UNIX for Dummies Questions & Answers

Hello. I am an older newbie trying to learn Unix. I have a task to perform and it entails counting lines of code. Currently, I am pointing to the directory where the files are contained and performing a 'find' on the file extensions (cpp, c, html, java, etc.) and piping that info with a 'wc -l'. But we all know that 'wc' counts every line and I only want to count lines of actual code.

What I would like to know is of the awk, sed, grep/egrep tools, which would be easier and accurate to truncate the files by removing comments, blank spaces and going through the entire line (some lines will have code three tabs in and such)? Once I get that, I want to do wc -l. Any help would be appreciated.

The tools are as accurate as one defines the regular expressions to identify code and parse it correctly. They are working usually flawless, which means they are always right. If something isn't working and I start to get grey hairs about a problem to solve, I can be sure, that sed, awk or grep is always right in the end
Easier... hm, maybe awk is looking a bit more structured and has commands like other languages have. sed is like ed with only 1 letter for a function and can look somewhat messed up if complex regexp's are written.
If you start with small examples, you will get used to both, I bet, usually it takes it's time though like with all things
You can do a lot of tasks with sed as you can do with awk. The easy thing with awk is, that it is by default designed to parse fields and has therefore variables like NF, FS, OFS, .. and some more and is very readable compared to sed.

So here a short list, when what to use (other people might use the tools with other priorities) basically:

grep --- search every line of a file for some content
egrep/grep -E --- for using regular expressions
sed --- substituting, filtering, line and multiline parsing, ...
awk --- parsing fields, lines and multiline parsing, substitution, ...

For your wc -l problem, maybe try something like this for examle (asuming your commented lines start with a hash (#):

Code:

root@isau02:/data/tmp/testfeld> cat infile
line of code 1
line of code 2
# comment 1
line of code 3
line of code 4
# comment 2
line of code 5
root@isau02:/data/tmp/testfeld> for file in `find . -type f -name "infile*" -print`; do echo $file `grep -cv ^# $file`; done
./infile 5

Edit: Ie. you have to go into the use of other regular expressions for comments in c/c++ like those /* ... */ and don't forget to escape them with usually backslashes.

Okay. Now here is my issue. I am trying to count lines of code with many different files. The previous programmer created an script that creates an AWK. The programmer has two lines of code like this:

Code:

 find <directory> -name <file types> | awk '{print \wc -l \"$1}

Is there an equivalent to this using egrep or grep or something else? Also, how do the arguments help with storage? Thank you very much.