gres

Globally search for a String and Substitute

A few things to note:

No, the acronym isn't quite right but it sounds better than gss.
We don't handle regexes on purpose; I often don't even want to use them at all and I find it very annoying to place backslashes everywhere when using sed. This may change in the future, but it also may not.
There is no distinction between replacing all occurences and the first occurence. I.e., gres always replaces globally.

Options:

-v for verbose mode

There is more commentary available at the top of the source file.

Why not just use `sed`?

A couple of reasons that I made my own script instead of just using sed:

With sed, you have to manually iterate through a directory to replace a string of text. E.g.:

$ mkdir tmp
$ echo "foo" > tmp/file
$ sed -i "" "s/foo/bar/" tmp
sed: tmp: in-place editing only works for regular files
$ for file in tmp/*
for> sed -i "" "s/foo/bar/" $file
$ cat tmp/file
bar

^ And that approach only does a search with a maxdepth of 1. So, if you don't want to manually cd into each subdirectory that you want to affect, then you have to utilize something like find with sed:

$ tree tmp # here i'd want to replace the text in each of these files:
tmp
├── foo
└── tmp
    ├── bar
    └── tmp
        └── baz

2 directories, 3 files
$ find tmp -type f -exec sed -i "" "s/original string/new string/" {} \;

^ I personally find that cumbersome, even though it handles the maxdepth problem. So, as you can see, recursively searching with sed sux, at least in my opinion.

With gres, you just feed it a directory and it does the recursion for you (a la grep):

$ mkdir tmp
$ echo "foo" > tmp/file
$ gres "s/foo/bar/" tmp
$ cat tmp/file 
bar

It also uses the stdlib's Find#find method (which is effectively the same as the find program), so you don't need to worry about the maxdepth, since it assumes the default behavior of find and recursively searches through directories and their subdirectories.

Another reason I did this was so that I could replace text in files without sed adding that newline at the end. I know that is part of the POSIX definition of a file, but that newline just irks me sometimes.

ylluminarious/gres

gres

Why not just use sed?

Why not just use `sed`?