Recursive string search

[Psykospot]

So, I would like to recursively search through a set of files in multiple directories for a specific set of strings and then output the data to a .csv file.

For instance, I'm looking for things in .idl files formatted like:

%msg_name{MSG_BLAH}
%msg_length{MSG_LEN_BLAH}
%msg_type{MSG_TYPE_BLAH}

And then output the data within the braces into a .csv file.

Looking for the files is just a matter of using find or awk, but I'm a bit lost from there, I think. Also, all the lines I need to look at are preceded by a "%," so that's how I'm searching for what I need.

Any ideas oh wise folk? Thanks in advance!

[Schmidt]

So, I would like to recursively search through a set of files in multiple directories for a specific set of strings and then output the data to a .csv file.

For instance, I'm looking for things in .idl files formatted like:

%msg_name{MSG_BLAH}
%msg_length{MSG_LEN_BLAH}
%msg_type{MSG_TYPE_BLAH}

And then output the data within the braces into a .csv file.

Looking for the files is just a matter of using find or awk, but I'm a bit lost from there, I think. Also, all the lines I need to look at are preceded by a "%," so that's how I'm searching for what I need.

Any ideas oh wise folk? Thanks in advance!

you can write a simple perl script, that will consist only one subroutine with about 10 lines of code here all you need:
'opendir' sub - see `perldoc -f opendir`

'readdir' sub - see `perldoc -f readdir`

file test operators : -d -f (ex. -d /usrdir returns true if usrdir is a directory, -f checks if argument is a regular file) - see `perldoc perlop`

then, in general, you make a subroutine that reads current (start) directory and proceeds regular files. If it meets a directory, then subroutine calls itself with this directory as argument. Of course you should save working dirs and restore it inside the subroutine. So, that's all