CODEWITHSUNDEEP
bashreplaceawkterminal
Bastian
Bastian

Reputation: 4738

Why does my find command not work?

For our IOS project I want to merge old style pragma lines like

#pragma mark -
#pragma mark Getters

into single lines like

#pragma mark - Getters

After doing some research on the Internet, I came up with the following terminal command:

find . -type f -exec awk '{if (sub(/#pragma mark -$/,"#pragma mark -")) printf "%s", $0; else print $0}' {} > {} \;

This is be the first command which would merge the two lines. I would then use another one to delete the redundant #pragma mark - characters. However, above command does not terminate. It does not change anything, at all.

Upvotes: 1

Views: 112

Answers (3)

Aaron Digulla
Aaron Digulla

Reputation: 328760

When you press enter, the shell first processes any redirections, in this case > {}. That happens before the shell creates the process for find since it needs to connect the output of the find process to {}. Which means you should find a file {} in the current folder.

I think you're better off with a loop in this case:

find . -type f -print0 | while IFS= read -r -d $'\0' file
do
   awk ... "$file" > "$file"
done

but there is a catch: Again, the shell will first do the redirection. That means it will create an empty $file as output for awk and then start awk which will then commence to read said empty file. A more simple way to achieve the same thing would be echo -n > "$file".

So you really need to write to a temporary file and then rename:

find . -type f -print0 | while IFS= read -r -d $'\0' file
do
   awk ... "$file" > "$file.tmp" && mv "$file.tmp" "$file"
done

And make sure you have a backup of everything before you run the command because it might go horribly wrong. For example, if you have a hidden folder from your version control, find will go in there and awk will then destroy a few important bits.

PS: If you use an IDE, enable regular expressions and search for #pragma mark -\n#pragma mark Getters. Now you can replace the two lines with a single one and your IDE will make 95% sure that the string isn't replaced in places where you don't want it to happen.

Upvotes: 3

djhaskin987
djhaskin987

Reputation: 10087

This should work:

#!/usr/bin/python
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
from __future__ import unicode_literals

import re
import sys

array = [] pragmas = [] with open("my.file", "r") as ins:
    i = 0
    lastPragmaLine = 0
    for line in ins:
        m = re.search(r'^\s*[#]\s*pragma\s+mark\s+(.*\S)\s*$', line)
        if m:
           pragmas.append(m.group(1))
           lastPragmaLine = i
        else:
           array.append(line.rstrip('\n\r'))
           i += 1
    array.insert(lastPragmaLine, '#pragma mark ' + ' '.join(pragmas))

with open("my.output.file", "w") as ins:
    for line in array:
        print(line, file=ins)

Copy the contents into foo.py and run python foo.py on your machine (assuming it has a python interpreter, which most do).

Upvotes: 0

Corentin Peuvrel
Corentin Peuvrel

Reputation: 340

If your awk version is >=4.1.0, you can do 

find . -type f -exec awk -i inplace '{if (sub(/#pragma mark -$/,"#pragma mark -")) printf "%s", $0; else print $0}' {} \;

Upvotes: 1

Related Questions