Complex line parse - need help.

Question

Could you help me to correct my parse code.

constant fixup privite ConfigAlarms = <U1 0>         /*  comment here*/

Slip of my code below, I don't know how to get the value of U1 and 0 and **/* comment here*/**.

if(/^\s*(constant)\s*(fixup|\/\*fixup\*\/|)\s*(private|)\s*(\w+)^ ...(Need help here.)
{
    $name1 = $1;  # for content
    $name2 = $2;  # for fixup
    $name3 = $3;  # for privite 
    $name4 = $4;  # for ConfigALarms
    $name5 = $5;  # for U1
    $name6 = $6;  # for 0
    $name7 = $7;  # for /*  comment here*/

Thank you for your guide

Updated the completed code

if(/^\s*(constant)\s*(fixup|\/\*fixup\*\/|)\s*(private|)\s*(\w+)\s+=\s+<([a-zA-Z0-9]+)\s+([0-9]+)>\s*\/\*\s*(.*?)\s*\*\/(\r|\n|\s)/)

Answer 1

Take your input:

constant fixup privite ConfigAlarms = <U1 0>         /*  comment here*/

And start replacing the variables with regular expressions while leaving the constants alone (also you need to escape Perl stuff):

constant fixup privite (\w+) = <([a-zA-Z0-9]+) ([0-9]+)>         \/\*  (.*?) \*\/

Then you can make it robust against spaces:

constant\s+fixup\s+privite\s(\w+)\s+=\s+<([a-zA-Z0-9]+)\s+([0-9]+)>\s+\/\*(.*?)\*\/

Your variables are now in $1, $2, and $3.

Answer 2

You have a mismatch between "private" in your regex and "privite" in your data.

After (\w+) there shouldn't be a ^; instead, you would have

\s*=\s*\<\s*(\S+)\s+(\S+)\s*\>\s*\/\*\s*(.*?)\s*\*\/

or something like that, assuming there are always two things inside the <> and those things can have anything but spaces. If those assumptions are incorrect, you need to provide more information about what your data may look like.

Answer 3

Ouch. First, there is an x modifier that will allow non-significant whitespace in your expression, so that you can make it more readable:

my $pattern = qr{
    ^\s*
    constant\s*
    #etc
}x;

(Notice the x modifier after the ending curly brace.) Next, as already demonstrated by the previous example, you can pick a delimiting character for the pattern, so that you don’t have to escape slashes in the expression:

my $pattern1 = qr/delimited by slashes/;
my $pattern2 = qr{delimited by curly braces};

And if you want to just group an expression without capturing it into $1, $2 etc., you can use the (?:…) construct:

if ('foo bar' =~ /(?:foo)\s*(bar)/) {
    say $1; # bar
}

Or, as of Perl 5.10, you can use named captures:

if ('foo bar' =~ /(?<name>foo)\s*bar/) {
    say $+{name}; # foo
}

This all should make your regexp much more readable. The main point of your question must have been answered by others by now :)

Answer 4

don't have to get too complicated with regex. what you just need are simple ones.

$string= "constant fixup privite ConfigAlarms = <U1 0>         /*  comment here*/";
@s = split / = / ,$string;
$s[-1] =~ s/<|>//g;  #get rid of the < and >
print $s[-1];

the above will give you

$ perl perl.pl
U1 0         /*  comment here*/

you can continue from here, either by splitting again , or using a simple group matching

@ss = split /\s+/ , $s[-1];
print Dumper(@ss);

first 2 element values will be U1 and 0, the rest you can join up (or splice) and will be your comments