Hi,
I have a bunch of codes (like 80), every code has several subroutines (There are altogether almost 500 subroutines). The task I was posed is to add C preprocessing to each subroutine (add #ifdef before each subroutine). I think this can be done using awk script. Could anybody help me on this? Please send email to julytli01@yahoo.com, thanks a lot!

Do you want to add and '#ifdef' statement only ?
You do you identify a subroutine ?

Give us an example of input code and requested result.

Look at the following example:

subroutine 1

...
return
end

what I want to do is to produce

#ifdef sub1
subroutine 1

...
return
end
#endif

But there are about 500 subroutines to do, so I want to use a smarter way to do that. I think the key issue in awk script is how to identify the string"subroutine", but right now I don't know how to do that.
Many thanks.

Try and adapt the following script (save your files before) :

Usage: the_script input_file(s)

With the following input file :
the result is :

Do I need to add the line" for code_file ..." in the script? I named the following script as awk.txt and the input file as input, then type:
./awk.txt input
It does not work, I replaced Specify input file(s) with "input", should I replace code_file with something? could you be more specific?
Many thanks.

You'll need the code block as posted in its entirety. 'code_file' is the name of a variable to which the values of your input parameters will be assigned (i.e. your input files). What error are you getting? Your script is probably not executable. Try 'chmod 755 awk.txt' and then run it again.

my script is executable. When I run it, it has the error information like:
"line 2: syntax error near unexpected token 'do". I checked with line 2 and did not find ', there is only do in line 2. I copied the whole script as you pasted and the only change is to change "specify input file(s)" to input, where my input is the simple two subroutines as you used.
Thanks a lot!

The error must be in the for statement.
Show us your modification.


Restor as posted, and execute the script with the input filename(s) as argument.

${*:?"Specify input file(s)"}
Substitute argument passed. If no argument on command line then print the error message "Specify input file(s)"

The whole content of the script I used is:

for code_file in ${*:?"Specify input file(s)"}
do
echo "Processing file $code_file ..."
temp_file=$code.tmp
awk '
$1=="subroutine" {
print "#ifdef sub" $2;
}
{
print $0;
prv_statement = new_statement;
new_statement = $1;
}
$1=="end" && prv_statement == "return" {
print "#endif";
}
' $code_file > $temp_file
if [ $? -eq 0 ]
then mv $temp_file $code_file
else echo "Failed !!!"
fi
done

exactly the same as yours. I saved the above script as awk.txt and change it to executable, then type ./awk.txt. It still has the error information as before: line 2 syntax error near unexpected token 'do.
Is there any format requirement on script? like in makefile, the command must start with a Tab.
I don't understand why.

There is no format requirement on script.
What is you loging shell ?

Modify the script and test :

Then it will have the following error information:
bash ./awk.txt bad interpreter: No such file or directory.
What's the shell requirement on running an awk script? Sorry to bother you so much, I am a beginner on this. thanks.

The script is written for sh, bash or ksh.
Awk can run whith every shell.

Enter the command : which sh

The result will be the full path of the shell interpreter for 'sh'
This is this path that must be specified in the first line of the script.

To determine what shell is yours, enter the command : set | grep -i shell