Shell Programming and Scripting

Dear all,

I am stuck while trying to add a string to multiple lines. Let me try to explain using an example:
Input:
--------
myExample_send ("MasterSends",
n,
"Data Type",
MPI_INT);

correct Output:
-----------------
#pragma my my_Example_send ("MasterSends",
#pragma my n,
#pragma my "Data Type",
#pragma my MPI_INT);


I search for "myExample_" and want to replace it with "#pragma my myExample_". My perl script works fine if the function call doesn't exceed one line. See below a section of the perl script I use:
******************************************
# Original string
$original = "myExample_";
# New string
$replacement = "#pragma my myExample_";
undef $/;

# Open each file otherwise print error message
foreach $file (@ARGV)
{
if ( ! open(INPUT,"<$file") )
{
print STDERR "Can't open file $file\n";
next;
}

# Read input file as one long record
$data=<INPUT>;
close INPUT;


# replace original string with new string
if ( $data =~ s/$original/$replacement/gi )
{
....
}
}
******************************************

However, if the function call exceeds one line, then obviously I get something similar to:
Current Output:
------------------
#pragma my myExample_send ("MasterSends",
n,
"Data Type",
MPI_INT);

Which would result in an error while compilation.

Can anybody "coach" me how to add "#pragma my " to all lines belonging to the myExample_send function call?

Thanks a lot in advance!

Hi,
You are reading only one line with $data.
Use @data instead. (This will read the entire file )

Thanks for the reply!
However, "@data" doesn't fix the problem I described above, does it?

@data=<INPUT>;
.
.

# replace original string with new string
foreach $data(@data) {
if ( $data =~ s/$original/$replacement/gi )
{
....
}
}

That's exactly what I've tried!
Nevertheless, the output is still:
------------------------------
#pragma my myExample_send ("MasterSends",
n,
"Data Type",
MPI_INT);
------------------------------

instead of:
------------------------------
#pragma my my_Example_send ("MasterSends",
#pragma my n,
#pragma my "Data Type",
#pragma my MPI_INT);
------------------------------

I think "undef $/;" already ensures that I read the entire file instead of reading line by line.

However, regarding the output I am looking for:
My thinking is that I also have to look for ";" and keep adding "#pragma my " until ";" appears.
But then I probably *have to* do it line by line, right?

Any thoughts out there?
Please let me know!
Thanks a lot!