Linux

I've been using this to search an entire directory recursively for a specific phrase in my code (html, css, php, javascript, etc.):

Code:

find dir_name -type f -exec grep -l "phrase" {} \;

The problem is that it searches ALL files in the directory 'dir_name', even binary ones such as large JPEG images. With a directory also containing large number of images dispersed everywhere, it dramatically reduces the efficiency. Sometimes it takes very long to get a response.

Is there any parameter I can easily use to make it only search text files?

Why not use the -name primary to find?
Check man find.

There is no find option for picking or excluding text files. By convention, JPEG files usually end in .jpgor.jpeg, but it is just a convention and even following convention it could be in uppercase or lowercase. If you know that the JPEG images in your file hierarchy all have the extension.jpg, you could exclude them from the search using:

Code:

find dir_name -type f ! -name '*.jpg' -exec grep -l "phrase" {} +

Note also the+instead of\;at the end of the -exec primary. That will produce the same output, but will invoke grep many fewer times (which will significantly improve performance when processing lots of files).

If your source files all use common naming conventions you could restrict your search to just those files using something like:

Code:

find dir_name -type f \( -name '*.htm' -o -name '*.html' ... \) -exec grep -l "phrase" {} +

where the...is replaced by other-o -name '*.ext'sequences specifying the rest of the filename extensions matching your set of source files.