That's different from sed, grep, perl, javascript, etc. (Think of awk regexps as always greedy and with the //s flag always on).

Setting RS to ^$ will cause awk to slurp the whole file as a single record, and then the //.*n pattern will match from the first // up to the last newline in the file; the import.*n branch won't even be considered.

There are probably smarter ways to do it, but this will insert the import NEW_IMPORT line after the first run of import lines from your file:

awk '!e&&/^import/,e=!/^import/{if(e)print "import NEW_IMPORT"}1' test.swift