File search question

[lvanhout]

Monks, I am reading a file looking for a specific flagged line. Sample 1 will loop through the entire file never activating the IF() statement. I fixed with sample 2 which stops reading the file at line 4 "testpoint\n". I do not understand why sample 1 doesn't work. I tried several variations including using chomp() . Am I missing something simple like a white space char?

sample 1:
while(<FILE>)
{    if($_ eq "testpoint\n")
    {    last;
    }
}

sample 2:
while(<FILE>)
{    if(/testpoint/)
    {    last;
    }
}

File contents being read:
<begin file>
This is a test file
data at the front of file
more test data
testpoint
data after the test
Last line
<end of file>

Lane

*note: fixed typo eg => eq happened while simplifying example
also fixed the ? => " happened when cutting a pasting from open office

[neilwatson]

You have eg instead of eq. Also, would using an regex not be a better solution? if (m/testpoint\n/){ ...

[davido]

Using 'eg' where you mean 'eq' is the first problem.

The second problem is that you're looking for exactly "testpoint\n", but in your question you told us that the line you want to stop on is "?testpoint?\n" (I assumed the \n since you're not chomping). Notice the '?' symbols? 'eq' tests for equality. Equality is not just the existance of certain text within a string, but the existance thereof to the exclusion of anything else.

Later in your post, you show example text without the '?' symbols. Which is it? You do have to be specific when you're using 'eq'. There could also be trailing whitespace not visually apparent after 'testpoint', which could be throwing off 'eq'.

[graff]

If the behavior is as you describe with the typos fixed, then the problem may have to do with line-termination issues. Maybe line 4 of the file actually contains "testpoint\r\n" (i.e. a CRLF) instead of just "\n". If the perl script is running on the same OS that created the text file, then this ouqht to work:

while (<FILE>) {
    chomp;
    if ( $_ eq "testpoint" ) {
        last;
    }
}

But if the text file was made on one OS and the script is running under a different OS (and if it's not gauranteed that the file has not gone through a dos2unix or unix2dos filter), then you're better off using the regex approach. To get effectively the same behevior as a working "eq", just anchor the beginning and ending of the regex:

while (<>) {
    if ( /^testpoint\s*$/ ) {
        last;
    }
}

[lvanhout]

I have embraced the regex approach. Thanks for the input.

Lane